Messages in this thread | | | Date | Wed, 4 Nov 2015 13:02:27 +0100 | From | Peter Zijlstra <> | Subject | Re: [PATCH v2] ipc/msg: Implement lockless pipelined wakeups |
| |
On Tue, Nov 03, 2015 at 04:03:29PM +0100, Sebastian Andrzej Siewior wrote: > @@ -932,58 +924,26 @@ long do_msgrcv(int msqid, void __user *buf, size_t bufsz, long msgtyp, int msgfl > rcu_read_lock(); > > /* Lockless receive, part 2:
This is a broken comment style, please fix that while you're there anyway.
Also, the comment above ("Lockless receive, part 1:") is broken too, not only in style, but it refers to rcu_read_unlock() as disabling preemption, and avoiding preemption, which is false.
> + * The work in pipelined_send() and expunge_all(): > + * - Set pointer to message > + * - Queue the receiver task for later wakeup > + * - Wake up the process after the lock is dropped. > * > + * Should the process wake up before this wakeup (due to a > + * signal) it will either see the message and continue ... > */ > > + msg = msr_d.r_msg;
Since this is a lockless read, it should very much be READ_ONCE(), esp. since you killed the volatile on its type.
| |