This series removes two smp_mb__before_spinlock() (ab)users and converts the
scheduler to use smp_mb__after_spinlock(), which provides more guarantees with
the same amount of barriers.

Reply via email to