Re: [PATCH 1/3] perf callchain: Convert children list to rbtree

[Jiri Olsa <jolsa@redhat.com>]

On Mon, Sep 23, 2013 at 06:29:15PM +0900, Namhyung Kim wrote:
> Hi,
> 
> On Sun, 22 Sep 2013 13:15:59 +0200, Jiri Olsa wrote:
> > On Tue, Sep 10, 2013 at 05:24:16PM +0900, Namhyung Kim wrote:
> >> From: Namhyung Kim <namhyung.kim@lge.com>
> >
> > SNIP
> >
> >> index 2b585bc308cf..1b22b6269213 100644
> >> --- a/tools/perf/util/callchain.h
> >> +++ b/tools/perf/util/callchain.h
> >> @@ -21,10 +21,9 @@ enum chain_order {
> >>  
> >>  struct callchain_node {
> >>  	struct callchain_node	*parent;
> >> -	struct list_head	siblings;
> >> -	struct list_head	children;
> >>  	struct list_head	val;
> >>  	struct rb_node		rb_node; /* to sort nodes in an rbtree */
> >> +	struct rb_root		rb_root_in; /* sorted tree of children */
> >>  	struct rb_root		rb_root; /* sorted tree of children */
> >>  	unsigned int		val_nr;
> >>  	u64			hit;
> >> @@ -86,8 +85,6 @@ extern __thread struct callchain_cursor callchain_cursor;
> >>  
> >>  static inline void callchain_init(struct callchain_root *root)
> >>  {
> >> -	INIT_LIST_HEAD(&root->node.siblings);
> >> -	INIT_LIST_HEAD(&root->node.children);
> >>  	INIT_LIST_HEAD(&root->node.val);
> >
> > don't we miss rb_root_in init here, like:
> >
> > 	root->node.rb_root_in = RB_ROOT;
> 
> The struct callchain_root in hist entry initialized to zero so
> technically it doesn't need to be set.  But I think it's better to have
> explicit initializer.

yep, I was writing some tests for this when I noticed that,
having callchain_root defined on stack

jirka