On PREEMPT_RT, non-HARD irq_work runs in per-CPU kthreads via
run_irq_workd(), so irq_work_sync() uses rcuwait to wait for
BUSY==0.

After irq_work_single() clears BUSY via atomic_cmpxchg(), it still
dereferences @work for irq_work_is_hard() and rcuwait_wake_up().
An irq_work_sync() caller on another CPU that enters after BUSY is
cleared can observe BUSY==0 immediately, return, and free the work
before those accesses complete — causing a use-after-free.

Fix this by wrapping run_irq_workd() in guard(rcu)() so that the
entire irq_work_single() execution is within an RCU read-side
critical section. Then add synchronize_rcu() in irq_work_sync()
after rcuwait_wait_event() to ensure the caller waits for the RCU
grace period before returning, preventing premature frees.

Fixes: 810979682ccc ("irq_work: Allow irq_work_sync() to sleep if irq_work() no IRQ support.")
Suggested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Suggested-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev>
---
 kernel/irq_work.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/kernel/irq_work.c b/kernel/irq_work.c
index 73f7e1fd4ab4..bf411656c316 100644
--- a/kernel/irq_work.c
+++ b/kernel/irq_work.c
@@ -292,6 +292,12 @@ void irq_work_sync(struct irq_work *work)
 	    !arch_irq_work_has_interrupt()) {
 		rcuwait_wait_event(&work->irqwait, !irq_work_is_busy(work),
 				   TASK_UNINTERRUPTIBLE);
+		/*
+		 * Ensure irq_work_single() does not access @work
+		 * after removing IRQ_WORK_BUSY. It is always
+		 * accessed within a RCU-read section.
+		 */
+		synchronize_rcu();
 		return;
 	}
 
@@ -302,6 +308,7 @@ EXPORT_SYMBOL_GPL(irq_work_sync);
 
 static void run_irq_workd(unsigned int cpu)
 {
+	guard(rcu)();
 	irq_work_run_list(this_cpu_ptr(&lazy_list));
 }
 
-- 
2.43.0