Re: ring_buffer_attach && cond_synchronize_rcu (Was: percpu-rwsem: Optimize readers and reduce global impact)

[Peter Zijlstra]

I made that the below.

Are you OK with the Changelog edits and the added SoB ?

---
Subject: perf: Fix ring_buffer_attach() RCU sync, again.
From: Oleg Nesterov <oleg@xxxxxxxxxx>
Date: Sat, 30 May 2015 22:04:25 +0200

While looking for other users of get_state/cond_sync. I Found
ring_buffer_attach() and it looks obviously buggy?

Don't we need to ensure that we have "synchronize" _between_
list_del() and list_add() ?

IOW. Suppose that ring_buffer_attach() preempts right_after
get_state_synchronize_rcu() and gp completes before spin_lock().

In this case cond_synchronize_rcu() does nothing and we reuse
->rb_entry without waiting for gp in between?

It also moves the ->rcu_pending check under "if (rb)", to make it
more readable imo.

Cc: tj@xxxxxxxxxx
Cc: mingo@xxxxxxxxxx
Cc: der.herr@xxxxxxx
Cc: dave@xxxxxxxxxxxx
Cc: torvalds@xxxxxxxxxxxxxxxxxxxx
Cc: josh@xxxxxxxxxxxxxxxx
Cc: "Paul E. McKenney" <paulmck@xxxxxxxxxxxxxxxxxx>
Cc: Alexander Shishkin <alexander.shishkin@xxxxxxxxxxxxxxx>
Fixes: b69cf53640da ("perf: Fix a race between ring_buffer_detach() and ring_buffer_attach()")
Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx>
Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Link: http://lkml.kernel.org/r/20150530200425.GA15748@xxxxxxxxxx
---
 kernel/events/core.c |   14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -4307,20 +4307,20 @@ static void ring_buffer_attach(struct pe
 		WARN_ON_ONCE(event->rcu_pending);
 
 		old_rb = event->rb;
-		event->rcu_batches = get_state_synchronize_rcu();
-		event->rcu_pending = 1;
-
 		spin_lock_irqsave(&old_rb->event_lock, flags);
 		list_del_rcu(&event->rb_entry);
 		spin_unlock_irqrestore(&old_rb->event_lock, flags);
-	}
 
-	if (event->rcu_pending && rb) {
-		cond_synchronize_rcu(event->rcu_batches);
-		event->rcu_pending = 0;
+		event->rcu_batches = get_state_synchronize_rcu();
+		event->rcu_pending = 1;
 	}
 
 	if (rb) {
+		if (event->rcu_pending) {
+			cond_synchronize_rcu(event->rcu_batches);
+			event->rcu_pending = 0;
+		}
+
 		spin_lock_irqsave(&rb->event_lock, flags);
 		list_add_rcu(&event->rb_entry, &rb->event_list);
 		spin_unlock_irqrestore(&rb->event_lock, flags);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/