Messages in this thread | | | Date | Fri, 27 Jul 2012 14:55:46 -0700 | Subject | Re: [PATCH 4/6] rbtree: faster augmented insert | From | Michel Lespinasse <> |
| |
On Fri, Jul 27, 2012 at 1:04 PM, Peter Zijlstra <peterz@infradead.org> wrote: > On Fri, 2012-07-20 at 05:31 -0700, Michel Lespinasse wrote: >> +static void augment_rotate(struct rb_node *rb_old, struct rb_node *rb_new) >> +{ >> + struct test_node *old = rb_entry(rb_old, struct test_node, rb); >> + struct test_node *new = rb_entry(rb_new, struct test_node, rb); >> + >> + /* Rotation doesn't change subtree's augmented value */ >> + new->augmented = old->augmented; >> + old->augmented = augment_recompute(old); >> +} > >> +static inline void augment_propagate(struct rb_node *rb) >> +{ >> + while (rb) { >> + struct test_node *node = rb_entry(rb, struct test_node, rb); >> + node->augmented = augment_recompute(node); >> + rb = rb_parent(&node->rb); >> + } >> +} > > So why do we have to introduce these two new function pointers to pass > along when they can both be trivially expressed in the old single > augment function?
Its because augment_rotate() needs to be a static function that we can take the address of and pass along as a callback to the tree rebalancing functions, while augment_propagate() needs to be an inline function that gets compiled within an __rb_erase() variant for a given type of augmented rbtree.
-- Michel "Walken" Lespinasse A program is never fully debugged until the last user dies.
| |