You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
kernel_samsung_sm7125/lib/rbtree.c

667 lines
18 KiB

/*
Red Black Trees
(C) 1999 Andrea Arcangeli <andrea@suse.de>
(C) 2002 David Woodhouse <dwmw2@infradead.org>
(C) 2012 Michel Lespinasse <walken@google.com>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
linux/lib/rbtree.c
*/
#include <linux/rbtree_augmented.h>
#include <linux/export.h>
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
/*
* red-black trees properties: http://en.wikipedia.org/wiki/Rbtree
*
* 1) A node is either red or black
* 2) The root is black
* 3) All leaves (NULL) are black
* 4) Both children of every red node are black
* 5) Every simple path from root to leaves contains the same number
* of black nodes.
*
* 4 and 5 give the O(log n) guarantee, since 4 implies you cannot have two
* consecutive red nodes in a path and every red node is therefore followed by
* a black. So if B is the number of black nodes on every simple path (as per
* 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. Unknown color nodes shall be drawn as red within
* parentheses and have some accompanying text comment.
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
*/
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
/*
* Notes on lockless lookups:
*
* All stores to the tree structure (rb_left and rb_right) must be done using
* WRITE_ONCE(). And we must not inadvertently cause (temporary) loops in the
* tree structure as seen in program order.
*
* These two requirements will allow lockless iteration of the tree -- not
* correct iteration mind you, tree rotations are not atomic so a lookup might
* miss entire subtrees.
*
* But they do guarantee that any such traversal will only see valid elements
* and that it will indeed complete -- does not get stuck in a loop.
*
* It also guarantees that if the lookup returns an element it is the 'correct'
* one. But not returning an element does _NOT_ mean it's not present.
*
* NOTE:
*
* Stores to __rb_parent_color are not important for simple lookups so those
* are left undone as of now. Nor did I check for loops involving parent
* pointers.
*/
static inline void rb_set_black(struct rb_node *rb)
{
rb->__rb_parent_color |= RB_BLACK;
}
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
static inline struct rb_node *rb_red_parent(struct rb_node *red)
{
return (struct rb_node *)red->__rb_parent_color;
}
/*
* Helper function for rotations:
* - old's parent and color get assigned to new
* - old gets assigned new as a parent and 'color' as a color.
*/
static inline void
__rb_rotate_set_parents(struct rb_node *old, struct rb_node *new,
struct rb_root *root, int color)
{
struct rb_node *parent = rb_parent(old);
new->__rb_parent_color = old->__rb_parent_color;
rb_set_parent_color(old, new, color);
__rb_change_child(old, new, parent, root);
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
}
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
static __always_inline void
__rb_insert(struct rb_node *node, struct rb_root *root,
rbtree: cache leftmost node internally Patch series "rbtree: Cache leftmost node internally", v4. A series to extending rbtrees to internally cache the leftmost node such that we can have fast overlap check optimization for all interval tree users[1]. The benefits of this series are that: (i) Unify users that do internal leftmost node caching. (ii) Optimize all interval tree users. (iii) Convert at least two new users (epoll and procfs) to the new interface. This patch (of 16): Red-black tree semantics imply that nodes with smaller or greater (or equal for duplicates) keys always be to the left and right, respectively. For the kernel this is extremely evident when considering our rb_first() semantics. Enabling lookups for the smallest node in the tree in O(1) can save a good chunk of cycles in not having to walk down the tree each time. To this end there are a few core users that explicitly do this, such as the scheduler and rtmutexes. There is also the desire for interval trees to have this optimization allowing faster overlap checking. This patch introduces a new 'struct rb_root_cached' which is just the root with a cached pointer to the leftmost node. The reason why the regular rb_root was not extended instead of adding a new structure was that this allows the user to have the choice between memory footprint and actual tree performance. The new wrappers on top of the regular rb_root calls are: - rb_first_cached(cached_root) -- which is a fast replacement for rb_first. - rb_insert_color_cached(node, cached_root, new) - rb_erase_cached(node, cached_root) In addition, augmented cached interfaces are also added for basic insertion and deletion operations; which becomes important for the interval tree changes. With the exception of the inserts, which adds a bool for updating the new leftmost, the interfaces are kept the same. To this end, porting rb users to the cached version becomes really trivial, and keeping current rbtree semantics for users that don't care about the optimization requires zero overhead. Link: http://lkml.kernel.org/r/20170719014603.19029-2-dave@stgolabs.net Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Reviewed-by: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years ago
bool newleft, struct rb_node **leftmost,
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
void (*augment_rotate)(struct rb_node *old, struct rb_node *new))
{
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
struct rb_node *parent = rb_red_parent(node), *gparent, *tmp;
rbtree: cache leftmost node internally Patch series "rbtree: Cache leftmost node internally", v4. A series to extending rbtrees to internally cache the leftmost node such that we can have fast overlap check optimization for all interval tree users[1]. The benefits of this series are that: (i) Unify users that do internal leftmost node caching. (ii) Optimize all interval tree users. (iii) Convert at least two new users (epoll and procfs) to the new interface. This patch (of 16): Red-black tree semantics imply that nodes with smaller or greater (or equal for duplicates) keys always be to the left and right, respectively. For the kernel this is extremely evident when considering our rb_first() semantics. Enabling lookups for the smallest node in the tree in O(1) can save a good chunk of cycles in not having to walk down the tree each time. To this end there are a few core users that explicitly do this, such as the scheduler and rtmutexes. There is also the desire for interval trees to have this optimization allowing faster overlap checking. This patch introduces a new 'struct rb_root_cached' which is just the root with a cached pointer to the leftmost node. The reason why the regular rb_root was not extended instead of adding a new structure was that this allows the user to have the choice between memory footprint and actual tree performance. The new wrappers on top of the regular rb_root calls are: - rb_first_cached(cached_root) -- which is a fast replacement for rb_first. - rb_insert_color_cached(node, cached_root, new) - rb_erase_cached(node, cached_root) In addition, augmented cached interfaces are also added for basic insertion and deletion operations; which becomes important for the interval tree changes. With the exception of the inserts, which adds a bool for updating the new leftmost, the interfaces are kept the same. To this end, porting rb users to the cached version becomes really trivial, and keeping current rbtree semantics for users that don't care about the optimization requires zero overhead. Link: http://lkml.kernel.org/r/20170719014603.19029-2-dave@stgolabs.net Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Reviewed-by: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years ago
if (newleft)
*leftmost = node;
while (true) {
/*
* Loop invariant: node is red.
*/
if (unlikely(!parent)) {
/*
* The inserted node is root. Either this is the
* first node, or we recursed at Case 1 below and
* are no longer violating 4).
*/
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
rb_set_parent_color(node, NULL, RB_BLACK);
break;
}
/*
* If there is a black parent, we are done.
* Otherwise, take some corrective action as,
* per 4), we don't want a red root or two
* consecutive red nodes.
*/
if(rb_is_black(parent))
break;
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
gparent = rb_red_parent(parent);
tmp = gparent->rb_right;
if (parent != tmp) { /* parent == gparent->rb_left */
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
if (tmp && rb_is_red(tmp)) {
/*
* Case 1 - node's uncle is red (color flips).
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
*
* G g
* / \ / \
* p u --> P U
* / /
* n n
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
*
* However, since g's parent might be red, and
* 4) does not allow this, we need to recurse
* at g.
*/
rb_set_parent_color(tmp, gparent, RB_BLACK);
rb_set_parent_color(parent, gparent, RB_BLACK);
node = gparent;
parent = rb_parent(node);
rb_set_parent_color(node, parent, RB_RED);
continue;
}
tmp = parent->rb_right;
if (node == tmp) {
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
/*
* Case 2 - node's uncle is black and node is
* the parent's right child (left rotate at parent).
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
*
* G G
* / \ / \
* p U --> n U
* \ /
* n p
*
* This still leaves us in violation of 4), the
* continuation into Case 3 will fix that.
*/
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
tmp = node->rb_left;
WRITE_ONCE(parent->rb_right, tmp);
WRITE_ONCE(node->rb_left, parent);
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
if (tmp)
rb_set_parent_color(tmp, parent,
RB_BLACK);
rb_set_parent_color(parent, node, RB_RED);
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
augment_rotate(parent, node);
parent = node;
tmp = node->rb_right;
}
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
/*
* Case 3 - node's uncle is black and node is
* the parent's left child (right rotate at gparent).
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
*
* G P
* / \ / \
* p U --> n g
* / \
* n U
*/
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
WRITE_ONCE(gparent->rb_left, tmp); /* == parent->rb_right */
WRITE_ONCE(parent->rb_right, gparent);
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
if (tmp)
rb_set_parent_color(tmp, gparent, RB_BLACK);
__rb_rotate_set_parents(gparent, parent, root, RB_RED);
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
augment_rotate(gparent, parent);
break;
} else {
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
tmp = gparent->rb_left;
if (tmp && rb_is_red(tmp)) {
/* Case 1 - color flips */
rb_set_parent_color(tmp, gparent, RB_BLACK);
rb_set_parent_color(parent, gparent, RB_BLACK);
node = gparent;
parent = rb_parent(node);
rb_set_parent_color(node, parent, RB_RED);
continue;
}
tmp = parent->rb_left;
if (node == tmp) {
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
/* Case 2 - right rotate at parent */
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
tmp = node->rb_right;
WRITE_ONCE(parent->rb_left, tmp);
WRITE_ONCE(node->rb_right, parent);
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
if (tmp)
rb_set_parent_color(tmp, parent,
RB_BLACK);
rb_set_parent_color(parent, node, RB_RED);
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
augment_rotate(parent, node);
parent = node;
tmp = node->rb_left;
}
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
/* Case 3 - left rotate at gparent */
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
WRITE_ONCE(gparent->rb_right, tmp); /* == parent->rb_left */
WRITE_ONCE(parent->rb_left, gparent);
rbtree: low level optimizations in rb_insert_color() - Use the newly introduced rb_set_parent_color() function to flip the color of nodes whose parent is already known. - Optimize rb_parent() when the node is known to be red - there is no need to mask out the color in that case. - Flipping gparent's color to red requires us to fetch its rb_parent_color field, so we can reuse it as the parent value for the next loop iteration. - Do not use __rb_rotate_left() and __rb_rotate_right() to handle tree rotations: we already have pointers to all relevant nodes, and know their colors (either because we want to adjust it, or because we've tested it, or we can deduce it as black due to the node proximity to a known red node). So we can generate more efficient code by making use of the node pointers we already have, and setting both the parent and color attributes for nodes all at once. Also in Case 2, some node attributes don't have to be set because we know another tree rotation (Case 3) will always follow and override them. Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
if (tmp)
rb_set_parent_color(tmp, gparent, RB_BLACK);
__rb_rotate_set_parents(gparent, parent, root, RB_RED);
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
augment_rotate(gparent, parent);
break;
}
}
}
/*
* Inline version for rb_erase() use - we want to be able to inline
* and eliminate the dummy_rotate callback there
*/
static __always_inline void
____rb_erase_color(struct rb_node *parent, struct rb_root *root,
void (*augment_rotate)(struct rb_node *old, struct rb_node *new))
{
struct rb_node *node = NULL, *sibling, *tmp1, *tmp2;
while (true) {
/*
* Loop invariants:
* - node is black (or NULL on first iteration)
* - node is not the root (parent is not NULL)
* - All leaf paths going through parent and node have a
* black node count that is 1 lower than other leaf paths.
*/
sibling = parent->rb_right;
if (node != sibling) { /* node == parent->rb_left */
if (rb_is_red(sibling)) {
/*
* Case 1 - left rotate at parent
*
* P S
* / \ / \
* N s --> p Sr
* / \ / \
* Sl Sr N Sl
*/
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
tmp1 = sibling->rb_left;
WRITE_ONCE(parent->rb_right, tmp1);
WRITE_ONCE(sibling->rb_left, parent);
rb_set_parent_color(tmp1, parent, RB_BLACK);
__rb_rotate_set_parents(parent, sibling, root,
RB_RED);
augment_rotate(parent, sibling);
sibling = tmp1;
}
tmp1 = sibling->rb_right;
if (!tmp1 || rb_is_black(tmp1)) {
tmp2 = sibling->rb_left;
if (!tmp2 || rb_is_black(tmp2)) {
/*
* 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) which
* can be fixed by flipping p to black
* if it was red, or by recursing at p.
* p is red when coming from Case 1.
*/
rb_set_parent_color(sibling, parent,
RB_RED);
if (rb_is_red(parent))
rb_set_black(parent);
else {
node = parent;
parent = rb_parent(node);
if (parent)
continue;
}
break;
}
/*
* Case 3 - right rotate at sibling
* (p could be either color here)
*
* (p) (p)
* / \ / \
* N S --> N sl
* / \ \
* sl Sr S
* \
* Sr
*
* Note: p might be red, and then both
* p and sl are red after rotation(which
* breaks property 4). This is fixed in
* Case 4 (in __rb_rotate_set_parents()
* which set sl the color of p
* and set p RB_BLACK)
*
* (p) (sl)
* / \ / \
* N sl --> P S
* \ / \
* S N Sr
* \
* Sr
*/
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
tmp1 = tmp2->rb_right;
WRITE_ONCE(sibling->rb_left, tmp1);
WRITE_ONCE(tmp2->rb_right, sibling);
WRITE_ONCE(parent->rb_right, tmp2);
if (tmp1)
rb_set_parent_color(tmp1, sibling,
RB_BLACK);
augment_rotate(sibling, tmp2);
tmp1 = sibling;
sibling = tmp2;
}
/*
* 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)
*/
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
tmp2 = sibling->rb_left;
WRITE_ONCE(parent->rb_right, tmp2);
WRITE_ONCE(sibling->rb_left, parent);
rb_set_parent_color(tmp1, sibling, RB_BLACK);
if (tmp2)
rb_set_parent(tmp2, parent);
__rb_rotate_set_parents(parent, sibling, root,
RB_BLACK);
augment_rotate(parent, sibling);
break;
} else {
sibling = parent->rb_left;
if (rb_is_red(sibling)) {
/* Case 1 - right rotate at parent */
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
tmp1 = sibling->rb_right;
WRITE_ONCE(parent->rb_left, tmp1);
WRITE_ONCE(sibling->rb_right, parent);
rb_set_parent_color(tmp1, parent, RB_BLACK);
__rb_rotate_set_parents(parent, sibling, root,
RB_RED);
augment_rotate(parent, sibling);
sibling = tmp1;
}
tmp1 = sibling->rb_left;
if (!tmp1 || rb_is_black(tmp1)) {
tmp2 = sibling->rb_right;
if (!tmp2 || rb_is_black(tmp2)) {
/* Case 2 - sibling color flip */
rb_set_parent_color(sibling, parent,
RB_RED);
if (rb_is_red(parent))
rb_set_black(parent);
else {
node = parent;
parent = rb_parent(node);
if (parent)
continue;
}
break;
}
/* Case 3 - left rotate at sibling */
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
tmp1 = tmp2->rb_left;
WRITE_ONCE(sibling->rb_right, tmp1);
WRITE_ONCE(tmp2->rb_left, sibling);
WRITE_ONCE(parent->rb_left, tmp2);
if (tmp1)
rb_set_parent_color(tmp1, sibling,
RB_BLACK);
augment_rotate(sibling, tmp2);
tmp1 = sibling;
sibling = tmp2;
}
/* Case 4 - right rotate at parent + color flips */
rbtree: Make lockless searches non-fatal Change the insert and erase code such that lockless searches are non-fatal. In and of itself an rbtree cannot be correctly searched while in-modification, we can however provide weaker guarantees that will allow the rbtree to be used in conjunction with other techniques, such as latches; see 9b0fd802e8c0 ("seqcount: Add raw_write_seqcount_latch()"). For this to work we need the following guarantees from the rbtree code: 1) a lockless reader must not see partial stores, this would allow it to observe nodes that are invalid memory. 2) there must not be (temporary) loops in the tree structure in the modifier's program order, this would cause a lookup which interrupts the modifier to get stuck indefinitely. For 1) we must use WRITE_ONCE() for all updates to the tree structure; in particular this patch only does rb_{left,right} as those are the only element required for simple searches. It generates slightly worse code, probably because volatile. But in pointer chasing heavy code a few instructions more should not matter. For 2) I have carefully audited the code and drawn every intermediate link state and not found a loop. Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Reviewed-by: Michel Lespinasse <walken@google.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
tmp2 = sibling->rb_right;
WRITE_ONCE(parent->rb_left, tmp2);
WRITE_ONCE(sibling->rb_right, parent);
rb_set_parent_color(tmp1, sibling, RB_BLACK);
if (tmp2)
rb_set_parent(tmp2, parent);
__rb_rotate_set_parents(parent, sibling, root,
RB_BLACK);
augment_rotate(parent, sibling);
break;
}
}
}
/* Non-inline version for rb_erase_augmented() use */
void __rb_erase_color(struct rb_node *parent, struct rb_root *root,
void (*augment_rotate)(struct rb_node *old, struct rb_node *new))
{
____rb_erase_color(parent, root, augment_rotate);
}
EXPORT_SYMBOL(__rb_erase_color);
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
/*
* Non-augmented rbtree manipulation functions.
*
* We use dummy augmented callbacks here, and have the compiler optimize them
* out of the rb_insert_color() and rb_erase() function definitions.
*/
static inline void dummy_propagate(struct rb_node *node, struct rb_node *stop) {}
static inline void dummy_copy(struct rb_node *old, struct rb_node *new) {}
static inline void dummy_rotate(struct rb_node *old, struct rb_node *new) {}
static const struct rb_augment_callbacks dummy_callbacks = {
.propagate = dummy_propagate,
.copy = dummy_copy,
.rotate = dummy_rotate
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
};
void rb_insert_color(struct rb_node *node, struct rb_root *root)
{
rbtree: cache leftmost node internally Patch series "rbtree: Cache leftmost node internally", v4. A series to extending rbtrees to internally cache the leftmost node such that we can have fast overlap check optimization for all interval tree users[1]. The benefits of this series are that: (i) Unify users that do internal leftmost node caching. (ii) Optimize all interval tree users. (iii) Convert at least two new users (epoll and procfs) to the new interface. This patch (of 16): Red-black tree semantics imply that nodes with smaller or greater (or equal for duplicates) keys always be to the left and right, respectively. For the kernel this is extremely evident when considering our rb_first() semantics. Enabling lookups for the smallest node in the tree in O(1) can save a good chunk of cycles in not having to walk down the tree each time. To this end there are a few core users that explicitly do this, such as the scheduler and rtmutexes. There is also the desire for interval trees to have this optimization allowing faster overlap checking. This patch introduces a new 'struct rb_root_cached' which is just the root with a cached pointer to the leftmost node. The reason why the regular rb_root was not extended instead of adding a new structure was that this allows the user to have the choice between memory footprint and actual tree performance. The new wrappers on top of the regular rb_root calls are: - rb_first_cached(cached_root) -- which is a fast replacement for rb_first. - rb_insert_color_cached(node, cached_root, new) - rb_erase_cached(node, cached_root) In addition, augmented cached interfaces are also added for basic insertion and deletion operations; which becomes important for the interval tree changes. With the exception of the inserts, which adds a bool for updating the new leftmost, the interfaces are kept the same. To this end, porting rb users to the cached version becomes really trivial, and keeping current rbtree semantics for users that don't care about the optimization requires zero overhead. Link: http://lkml.kernel.org/r/20170719014603.19029-2-dave@stgolabs.net Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Reviewed-by: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years ago
__rb_insert(node, root, false, NULL, dummy_rotate);
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
}
EXPORT_SYMBOL(rb_insert_color);
void rb_erase(struct rb_node *node, struct rb_root *root)
{
struct rb_node *rebalance;
rbtree: cache leftmost node internally Patch series "rbtree: Cache leftmost node internally", v4. A series to extending rbtrees to internally cache the leftmost node such that we can have fast overlap check optimization for all interval tree users[1]. The benefits of this series are that: (i) Unify users that do internal leftmost node caching. (ii) Optimize all interval tree users. (iii) Convert at least two new users (epoll and procfs) to the new interface. This patch (of 16): Red-black tree semantics imply that nodes with smaller or greater (or equal for duplicates) keys always be to the left and right, respectively. For the kernel this is extremely evident when considering our rb_first() semantics. Enabling lookups for the smallest node in the tree in O(1) can save a good chunk of cycles in not having to walk down the tree each time. To this end there are a few core users that explicitly do this, such as the scheduler and rtmutexes. There is also the desire for interval trees to have this optimization allowing faster overlap checking. This patch introduces a new 'struct rb_root_cached' which is just the root with a cached pointer to the leftmost node. The reason why the regular rb_root was not extended instead of adding a new structure was that this allows the user to have the choice between memory footprint and actual tree performance. The new wrappers on top of the regular rb_root calls are: - rb_first_cached(cached_root) -- which is a fast replacement for rb_first. - rb_insert_color_cached(node, cached_root, new) - rb_erase_cached(node, cached_root) In addition, augmented cached interfaces are also added for basic insertion and deletion operations; which becomes important for the interval tree changes. With the exception of the inserts, which adds a bool for updating the new leftmost, the interfaces are kept the same. To this end, porting rb users to the cached version becomes really trivial, and keeping current rbtree semantics for users that don't care about the optimization requires zero overhead. Link: http://lkml.kernel.org/r/20170719014603.19029-2-dave@stgolabs.net Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Reviewed-by: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years ago
rebalance = __rb_erase_augmented(node, root,
NULL, &dummy_callbacks);
if (rebalance)
____rb_erase_color(rebalance, root, dummy_rotate);
}
EXPORT_SYMBOL(rb_erase);
rbtree: cache leftmost node internally Patch series "rbtree: Cache leftmost node internally", v4. A series to extending rbtrees to internally cache the leftmost node such that we can have fast overlap check optimization for all interval tree users[1]. The benefits of this series are that: (i) Unify users that do internal leftmost node caching. (ii) Optimize all interval tree users. (iii) Convert at least two new users (epoll and procfs) to the new interface. This patch (of 16): Red-black tree semantics imply that nodes with smaller or greater (or equal for duplicates) keys always be to the left and right, respectively. For the kernel this is extremely evident when considering our rb_first() semantics. Enabling lookups for the smallest node in the tree in O(1) can save a good chunk of cycles in not having to walk down the tree each time. To this end there are a few core users that explicitly do this, such as the scheduler and rtmutexes. There is also the desire for interval trees to have this optimization allowing faster overlap checking. This patch introduces a new 'struct rb_root_cached' which is just the root with a cached pointer to the leftmost node. The reason why the regular rb_root was not extended instead of adding a new structure was that this allows the user to have the choice between memory footprint and actual tree performance. The new wrappers on top of the regular rb_root calls are: - rb_first_cached(cached_root) -- which is a fast replacement for rb_first. - rb_insert_color_cached(node, cached_root, new) - rb_erase_cached(node, cached_root) In addition, augmented cached interfaces are also added for basic insertion and deletion operations; which becomes important for the interval tree changes. With the exception of the inserts, which adds a bool for updating the new leftmost, the interfaces are kept the same. To this end, porting rb users to the cached version becomes really trivial, and keeping current rbtree semantics for users that don't care about the optimization requires zero overhead. Link: http://lkml.kernel.org/r/20170719014603.19029-2-dave@stgolabs.net Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Reviewed-by: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years ago
void rb_insert_color_cached(struct rb_node *node,
struct rb_root_cached *root, bool leftmost)
{
__rb_insert(node, &root->rb_root, leftmost,
&root->rb_leftmost, dummy_rotate);
}
EXPORT_SYMBOL(rb_insert_color_cached);
void rb_erase_cached(struct rb_node *node, struct rb_root_cached *root)
{
struct rb_node *rebalance;
rebalance = __rb_erase_augmented(node, &root->rb_root,
&root->rb_leftmost, &dummy_callbacks);
if (rebalance)
____rb_erase_color(rebalance, &root->rb_root, dummy_rotate);
}
EXPORT_SYMBOL(rb_erase_cached);
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
/*
* Augmented rbtree manipulation functions.
*
* This instantiates the same __always_inline functions as in the non-augmented
* case, but this time with user-defined callbacks.
*/
void __rb_insert_augmented(struct rb_node *node, struct rb_root *root,
rbtree: cache leftmost node internally Patch series "rbtree: Cache leftmost node internally", v4. A series to extending rbtrees to internally cache the leftmost node such that we can have fast overlap check optimization for all interval tree users[1]. The benefits of this series are that: (i) Unify users that do internal leftmost node caching. (ii) Optimize all interval tree users. (iii) Convert at least two new users (epoll and procfs) to the new interface. This patch (of 16): Red-black tree semantics imply that nodes with smaller or greater (or equal for duplicates) keys always be to the left and right, respectively. For the kernel this is extremely evident when considering our rb_first() semantics. Enabling lookups for the smallest node in the tree in O(1) can save a good chunk of cycles in not having to walk down the tree each time. To this end there are a few core users that explicitly do this, such as the scheduler and rtmutexes. There is also the desire for interval trees to have this optimization allowing faster overlap checking. This patch introduces a new 'struct rb_root_cached' which is just the root with a cached pointer to the leftmost node. The reason why the regular rb_root was not extended instead of adding a new structure was that this allows the user to have the choice between memory footprint and actual tree performance. The new wrappers on top of the regular rb_root calls are: - rb_first_cached(cached_root) -- which is a fast replacement for rb_first. - rb_insert_color_cached(node, cached_root, new) - rb_erase_cached(node, cached_root) In addition, augmented cached interfaces are also added for basic insertion and deletion operations; which becomes important for the interval tree changes. With the exception of the inserts, which adds a bool for updating the new leftmost, the interfaces are kept the same. To this end, porting rb users to the cached version becomes really trivial, and keeping current rbtree semantics for users that don't care about the optimization requires zero overhead. Link: http://lkml.kernel.org/r/20170719014603.19029-2-dave@stgolabs.net Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Reviewed-by: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years ago
bool newleft, struct rb_node **leftmost,
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
void (*augment_rotate)(struct rb_node *old, struct rb_node *new))
{
rbtree: cache leftmost node internally Patch series "rbtree: Cache leftmost node internally", v4. A series to extending rbtrees to internally cache the leftmost node such that we can have fast overlap check optimization for all interval tree users[1]. The benefits of this series are that: (i) Unify users that do internal leftmost node caching. (ii) Optimize all interval tree users. (iii) Convert at least two new users (epoll and procfs) to the new interface. This patch (of 16): Red-black tree semantics imply that nodes with smaller or greater (or equal for duplicates) keys always be to the left and right, respectively. For the kernel this is extremely evident when considering our rb_first() semantics. Enabling lookups for the smallest node in the tree in O(1) can save a good chunk of cycles in not having to walk down the tree each time. To this end there are a few core users that explicitly do this, such as the scheduler and rtmutexes. There is also the desire for interval trees to have this optimization allowing faster overlap checking. This patch introduces a new 'struct rb_root_cached' which is just the root with a cached pointer to the leftmost node. The reason why the regular rb_root was not extended instead of adding a new structure was that this allows the user to have the choice between memory footprint and actual tree performance. The new wrappers on top of the regular rb_root calls are: - rb_first_cached(cached_root) -- which is a fast replacement for rb_first. - rb_insert_color_cached(node, cached_root, new) - rb_erase_cached(node, cached_root) In addition, augmented cached interfaces are also added for basic insertion and deletion operations; which becomes important for the interval tree changes. With the exception of the inserts, which adds a bool for updating the new leftmost, the interfaces are kept the same. To this end, porting rb users to the cached version becomes really trivial, and keeping current rbtree semantics for users that don't care about the optimization requires zero overhead. Link: http://lkml.kernel.org/r/20170719014603.19029-2-dave@stgolabs.net Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Reviewed-by: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years ago
__rb_insert(node, root, newleft, leftmost, augment_rotate);
rbtree: faster augmented rbtree manipulation Introduce new augmented rbtree APIs that allow minimal recalculation of augmented node information. A new callback is added to the rbtree insertion and erase rebalancing functions, to be called on each tree rotations. Such rotations preserve the subtree's root augmented value, but require recalculation of the one child that was previously located at the subtree root. In the insertion case, the handcoded search phase must be updated to maintain the augmented information on insertion, and then the rbtree coloring/rebalancing algorithms keep it up to date. In the erase case, things are more complicated since it is library code that manipulates the rbtree in order to remove internal nodes. This requires a couple additional callbacks to copy a subtree's augmented value when a new root is stitched in, and to recompute augmented values down the ancestry path when a node is removed from the tree. In order to preserve maximum speed for the non-augmented case, we provide two versions of each tree manipulation function. rb_insert_augmented() is the augmented equivalent of rb_insert_color(), and rb_erase_augmented() is the augmented equivalent of rb_erase(). Signed-off-by: Michel Lespinasse <walken@google.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
}
EXPORT_SYMBOL(__rb_insert_augmented);
/*
* This function returns the first node (in sort order) of the tree.
*/
struct rb_node *rb_first(const struct rb_root *root)
{
struct rb_node *n;
n = root->rb_node;
if (!n)
return NULL;
while (n->rb_left)
n = n->rb_left;
return n;
}
EXPORT_SYMBOL(rb_first);
struct rb_node *rb_last(const struct rb_root *root)
{
struct rb_node *n;
n = root->rb_node;
if (!n)
return NULL;
while (n->rb_right)
n = n->rb_right;
return n;
}
EXPORT_SYMBOL(rb_last);
struct rb_node *rb_next(const struct rb_node *node)
{
struct rb_node *parent;
rbtree: empty nodes have no color Empty nodes have no color. We can make use of this property to simplify the code emitted by the RB_EMPTY_NODE and RB_CLEAR_NODE macros. Also, we can get rid of the rb_init_node function which had been introduced by commit 88d19cf37952 ("timers: Add rb_init_node() to allow for stack allocated rb nodes") to avoid some issue with the empty node's color not being initialized. I'm not sure what the RB_EMPTY_NODE checks in rb_prev() / rb_next() are doing there, though. axboe introduced them in commit 10fd48f2376d ("rbtree: fixed reversed RB_EMPTY_NODE and rb_next/prev"). The way I see it, the 'empty node' abstraction is only used by rbtree users to flag nodes that they haven't inserted in any rbtree, so asking the predecessor or successor of such nodes doesn't make any sense. One final rb_init_node() caller was recently added in sysctl code to implement faster sysctl name lookups. This code doesn't make use of RB_EMPTY_NODE at all, and from what I could see it only called rb_init_node() under the mistaken assumption that such initialization was required before node insertion. [sfr@canb.auug.org.au: fix net/ceph/osd_client.c build] Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: John Stultz <john.stultz@linaro.org> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
if (RB_EMPTY_NODE(node))
return NULL;
/*
* If we have a right-hand child, go down and then left as far
* as we can.
*/
if (node->rb_right) {
rbtree: cache leftmost node internally Patch series "rbtree: Cache leftmost node internally", v4. A series to extending rbtrees to internally cache the leftmost node such that we can have fast overlap check optimization for all interval tree users[1]. The benefits of this series are that: (i) Unify users that do internal leftmost node caching. (ii) Optimize all interval tree users. (iii) Convert at least two new users (epoll and procfs) to the new interface. This patch (of 16): Red-black tree semantics imply that nodes with smaller or greater (or equal for duplicates) keys always be to the left and right, respectively. For the kernel this is extremely evident when considering our rb_first() semantics. Enabling lookups for the smallest node in the tree in O(1) can save a good chunk of cycles in not having to walk down the tree each time. To this end there are a few core users that explicitly do this, such as the scheduler and rtmutexes. There is also the desire for interval trees to have this optimization allowing faster overlap checking. This patch introduces a new 'struct rb_root_cached' which is just the root with a cached pointer to the leftmost node. The reason why the regular rb_root was not extended instead of adding a new structure was that this allows the user to have the choice between memory footprint and actual tree performance. The new wrappers on top of the regular rb_root calls are: - rb_first_cached(cached_root) -- which is a fast replacement for rb_first. - rb_insert_color_cached(node, cached_root, new) - rb_erase_cached(node, cached_root) In addition, augmented cached interfaces are also added for basic insertion and deletion operations; which becomes important for the interval tree changes. With the exception of the inserts, which adds a bool for updating the new leftmost, the interfaces are kept the same. To this end, porting rb users to the cached version becomes really trivial, and keeping current rbtree semantics for users that don't care about the optimization requires zero overhead. Link: http://lkml.kernel.org/r/20170719014603.19029-2-dave@stgolabs.net Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Reviewed-by: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years ago
node = node->rb_right;
while (node->rb_left)
node=node->rb_left;
return (struct rb_node *)node;
}
/*
* No right-hand children. Everything down and left is smaller than us,
* so any 'next' node must be in the general direction of our parent.
* Go up the tree; any time the ancestor is a right-hand child of its
* parent, keep going up. First time it's a left-hand child of its
* parent, said parent is our 'next' node.
*/
while ((parent = rb_parent(node)) && node == parent->rb_right)
node = parent;
return parent;
}
EXPORT_SYMBOL(rb_next);
struct rb_node *rb_prev(const struct rb_node *node)
{
struct rb_node *parent;
rbtree: empty nodes have no color Empty nodes have no color. We can make use of this property to simplify the code emitted by the RB_EMPTY_NODE and RB_CLEAR_NODE macros. Also, we can get rid of the rb_init_node function which had been introduced by commit 88d19cf37952 ("timers: Add rb_init_node() to allow for stack allocated rb nodes") to avoid some issue with the empty node's color not being initialized. I'm not sure what the RB_EMPTY_NODE checks in rb_prev() / rb_next() are doing there, though. axboe introduced them in commit 10fd48f2376d ("rbtree: fixed reversed RB_EMPTY_NODE and rb_next/prev"). The way I see it, the 'empty node' abstraction is only used by rbtree users to flag nodes that they haven't inserted in any rbtree, so asking the predecessor or successor of such nodes doesn't make any sense. One final rb_init_node() caller was recently added in sysctl code to implement faster sysctl name lookups. This code doesn't make use of RB_EMPTY_NODE at all, and from what I could see it only called rb_init_node() under the mistaken assumption that such initialization was required before node insertion. [sfr@canb.auug.org.au: fix net/ceph/osd_client.c build] Signed-off-by: Michel Lespinasse <walken@google.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Acked-by: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Daniel Santos <daniel.santos@pobox.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: John Stultz <john.stultz@linaro.org> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years ago
if (RB_EMPTY_NODE(node))
return NULL;
/*
* If we have a left-hand child, go down and then right as far
* as we can.
*/
if (node->rb_left) {
rbtree: cache leftmost node internally Patch series "rbtree: Cache leftmost node internally", v4. A series to extending rbtrees to internally cache the leftmost node such that we can have fast overlap check optimization for all interval tree users[1]. The benefits of this series are that: (i) Unify users that do internal leftmost node caching. (ii) Optimize all interval tree users. (iii) Convert at least two new users (epoll and procfs) to the new interface. This patch (of 16): Red-black tree semantics imply that nodes with smaller or greater (or equal for duplicates) keys always be to the left and right, respectively. For the kernel this is extremely evident when considering our rb_first() semantics. Enabling lookups for the smallest node in the tree in O(1) can save a good chunk of cycles in not having to walk down the tree each time. To this end there are a few core users that explicitly do this, such as the scheduler and rtmutexes. There is also the desire for interval trees to have this optimization allowing faster overlap checking. This patch introduces a new 'struct rb_root_cached' which is just the root with a cached pointer to the leftmost node. The reason why the regular rb_root was not extended instead of adding a new structure was that this allows the user to have the choice between memory footprint and actual tree performance. The new wrappers on top of the regular rb_root calls are: - rb_first_cached(cached_root) -- which is a fast replacement for rb_first. - rb_insert_color_cached(node, cached_root, new) - rb_erase_cached(node, cached_root) In addition, augmented cached interfaces are also added for basic insertion and deletion operations; which becomes important for the interval tree changes. With the exception of the inserts, which adds a bool for updating the new leftmost, the interfaces are kept the same. To this end, porting rb users to the cached version becomes really trivial, and keeping current rbtree semantics for users that don't care about the optimization requires zero overhead. Link: http://lkml.kernel.org/r/20170719014603.19029-2-dave@stgolabs.net Signed-off-by: Davidlohr Bueso <dbueso@suse.de> Reviewed-by: Jan Kara <jack@suse.cz> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years ago
node = node->rb_left;
while (node->rb_right)
node=node->rb_right;
return (struct rb_node *)node;
}
/*
* No left-hand children. Go up till we find an ancestor which
* is a right-hand child of its parent.
*/
while ((parent = rb_parent(node)) && node == parent->rb_left)
node = parent;
return parent;
}
EXPORT_SYMBOL(rb_prev);
void rb_replace_node(struct rb_node *victim, struct rb_node *new,
struct rb_root *root)
{
struct rb_node *parent = rb_parent(victim);
/* Copy the pointers/colour from the victim to the replacement */
*new = *victim;
/* Set the surrounding nodes to point to the replacement */
if (victim->rb_left)
rb_set_parent(victim->rb_left, new);
if (victim->rb_right)
rb_set_parent(victim->rb_right, new);
__rb_change_child(victim, new, parent, root);
}
EXPORT_SYMBOL(rb_replace_node);
void rb_replace_node_rcu(struct rb_node *victim, struct rb_node *new,
struct rb_root *root)
{
struct rb_node *parent = rb_parent(victim);
/* Copy the pointers/colour from the victim to the replacement */
*new = *victim;
/* Set the surrounding nodes to point to the replacement */
if (victim->rb_left)
rb_set_parent(victim->rb_left, new);
if (victim->rb_right)
rb_set_parent(victim->rb_right, new);
/* Set the parent's pointer to the new node last after an RCU barrier
* so that the pointers onwards are seen to be set correctly when doing
* an RCU walk over the tree.
*/
__rb_change_child_rcu(victim, new, parent, root);
}
EXPORT_SYMBOL(rb_replace_node_rcu);
rbtree: add postorder iteration functions Postorder iteration yields all of a node's children prior to yielding the node itself, and this particular implementation also avoids examining the leaf links in a node after that node has been yielded. In what I expect will be its most common usage, postorder iteration allows the deletion of every node in an rbtree without modifying the rbtree nodes (no _requirement_ that they be nulled) while avoiding referencing child nodes after they have been "deleted" (most commonly, freed). I have only updated zswap to use this functionality at this point, but numerous bits of code (most notably in the filesystem drivers) use a hand rolled postorder iteration that NULLs child links as it traverses the tree. Each of those instances could be replaced with this common implementation. 1 & 2 add rbtree postorder iteration functions. 3 adds testing of the iteration to the rbtree runtime tests 4 allows building the rbtree runtime tests as builtins 5 updates zswap. This patch: Add postorder iteration functions for rbtree. These are useful for safely freeing an entire rbtree without modifying the tree at all. Signed-off-by: Cody P Schafer <cody@linux.vnet.ibm.com> Reviewed-by: Seth Jennings <sjenning@linux.vnet.ibm.com> Cc: David Woodhouse <David.Woodhouse@intel.com> Cc: Rik van Riel <riel@redhat.com> Cc: Michel Lespinasse <walken@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years ago
static struct rb_node *rb_left_deepest_node(const struct rb_node *node)
{
for (;;) {
if (node->rb_left)
node = node->rb_left;
else if (node->rb_right)
node = node->rb_right;
else
return (struct rb_node *)node;
}
}
struct rb_node *rb_next_postorder(const struct rb_node *node)
{
const struct rb_node *parent;
if (!node)
return NULL;
parent = rb_parent(node);
/* If we're sitting on node, we've already seen our children */
if (parent && node == parent->rb_left && parent->rb_right) {
/* If we are the parent's left node, go to the parent's right
* node then all the way down to the left */
return rb_left_deepest_node(parent->rb_right);
} else
/* Otherwise we are the parent's right node, and the parent
* should be next */
return (struct rb_node *)parent;
}
EXPORT_SYMBOL(rb_next_postorder);
struct rb_node *rb_first_postorder(const struct rb_root *root)
{
if (!root->rb_node)
return NULL;
return rb_left_deepest_node(root->rb_node);
}
EXPORT_SYMBOL(rb_first_postorder);