On Tue, Jun 19, 2018 at 04:16:42PM +0200, Peter Zijlstra wrote: > On Tue, Jun 19, 2018 at 02:38:45PM +0100, Quentin Perret wrote: > > But maybe I could use something simpler than a lock in this case ? > > Would WRITE_ONCE/READ_ONCE be enough to ensure that atomicity for > > example ? > > Yes, since its a single pointer, smp_store_release() + READ_ONCE() > should be sufficient (these are the foundations of RCU).
Note that per_cpu()/this_cpu_read() and friends should imply READ_ONCE().
