On Fri, Nov 11, 2016 at 02:00:34PM +0100, Peter Zijlstra wrote: > +static inline bool refcount_sub_and_test(int i, refcount_t *r) > +{ > + unsigned int old, new, val = atomic_read(&r->refs); > + > + for (;;) {
regardless of the sub_and_test vs inc_and_test issue, this should probably also have: if (val == UINT_MAX) return false; such that we stay saturated. If for some reason someone can trigger more dec's than inc's, we'd be hosed. > + new = val - i; > + if (new > val) > + BUG(); /* underflow */ > + > + old = atomic_cmpxchg_release(&r->refs, val, new); > + if (old == val) > + break; > + > + val = old; > + } > + > + return !new; > +}