[PATCH] rbtree: introduce three helpers about sort-order iteration

[PATCH] rbtree: introduce three helpers about sort-order iteration

[qiwuchen55]

From: chenqiwu <chenqiwu@xiaomi.com>

I noticed that there are many relatively common usages about the
sort-order iteration for rbtree, like this:

[drivers/android/binder.c]
for (n = rb_first(&proc->nodes); n != NULL; n = rb_next(n)) {
	struct binder_node *node = rb_entry(n, struct binder_node,
					    rb_node);
	......
}

This patch introduced three helpers to simplify sort-order iteration
for rbtree.

Signed-off-by: chenqiwu <chenqiwu@xiaomi.com>
---
 include/linux/rbtree.h | 45 +++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 43 insertions(+), 2 deletions(-)

diff --git a/include/linux/rbtree.h b/include/linux/rbtree.h
index 1fd61a9..26b4359 100644
--- a/include/linux/rbtree.h
+++ b/include/linux/rbtree.h
@@ -90,11 +90,52 @@ static inline void rb_link_node_rcu(struct rb_node *node, struct rb_node *parent
 	})
 
 /**
+ * rbtree_for_each - iterate in sort-order over a rbtree.
+ * @pos:	the 'rb_node *' to use as a loop cursor.
+ * @root:	'rb_root *' of the rbtree.
+ */
+#define rbtree_for_each(pos, root) \
+	for (pos = rb_first(root); pos; pos = rb_next(pos))
+
+/**
+ * rbtree_for_each_entry - iterate in sort-order over rb_root of given type.
+ * @pos:	the 'type *' to use as a loop cursor.
+ * @root:	'rb_root *' of the rbtree.
+ * @field:	the name of the rb_node field within 'type'.
+ */
+#define rbtree_for_each_entry(pos, root, field) \
+	for (pos = rb_entry_safe(rb_first(root), typeof(*pos), field); \
+	     pos; \
+	     pos = rb_entry_safe(rb_next(&pos->field), typeof(*pos), field))
+
+/**
+ * rbtree_for_each_entry_safe - iterate in sort-order over of given type
+ * allowing the backing memory of @pos to be invalidated.
+ * @pos:	the 'type *' to use as a loop cursor.
+ * @n:		another 'type *' to use as temporary storage.
+ * @root:	'rb_root *' of the rbtree.
+ * @field:	the name of the rb_node field within 'type'.
+ *
+ * rbtree_order_for_each_entry_safe() provides a similar guarantee as
+ * list_for_each_entry_safe() and allows the sort-order iteration to
+ * continue independent of changes to @pos by the body of the loop.
+ *
+ * Note, however, that it cannot handle other modifications that re-order the
+ * rbtree it is iterating over. This includes calling rb_erase() on @pos, as
+ * rb_erase() may rebalance the tree, causing us to miss some nodes.
+ */
+#define rbtree_for_each_entry_safe(pos, n, root, field) \
+	for (pos = rb_entry_safe(rb_first(root), typeof(*pos), field); \
+	     pos && ({ n = rb_entry_safe(rb_next(&pos->field), \
+			typeof(*pos), field); 1; }); \
+	     pos = n)
+
+/**
  * rbtree_postorder_for_each_entry_safe - iterate in post-order over rb_root of
- * given type allowing the backing memory of @pos to be invalidated
+ * given type allowing the backing memory of @pos to be invalidated.
  *
  * @pos:	the 'type *' to use as a loop cursor.
- * @n:		another 'type *' to use as temporary storage
+ * @n:		another 'type *' to use as temporary storage.
  * @root:	'rb_root *' of the rbtree.
  * @field:	the name of the rb_node field within 'type'.
  *
-- 
1.9.1

Re: [PATCH] rbtree: introduce three helpers about sort-order iteration

[Peter Zijlstra]

On Tue, Mar 10, 2020 at 12:02:14AM +0800, qiwuchen55@gmail.com wrote:
> +/**
> + * rbtree_for_each_entry_safe - iterate in sort-order over of given type
> + * allowing the backing memory of @pos to be invalidated.
> + * @pos:	the 'type *' to use as a loop cursor.
> + * @n:		another 'type *' to use as temporary storage.
> + * @root:	'rb_root *' of the rbtree.
> + * @field:	the name of the rb_node field within 'type'.
> + *
> + * rbtree_order_for_each_entry_safe() provides a similar guarantee as
> + * list_for_each_entry_safe() and allows the sort-order iteration to
> + * continue independent of changes to @pos by the body of the loop.
> + *
> + * Note, however, that it cannot handle other modifications that re-order the
> + * rbtree it is iterating over. This includes calling rb_erase() on @pos, as
> + * rb_erase() may rebalance the tree, causing us to miss some nodes.
> + */
> +#define rbtree_for_each_entry_safe(pos, n, root, field) \
> +	for (pos = rb_entry_safe(rb_first(root), typeof(*pos), field); \
> +	     pos && ({ n = rb_entry_safe(rb_next(&pos->field), \
> +			typeof(*pos), field); 1; }); \
> +	     pos = n)

Since this cannot deal with rb_erase(), what exactly is the purpose of
this thing?

Re: [PATCH] rbtree: introduce three helpers about sort-order iteration

[chenqiwu]

On Mon, Mar 09, 2020 at 05:55:11PM +0100, Peter Zijlstra wrote:
> On Tue, Mar 10, 2020 at 12:02:14AM +0800, qiwuchen55@gmail.com wrote:
> > +/**
> > + * rbtree_for_each_entry_safe - iterate in sort-order over of given type
> > + * allowing the backing memory of @pos to be invalidated.
> > + * @pos:	the 'type *' to use as a loop cursor.
> > + * @n:		another 'type *' to use as temporary storage.
> > + * @root:	'rb_root *' of the rbtree.
> > + * @field:	the name of the rb_node field within 'type'.
> > + *
> > + * rbtree_order_for_each_entry_safe() provides a similar guarantee as
> > + * list_for_each_entry_safe() and allows the sort-order iteration to
> > + * continue independent of changes to @pos by the body of the loop.
> > + *
> > + * Note, however, that it cannot handle other modifications that re-order the
> > + * rbtree it is iterating over. This includes calling rb_erase() on @pos, as
> > + * rb_erase() may rebalance the tree, causing us to miss some nodes.
> > + */
> > +#define rbtree_for_each_entry_safe(pos, n, root, field) \
> > +	for (pos = rb_entry_safe(rb_first(root), typeof(*pos), field); \
> > +	     pos && ({ n = rb_entry_safe(rb_next(&pos->field), \
> > +			typeof(*pos), field); 1; }); \
> > +	     pos = n)
> 
> Since this cannot deal with rb_erase(), what exactly is the purpose of
> this thing?

It's just a copy of rbtree_postorder_for_each_entry_safe() for
the usage scenario of sort-order iteration. It can be used for
walking the tree and free all entries of the given type.