Re: [PATCH 0/6] locking/rtmutex: Avoid PI state recursion through sched_submit_work()

[Sebastian Andrzej Siewior <bigeasy@linutronix.de>]

On 2023-08-16 16:58:18 [+0200], Peter Zijlstra wrote:
> I've ended up with the below, but it is quite horrible.. but let me go
> stare at the futex wreckage before trying to clean things up.

What about

diff --git a/kernel/locking/rwbase_rt.c b/kernel/locking/rwbase_rt.c
index f8a194e7ec9e9..b5e881250fec5 100644
--- a/kernel/locking/rwbase_rt.c
+++ b/kernel/locking/rwbase_rt.c
@@ -241,6 +241,8 @@ static int __sched rwbase_write_lock(struct rwbase_rt *rwb,
 	/* Force readers into slow path */
 	atomic_sub(READER_BIAS, &rwb->readers);
 
+	rt_mutex_pre_schedule();
+
 	raw_spin_lock_irqsave(&rtm->wait_lock, flags);
 	if (__rwbase_write_trylock(rwb))
 		goto out_unlock;
@@ -252,6 +254,7 @@ static int __sched rwbase_write_lock(struct rwbase_rt *rwb,
 		if (rwbase_signal_pending_state(state, current)) {
 			rwbase_restore_current_state();
 			__rwbase_write_unlock(rwb, 0, flags);
+			rt_mutex_post_schedule();
 			trace_contention_end(rwb, -EINTR);
 			return -EINTR;
 		}
@@ -270,6 +273,7 @@ static int __sched rwbase_write_lock(struct rwbase_rt *rwb,
 
 out_unlock:
 	raw_spin_unlock_irqrestore(&rtm->wait_lock, flags);
+	rt_mutex_post_schedule();
 	return 0;
 }
 
Sebastian