linux_old1/include/linux/rbtree_augmented.h

293 lines
8.7 KiB
C
Raw Normal View History

/*
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/include/linux/rbtree_augmented.h
*/
#ifndef _LINUX_RBTREE_AUGMENTED_H
#define _LINUX_RBTREE_AUGMENTED_H
#include <linux/compiler.h>
#include <linux/rbtree.h>
#include <linux/rcupdate.h>
/*
* Please note - only struct rb_augment_callbacks and the prototypes for
* rb_insert_augmented() and rb_erase_augmented() are intended to be public.
* The rest are implementation details you are not expected to depend on.
*
* See Documentation/rbtree.txt for documentation and samples.
*/
struct rb_augment_callbacks {
void (*propagate)(struct rb_node *node, struct rb_node *stop);
void (*copy)(struct rb_node *old, struct rb_node *new);
void (*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>
2017-09-09 07:14:36 +08:00
extern void __rb_insert_augmented(struct rb_node *node,
struct rb_root *root,
bool newleft, struct rb_node **leftmost,
void (*augment_rotate)(struct rb_node *old, struct rb_node *new));
/*
* Fixup the rbtree and update the augmented information when rebalancing.
*
* On insertion, the user must update the augmented information on the path
* leading to the inserted node, then call rb_link_node() as usual and
* rb_augment_inserted() instead of the usual rb_insert_color() call.
* If rb_augment_inserted() rebalances the rbtree, it will callback into
* a user provided function to update the augmented information on the
* affected subtrees.
*/
static inline void
rb_insert_augmented(struct rb_node *node, struct rb_root *root,
const struct rb_augment_callbacks *augment)
{
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>
2017-09-09 07:14:36 +08:00
__rb_insert_augmented(node, root, false, NULL, augment->rotate);
}
static inline void
rb_insert_augmented_cached(struct rb_node *node,
struct rb_root_cached *root, bool newleft,
const struct rb_augment_callbacks *augment)
{
__rb_insert_augmented(node, &root->rb_root,
newleft, &root->rb_leftmost, augment->rotate);
}
#define RB_DECLARE_CALLBACKS(rbstatic, rbname, rbstruct, rbfield, \
rbtype, rbaugmented, rbcompute) \
static inline void \
rbname ## _propagate(struct rb_node *rb, struct rb_node *stop) \
{ \
while (rb != stop) { \
rbstruct *node = rb_entry(rb, rbstruct, rbfield); \
rbtype augmented = rbcompute(node); \
if (node->rbaugmented == augmented) \
break; \
node->rbaugmented = augmented; \
rb = rb_parent(&node->rbfield); \
} \
} \
static inline void \
rbname ## _copy(struct rb_node *rb_old, struct rb_node *rb_new) \
{ \
rbstruct *old = rb_entry(rb_old, rbstruct, rbfield); \
rbstruct *new = rb_entry(rb_new, rbstruct, rbfield); \
new->rbaugmented = old->rbaugmented; \
} \
static void \
rbname ## _rotate(struct rb_node *rb_old, struct rb_node *rb_new) \
{ \
rbstruct *old = rb_entry(rb_old, rbstruct, rbfield); \
rbstruct *new = rb_entry(rb_new, rbstruct, rbfield); \
new->rbaugmented = old->rbaugmented; \
old->rbaugmented = rbcompute(old); \
} \
rbstatic const struct rb_augment_callbacks rbname = { \
.propagate = rbname ## _propagate, \
.copy = rbname ## _copy, \
.rotate = rbname ## _rotate \
};
#define RB_RED 0
#define RB_BLACK 1
#define __rb_parent(pc) ((struct rb_node *)(pc & ~3))
#define __rb_color(pc) ((pc) & 1)
#define __rb_is_black(pc) __rb_color(pc)
#define __rb_is_red(pc) (!__rb_color(pc))
#define rb_color(rb) __rb_color((rb)->__rb_parent_color)
#define rb_is_red(rb) __rb_is_red((rb)->__rb_parent_color)
#define rb_is_black(rb) __rb_is_black((rb)->__rb_parent_color)
static inline void rb_set_parent(struct rb_node *rb, struct rb_node *p)
{
rb->__rb_parent_color = rb_color(rb) | (unsigned long)p;
}
static inline void rb_set_parent_color(struct rb_node *rb,
struct rb_node *p, int color)
{
rb->__rb_parent_color = (unsigned long)p | color;
}
static inline void
__rb_change_child(struct rb_node *old, struct rb_node *new,
struct rb_node *parent, struct rb_root *root)
{
if (parent) {
if (parent->rb_left == old)
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>
2015-05-27 09:39:36 +08:00
WRITE_ONCE(parent->rb_left, new);
else
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>
2015-05-27 09:39:36 +08:00
WRITE_ONCE(parent->rb_right, new);
} else
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>
2015-05-27 09:39:36 +08:00
WRITE_ONCE(root->rb_node, new);
}
static inline void
__rb_change_child_rcu(struct rb_node *old, struct rb_node *new,
struct rb_node *parent, struct rb_root *root)
{
if (parent) {
if (parent->rb_left == old)
rcu_assign_pointer(parent->rb_left, new);
else
rcu_assign_pointer(parent->rb_right, new);
} else
rcu_assign_pointer(root->rb_node, new);
}
extern void __rb_erase_color(struct rb_node *parent, struct rb_root *root,
void (*augment_rotate)(struct rb_node *old, struct rb_node *new));
static __always_inline struct rb_node *
__rb_erase_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>
2017-09-09 07:14:36 +08:00
struct rb_node **leftmost,
const struct rb_augment_callbacks *augment)
{
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>
2015-05-27 09:39:36 +08:00
struct rb_node *child = node->rb_right;
struct rb_node *tmp = node->rb_left;
struct rb_node *parent, *rebalance;
unsigned long pc;
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>
2017-09-09 07:14:36 +08:00
if (leftmost && node == *leftmost)
*leftmost = rb_next(node);
if (!tmp) {
/*
* Case 1: node to erase has no more than 1 child (easy!)
*
* Note that if there is one child it must be red due to 5)
* and node must be black due to 4). We adjust colors locally
* so as to bypass __rb_erase_color() later on.
*/
pc = node->__rb_parent_color;
parent = __rb_parent(pc);
__rb_change_child(node, child, parent, root);
if (child) {
child->__rb_parent_color = pc;
rebalance = NULL;
} else
rebalance = __rb_is_black(pc) ? parent : NULL;
tmp = parent;
} else if (!child) {
/* Still case 1, but this time the child is node->rb_left */
tmp->__rb_parent_color = pc = node->__rb_parent_color;
parent = __rb_parent(pc);
__rb_change_child(node, tmp, parent, root);
rebalance = NULL;
tmp = parent;
} else {
struct rb_node *successor = child, *child2;
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>
2015-05-27 09:39:36 +08:00
tmp = child->rb_left;
if (!tmp) {
/*
* Case 2: node's successor is its right child
*
* (n) (s)
* / \ / \
* (x) (s) -> (x) (c)
* \
* (c)
*/
parent = successor;
child2 = successor->rb_right;
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>
2015-05-27 09:39:36 +08:00
augment->copy(node, successor);
} else {
/*
* Case 3: node's successor is leftmost under
* node's right child subtree
*
* (n) (s)
* / \ / \
* (x) (y) -> (x) (y)
* / /
* (p) (p)
* / /
* (s) (c)
* \
* (c)
*/
do {
parent = successor;
successor = tmp;
tmp = tmp->rb_left;
} while (tmp);
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>
2015-05-27 09:39:36 +08:00
child2 = successor->rb_right;
WRITE_ONCE(parent->rb_left, child2);
WRITE_ONCE(successor->rb_right, child);
rb_set_parent(child, successor);
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>
2015-05-27 09:39:36 +08:00
augment->copy(node, successor);
augment->propagate(parent, successor);
}
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>
2015-05-27 09:39:36 +08:00
tmp = node->rb_left;
WRITE_ONCE(successor->rb_left, tmp);
rb_set_parent(tmp, successor);
pc = node->__rb_parent_color;
tmp = __rb_parent(pc);
__rb_change_child(node, successor, tmp, root);
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>
2015-05-27 09:39:36 +08:00
if (child2) {
successor->__rb_parent_color = pc;
rb_set_parent_color(child2, parent, RB_BLACK);
rebalance = NULL;
} else {
unsigned long pc2 = successor->__rb_parent_color;
successor->__rb_parent_color = pc;
rebalance = __rb_is_black(pc2) ? parent : NULL;
}
tmp = successor;
}
augment->propagate(tmp, NULL);
return rebalance;
}
static __always_inline void
rb_erase_augmented(struct rb_node *node, struct rb_root *root,
const struct rb_augment_callbacks *augment)
{
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>
2017-09-09 07:14:36 +08:00
struct rb_node *rebalance = __rb_erase_augmented(node, root,
NULL, augment);
if (rebalance)
__rb_erase_color(rebalance, root, augment->rotate);
}
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>
2017-09-09 07:14:36 +08:00
static __always_inline void
rb_erase_augmented_cached(struct rb_node *node, struct rb_root_cached *root,
const struct rb_augment_callbacks *augment)
{
struct rb_node *rebalance = __rb_erase_augmented(node, &root->rb_root,
&root->rb_leftmost,
augment);
if (rebalance)
__rb_erase_color(rebalance, &root->rb_root, augment->rotate);
}
#endif /* _LINUX_RBTREE_AUGMENTED_H */