From: Michel Lespinasse <walken@google.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: aarcange@redhat.com, dwmw2@infradead.org, riel@redhat.com,
daniel.santos@pobox.com, axboe@kernel.dk, ebiederm@xmission.com,
linux-mm@kvack.org, akpm@linux-foundation.org,
linux-kernel@vger.kernel.org, torvalds@linux-foundation.org
Subject: Re: [PATCH 00/13] rbtree updates
Date: Wed, 11 Jul 2012 18:12:08 -0700 [thread overview]
Message-ID: <20120712011208.GA1152@google.com> (raw)
In-Reply-To: <1342012996.3462.154.camel@twins>
On Wed, Jul 11, 2012 at 6:23 AM, Peter Zijlstra <peterz@infradead.org> wrote:
> Looks nice.. How about something like the below on top.. I couldn't
> immediately find a sane reason for the grand-parent to always be red in
> the insertion case.
Do you mean the case you marked XXX ? it is actually parent that is
red, which we know because we tested that a few lines earlier.
> @@ -85,12 +104,27 @@ void rb_insert_color(struct rb_node *nod
> } else if (rb_is_black(parent))
> break;
>
> + /*
> + * XXX
> + */
> gparent = rb_red_parent(parent);
See :)
> if (parent == gparent->rb_left) {
> tmp = gparent->rb_right;
> if (tmp && rb_is_red(tmp)) {
> - /* Case 1 - color flips */
> + /*
> + * Case 1 - color flips
> + *
> + * G g
> + * / \ / \
> + * p u --> P U
> + * / /
> + * n N
> + *
> + * However, since g's parent might be red, and
> + * 4) does not allow this, we need to recurse
> + * at g.
> + */
I like these diagrams - I initially didn't think they'd work well, given the need for colors etc, but I now see that it's workable.
In __rb_erase_color(), some of the cases are more complicated than you drew however, because some node colors aren't known.
This is what I ended up with:
* 5), then the longest possible path due to 4 is 2B.
*
* We shall indicate color with case, where black nodes are uppercase and red
- * nodes will be lowercase.
+ * nodes will be lowercase. Unknown color nodes shall be drawn as red with
+ * some accompanying text comment.
*/
+ /*
+ * Case 2 - sibling color flip
+ * (p could be either color here)
+ *
+ * p p
+ * / \ / \
+ * N S --> N s
+ * / \ / \
+ * Sl Sr Sl Sr
+ *
+ * This leaves us violating 5), so
+ * recurse at p. If p is red, the
+ * recursion will just flip it to black
+ * and exit. If coming from Case 1,
+ * p is known to be red.
+ */
+ /*
+ * Case 3 - right rotate at sibling
+ * (p could be either color here)
+ *
+ * p p
+ * / \ / \
+ * N S --> N Sl
+ * / \ \
+ * sl Sr s
+ * \
+ * Sr
+ */
+ /*
+ * Case 4 - left rotate at parent + color flips
+ * (p and sl could be either color here.
+ * After rotation, p becomes black, s acquires
+ * p's color, and sl keeps its color)
+ *
+ * p s
+ * / \ / \
+ * N S --> P Sr
+ * / \ / \
+ * sl sr N sl
+ */
--
Michel "Walken" Lespinasse
A program is never fully debugged until the last user dies.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
WARNING: multiple messages have this Message-ID (diff)
From: Michel Lespinasse <walken@google.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: aarcange@redhat.com, dwmw2@infradead.org, riel@redhat.com,
daniel.santos@pobox.com, axboe@kernel.dk, ebiederm@xmission.com,
linux-mm@kvack.org, akpm@linux-foundation.org,
linux-kernel@vger.kernel.org, torvalds@linux-foundation.org
Subject: Re: [PATCH 00/13] rbtree updates
Date: Wed, 11 Jul 2012 18:12:08 -0700 [thread overview]
Message-ID: <20120712011208.GA1152@google.com> (raw)
In-Reply-To: <1342012996.3462.154.camel@twins>
On Wed, Jul 11, 2012 at 6:23 AM, Peter Zijlstra <peterz@infradead.org> wrote:
> Looks nice.. How about something like the below on top.. I couldn't
> immediately find a sane reason for the grand-parent to always be red in
> the insertion case.
Do you mean the case you marked XXX ? it is actually parent that is
red, which we know because we tested that a few lines earlier.
> @@ -85,12 +104,27 @@ void rb_insert_color(struct rb_node *nod
> } else if (rb_is_black(parent))
> break;
>
> + /*
> + * XXX
> + */
> gparent = rb_red_parent(parent);
See :)
> if (parent == gparent->rb_left) {
> tmp = gparent->rb_right;
> if (tmp && rb_is_red(tmp)) {
> - /* Case 1 - color flips */
> + /*
> + * Case 1 - color flips
> + *
> + * G g
> + * / \ / \
> + * p u --> P U
> + * / /
> + * n N
> + *
> + * However, since g's parent might be red, and
> + * 4) does not allow this, we need to recurse
> + * at g.
> + */
I like these diagrams - I initially didn't think they'd work well, given the need for colors etc, but I now see that it's workable.
In __rb_erase_color(), some of the cases are more complicated than you drew however, because some node colors aren't known.
This is what I ended up with:
* 5), then the longest possible path due to 4 is 2B.
*
* We shall indicate color with case, where black nodes are uppercase and red
- * nodes will be lowercase.
+ * nodes will be lowercase. Unknown color nodes shall be drawn as red with
+ * some accompanying text comment.
*/
+ /*
+ * Case 2 - sibling color flip
+ * (p could be either color here)
+ *
+ * p p
+ * / \ / \
+ * N S --> N s
+ * / \ / \
+ * Sl Sr Sl Sr
+ *
+ * This leaves us violating 5), so
+ * recurse at p. If p is red, the
+ * recursion will just flip it to black
+ * and exit. If coming from Case 1,
+ * p is known to be red.
+ */
+ /*
+ * Case 3 - right rotate at sibling
+ * (p could be either color here)
+ *
+ * p p
+ * / \ / \
+ * N S --> N Sl
+ * / \ \
+ * sl Sr s
+ * \
+ * Sr
+ */
+ /*
+ * Case 4 - left rotate at parent + color flips
+ * (p and sl could be either color here.
+ * After rotation, p becomes black, s acquires
+ * p's color, and sl keeps its color)
+ *
+ * p s
+ * / \ / \
+ * N S --> P Sr
+ * / \ / \
+ * sl sr N sl
+ */
--
Michel "Walken" Lespinasse
A program is never fully debugged until the last user dies.
next prev parent reply other threads:[~2012-07-12 1:12 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-07-09 23:35 [PATCH 00/13] rbtree updates Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-09 23:35 ` [PATCH 01/13] rbtree: reference Documentation/rbtree.txt for usage instructions Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-10 1:45 ` Rik van Riel
2012-07-10 1:45 ` Rik van Riel
2012-07-09 23:35 ` [PATCH 02/13] rbtree: empty nodes have no color Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-10 10:59 ` Daniel Santos
2012-07-10 10:59 ` Daniel Santos
2012-07-10 23:10 ` Michel Lespinasse
2012-07-10 23:10 ` Michel Lespinasse
2012-07-09 23:35 ` [PATCH 03/13] rbtree: fix incorrect rbtree node insertion in fs/proc/proc_sysctl.c Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-09 23:35 ` [PATCH 04/13] rbtree: move some implementation details from rbtree.h to rbtree.c Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-10 12:19 ` Michal Nazarewicz
2012-07-10 12:19 ` Michal Nazarewicz
2012-07-10 23:12 ` Michel Lespinasse
2012-07-10 23:12 ` Michel Lespinasse
2012-07-11 15:48 ` Michal Nazarewicz
2012-07-11 15:48 ` Michal Nazarewicz
2012-07-09 23:35 ` [PATCH 05/13] rbtree: performance and correctness test Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-10 12:27 ` Michal Nazarewicz
2012-07-10 12:27 ` Michal Nazarewicz
2012-07-10 23:18 ` Michel Lespinasse
2012-07-10 23:18 ` Michel Lespinasse
2012-07-11 6:14 ` Michel Lespinasse
2012-07-11 6:14 ` Michel Lespinasse
2012-07-11 19:30 ` Daniel Santos
2012-07-11 19:30 ` Daniel Santos
2012-07-09 23:35 ` [PATCH 06/13] rbtree: break out of rb_insert_color loop after tree rotation Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-09 23:35 ` [PATCH 07/13] rbtree: adjust root color in rb_insert_color() only when necessary Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-09 23:35 ` [PATCH 08/13] rbtree: optimize tree rotations in rb_insert_color() Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-09 23:35 ` [PATCH 09/13] rbtree: optimize color flips and parent fetching " Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-09 23:35 ` [PATCH 10/13] rbtree: adjust node color in __rb_erase_color() only when necessary Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-09 23:35 ` [PATCH 11/13] rbtree: optimize case selection logic in __rb_erase_color() Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-09 23:35 ` [PATCH 12/13] rbtree: optimize tree rotations " Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-09 23:35 ` [PATCH 13/13] rbtree: optimize color flips " Michel Lespinasse
2012-07-09 23:35 ` Michel Lespinasse
2012-07-11 13:23 ` [PATCH 00/13] rbtree updates Peter Zijlstra
2012-07-11 13:23 ` Peter Zijlstra
2012-07-12 1:12 ` Michel Lespinasse [this message]
2012-07-12 1:12 ` Michel Lespinasse
2012-07-12 14:09 ` Peter Zijlstra
2012-07-12 14:09 ` Peter Zijlstra
2012-07-12 14:12 ` Peter Zijlstra
2012-07-12 14:12 ` Peter Zijlstra
2012-07-13 0:39 ` Michel Lespinasse
2012-07-13 0:39 ` Michel Lespinasse
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20120712011208.GA1152@google.com \
--to=walken@google.com \
--cc=aarcange@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=axboe@kernel.dk \
--cc=daniel.santos@pobox.com \
--cc=dwmw2@infradead.org \
--cc=ebiederm@xmission.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=peterz@infradead.org \
--cc=riel@redhat.com \
--cc=torvalds@linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.