* [PATCH 1/9] rbtree: Cache leftmost node internally
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
@ 2017-06-29 17:15 ` Davidlohr Bueso
2017-07-15 10:54 ` Christoph Hellwig
2017-06-29 17:15 ` [PATCH 2/9] sched/fair: Replace cfs_rq->rb_leftmost Davidlohr Bueso
` (9 subsequent siblings)
10 siblings, 1 reply; 15+ messages in thread
From: Davidlohr Bueso @ 2017-06-29 17:15 UTC (permalink / raw)
To: mingo, peterz, akpm
Cc: torvalds, jack, kirill.shutemov, ldufour, mhocko, mgorman, dave,
linux-kernel, Davidlohr Bueso
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.
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
include/linux/rbtree.h | 11 +++++++++++
include/linux/rbtree_augmented.h | 33 ++++++++++++++++++++++++++++++---
lib/rbtree.c | 34 +++++++++++++++++++++++++++++-----
3 files changed, 70 insertions(+), 8 deletions(-)
diff --git a/include/linux/rbtree.h b/include/linux/rbtree.h
index e585018498d5..71c9ee585454 100644
--- a/include/linux/rbtree.h
+++ b/include/linux/rbtree.h
@@ -44,10 +44,15 @@ struct rb_root {
struct rb_node *rb_node;
};
+struct rb_root_cached {
+ struct rb_root rb_root;
+ struct rb_node *rb_leftmost;
+};
#define rb_parent(r) ((struct rb_node *)((r)->__rb_parent_color & ~3))
#define RB_ROOT (struct rb_root) { NULL, }
+#define RB_ROOT_CACHED (struct rb_root_cached) { {NULL, }, NULL }
#define rb_entry(ptr, type, member) container_of(ptr, type, member)
#define RB_EMPTY_ROOT(root) (READ_ONCE((root)->rb_node) == NULL)
@@ -69,6 +74,12 @@ extern struct rb_node *rb_prev(const struct rb_node *);
extern struct rb_node *rb_first(const struct rb_root *);
extern struct rb_node *rb_last(const struct rb_root *);
+extern void rb_insert_color_cached(struct rb_node *,
+ struct rb_root_cached *, bool);
+extern void rb_erase_cached(struct rb_node *node, struct rb_root_cached *);
+/* Same as rb_first(), but O(1) */
+#define rb_first_cached(root) (root)->rb_leftmost
+
/* Postorder iteration - always visit the parent after its children */
extern struct rb_node *rb_first_postorder(const struct rb_root *);
extern struct rb_node *rb_next_postorder(const struct rb_node *);
diff --git a/include/linux/rbtree_augmented.h b/include/linux/rbtree_augmented.h
index 9702b6e183bc..6bfd2b581f75 100644
--- a/include/linux/rbtree_augmented.h
+++ b/include/linux/rbtree_augmented.h
@@ -41,7 +41,9 @@ struct rb_augment_callbacks {
void (*rotate)(struct rb_node *old, struct rb_node *new);
};
-extern void __rb_insert_augmented(struct rb_node *node, struct rb_root *root,
+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.
@@ -57,7 +59,16 @@ static inline void
rb_insert_augmented(struct rb_node *node, struct rb_root *root,
const struct rb_augment_callbacks *augment)
{
- __rb_insert_augmented(node, root, augment->rotate);
+ __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, \
@@ -150,6 +161,7 @@ extern void __rb_erase_color(struct rb_node *parent, struct rb_root *root,
static __always_inline struct rb_node *
__rb_erase_augmented(struct rb_node *node, struct rb_root *root,
+ struct rb_node **leftmost,
const struct rb_augment_callbacks *augment)
{
struct rb_node *child = node->rb_right;
@@ -157,6 +169,9 @@ __rb_erase_augmented(struct rb_node *node, struct rb_root *root,
struct rb_node *parent, *rebalance;
unsigned long pc;
+ if (leftmost && node == *leftmost)
+ *leftmost = rb_next(node);
+
if (!tmp) {
/*
* Case 1: node to erase has no more than 1 child (easy!)
@@ -256,9 +271,21 @@ static __always_inline void
rb_erase_augmented(struct rb_node *node, struct rb_root *root,
const struct rb_augment_callbacks *augment)
{
- struct rb_node *rebalance = __rb_erase_augmented(node, root, augment);
+ struct rb_node *rebalance = __rb_erase_augmented(node, root,
+ NULL, augment);
if (rebalance)
__rb_erase_color(rebalance, root, augment->rotate);
}
+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 */
diff --git a/lib/rbtree.c b/lib/rbtree.c
index 4ba2828a67c0..d102d9d2ffaa 100644
--- a/lib/rbtree.c
+++ b/lib/rbtree.c
@@ -95,10 +95,14 @@ __rb_rotate_set_parents(struct rb_node *old, struct rb_node *new,
static __always_inline void
__rb_insert(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))
{
struct rb_node *parent = rb_red_parent(node), *gparent, *tmp;
+ if (newleft)
+ *leftmost = node;
+
while (true) {
/*
* Loop invariant: node is red
@@ -434,19 +438,38 @@ static const struct rb_augment_callbacks dummy_callbacks = {
void rb_insert_color(struct rb_node *node, struct rb_root *root)
{
- __rb_insert(node, root, dummy_rotate);
+ __rb_insert(node, root, false, NULL, dummy_rotate);
}
EXPORT_SYMBOL(rb_insert_color);
void rb_erase(struct rb_node *node, struct rb_root *root)
{
struct rb_node *rebalance;
- rebalance = __rb_erase_augmented(node, root, &dummy_callbacks);
+ rebalance = __rb_erase_augmented(node, root,
+ NULL, &dummy_callbacks);
if (rebalance)
____rb_erase_color(rebalance, root, dummy_rotate);
}
EXPORT_SYMBOL(rb_erase);
+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);
+
/*
* Augmented rbtree manipulation functions.
*
@@ -455,9 +478,10 @@ EXPORT_SYMBOL(rb_erase);
*/
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))
{
- __rb_insert(node, root, augment_rotate);
+ __rb_insert(node, root, newleft, leftmost, augment_rotate);
}
EXPORT_SYMBOL(__rb_insert_augmented);
@@ -502,7 +526,7 @@ struct rb_node *rb_next(const struct rb_node *node)
* as we can.
*/
if (node->rb_right) {
- node = node->rb_right;
+ node = node->rb_right;
while (node->rb_left)
node=node->rb_left;
return (struct rb_node *)node;
@@ -534,7 +558,7 @@ struct rb_node *rb_prev(const struct rb_node *node)
* as we can.
*/
if (node->rb_left) {
- node = node->rb_left;
+ node = node->rb_left;
while (node->rb_right)
node=node->rb_right;
return (struct rb_node *)node;
--
2.12.0
^ permalink raw reply related [flat|nested] 15+ messages in thread* Re: [PATCH 1/9] rbtree: Cache leftmost node internally
2017-06-29 17:15 ` [PATCH 1/9] " Davidlohr Bueso
@ 2017-07-15 10:54 ` Christoph Hellwig
2017-07-15 15:28 ` Davidlohr Bueso
0 siblings, 1 reply; 15+ messages in thread
From: Christoph Hellwig @ 2017-07-15 10:54 UTC (permalink / raw)
To: Davidlohr Bueso
Cc: mingo, peterz, akpm, torvalds, jack, kirill.shutemov, ldufour,
mhocko, mgorman, linux-kernel, Davidlohr Bueso
Is the leftmost node so much more special than the rightmost one?
Just asking because I have some code that caches both of them, but
using your helpers would make it look very asymmetric..
^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: [PATCH 1/9] rbtree: Cache leftmost node internally
2017-07-15 10:54 ` Christoph Hellwig
@ 2017-07-15 15:28 ` Davidlohr Bueso
0 siblings, 0 replies; 15+ messages in thread
From: Davidlohr Bueso @ 2017-07-15 15:28 UTC (permalink / raw)
To: Christoph Hellwig
Cc: mingo, peterz, akpm, torvalds, jack, kirill.shutemov, ldufour,
mhocko, mgorman, linux-kernel, Davidlohr Bueso
On Sat, 15 Jul 2017, Christoph Hellwig wrote:
>Is the leftmost node so much more special than the rightmost one?
>
>Just asking because I have some code that caches both of them, but
>using your helpers would make it look very asymmetric..
Yeah, I thought of this, but ultimately decided that the leftmost
node was more used (rb_first vs rb_last calls) and probably not
worth adding the rightmost one which would endup enlarging the
structure even more.
That said, I do have some patches for the rbtree users to cache the
rightmost node, similarly to how it was done with the leftmost
before this patchset.
If folks don't mind enlarging rb_root_cached, I could definitely
add the rightmost, but it seems an overkill -- albeit asymmetric.
Thanks,
Davidlohr
^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH 2/9] sched/fair: Replace cfs_rq->rb_leftmost
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
2017-06-29 17:15 ` [PATCH 1/9] " Davidlohr Bueso
@ 2017-06-29 17:15 ` Davidlohr Bueso
2017-06-29 17:15 ` [PATCH 3/9] sched/deadline: Replace earliest dl and rq leftmost caching Davidlohr Bueso
` (8 subsequent siblings)
10 siblings, 0 replies; 15+ messages in thread
From: Davidlohr Bueso @ 2017-06-29 17:15 UTC (permalink / raw)
To: mingo, peterz, akpm
Cc: torvalds, jack, kirill.shutemov, ldufour, mhocko, mgorman, dave,
linux-kernel, Davidlohr Bueso
... with the generic rbtree flavor instead. No changes
in semantics whatsoever.
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
kernel/sched/debug.c | 2 +-
kernel/sched/fair.c | 35 +++++++++++------------------------
kernel/sched/sched.h | 3 +--
3 files changed, 13 insertions(+), 27 deletions(-)
diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 38f019324f1a..85df87666f6e 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -488,7 +488,7 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
SPLIT_NS(cfs_rq->exec_clock));
raw_spin_lock_irqsave(&rq->lock, flags);
- if (cfs_rq->rb_leftmost)
+ if (cfs_rq->tasks_timeline.rb_leftmost)
MIN_vruntime = (__pick_first_entity(cfs_rq))->vruntime;
last = __pick_last_entity(cfs_rq);
if (last)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 6f4f155adf5f..e59dbe5a8c30 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -523,8 +523,8 @@ static void update_min_vruntime(struct cfs_rq *cfs_rq)
curr = NULL;
}
- if (cfs_rq->rb_leftmost) {
- struct sched_entity *se = rb_entry(cfs_rq->rb_leftmost,
+ if (cfs_rq->tasks_timeline.rb_leftmost) {
+ struct sched_entity *se = rb_entry(cfs_rq->tasks_timeline.rb_leftmost,
struct sched_entity,
run_node);
@@ -547,10 +547,10 @@ static void update_min_vruntime(struct cfs_rq *cfs_rq)
*/
static void __enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se)
{
- struct rb_node **link = &cfs_rq->tasks_timeline.rb_node;
+ struct rb_node **link = &cfs_rq->tasks_timeline.rb_root.rb_node;
struct rb_node *parent = NULL;
struct sched_entity *entry;
- int leftmost = 1;
+ bool leftmost = true;
/*
* Find the right place in the rbtree:
@@ -566,36 +566,23 @@ static void __enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se)
link = &parent->rb_left;
} else {
link = &parent->rb_right;
- leftmost = 0;
+ leftmost = false;
}
}
- /*
- * Maintain a cache of leftmost tree entries (it is frequently
- * used):
- */
- if (leftmost)
- cfs_rq->rb_leftmost = &se->run_node;
-
rb_link_node(&se->run_node, parent, link);
- rb_insert_color(&se->run_node, &cfs_rq->tasks_timeline);
+ rb_insert_color_cached(&se->run_node,
+ &cfs_rq->tasks_timeline, leftmost);
}
static void __dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se)
{
- if (cfs_rq->rb_leftmost == &se->run_node) {
- struct rb_node *next_node;
-
- next_node = rb_next(&se->run_node);
- cfs_rq->rb_leftmost = next_node;
- }
-
- rb_erase(&se->run_node, &cfs_rq->tasks_timeline);
+ rb_erase_cached(&se->run_node, &cfs_rq->tasks_timeline);
}
struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq)
{
- struct rb_node *left = cfs_rq->rb_leftmost;
+ struct rb_node *left = cfs_rq->tasks_timeline.rb_leftmost;
if (!left)
return NULL;
@@ -616,7 +603,7 @@ static struct sched_entity *__pick_next_entity(struct sched_entity *se)
#ifdef CONFIG_SCHED_DEBUG
struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq)
{
- struct rb_node *last = rb_last(&cfs_rq->tasks_timeline);
+ struct rb_node *last = rb_last(&cfs_rq->tasks_timeline.rb_root);
if (!last)
return NULL;
@@ -9161,7 +9148,7 @@ static void set_curr_task_fair(struct rq *rq)
void init_cfs_rq(struct cfs_rq *cfs_rq)
{
- cfs_rq->tasks_timeline = RB_ROOT;
+ cfs_rq->tasks_timeline = RB_ROOT_CACHED;
cfs_rq->min_vruntime = (u64)(-(1LL << 20));
#ifndef CONFIG_64BIT
cfs_rq->min_vruntime_copy = cfs_rq->min_vruntime;
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index eeef1a3086d1..bf4d3f7d29c7 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -426,8 +426,7 @@ struct cfs_rq {
u64 min_vruntime_copy;
#endif
- struct rb_root tasks_timeline;
- struct rb_node *rb_leftmost;
+ struct rb_root_cached tasks_timeline;
/*
* 'curr' points to currently running entity on this cfs_rq.
--
2.12.0
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH 3/9] sched/deadline: Replace earliest dl and rq leftmost caching
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
2017-06-29 17:15 ` [PATCH 1/9] " Davidlohr Bueso
2017-06-29 17:15 ` [PATCH 2/9] sched/fair: Replace cfs_rq->rb_leftmost Davidlohr Bueso
@ 2017-06-29 17:15 ` Davidlohr Bueso
2017-06-29 17:15 ` [PATCH 4/9] locking/rtmutex: Replace top-waiter and pi_waiters " Davidlohr Bueso
` (7 subsequent siblings)
10 siblings, 0 replies; 15+ messages in thread
From: Davidlohr Bueso @ 2017-06-29 17:15 UTC (permalink / raw)
To: mingo, peterz, akpm
Cc: torvalds, jack, kirill.shutemov, ldufour, mhocko, mgorman, dave,
linux-kernel, Davidlohr Bueso
... with the generic rbtree flavor instead. No changes
in semantics whatsoever.
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
kernel/sched/deadline.c | 50 +++++++++++++++++++------------------------------
kernel/sched/sched.h | 6 ++----
2 files changed, 21 insertions(+), 35 deletions(-)
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index a84299f44b5d..5d2ab09e7b9f 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -296,7 +296,7 @@ static inline int is_leftmost(struct task_struct *p, struct dl_rq *dl_rq)
{
struct sched_dl_entity *dl_se = &p->dl;
- return dl_rq->rb_leftmost == &dl_se->rb_node;
+ return dl_rq->root.rb_leftmost == &dl_se->rb_node;
}
void init_dl_bandwidth(struct dl_bandwidth *dl_b, u64 period, u64 runtime)
@@ -320,7 +320,7 @@ void init_dl_bw(struct dl_bw *dl_b)
void init_dl_rq(struct dl_rq *dl_rq)
{
- dl_rq->rb_root = RB_ROOT;
+ dl_rq->root = RB_ROOT_CACHED;
#ifdef CONFIG_SMP
/* zero means no -deadline tasks */
@@ -328,7 +328,7 @@ void init_dl_rq(struct dl_rq *dl_rq)
dl_rq->dl_nr_migratory = 0;
dl_rq->overloaded = 0;
- dl_rq->pushable_dl_tasks_root = RB_ROOT;
+ dl_rq->pushable_dl_tasks_root = RB_ROOT_CACHED;
#else
init_dl_bw(&dl_rq->dl_bw);
#endif
@@ -410,10 +410,10 @@ static void dec_dl_migration(struct sched_dl_entity *dl_se, struct dl_rq *dl_rq)
static void enqueue_pushable_dl_task(struct rq *rq, struct task_struct *p)
{
struct dl_rq *dl_rq = &rq->dl;
- struct rb_node **link = &dl_rq->pushable_dl_tasks_root.rb_node;
+ struct rb_node **link = &dl_rq->pushable_dl_tasks_root.rb_root.rb_node;
struct rb_node *parent = NULL;
struct task_struct *entry;
- int leftmost = 1;
+ bool leftmost = true;
BUG_ON(!RB_EMPTY_NODE(&p->pushable_dl_tasks));
@@ -425,17 +425,16 @@ static void enqueue_pushable_dl_task(struct rq *rq, struct task_struct *p)
link = &parent->rb_left;
else {
link = &parent->rb_right;
- leftmost = 0;
+ leftmost = false;
}
}
- if (leftmost) {
- dl_rq->pushable_dl_tasks_leftmost = &p->pushable_dl_tasks;
+ if (leftmost)
dl_rq->earliest_dl.next = p->dl.deadline;
- }
rb_link_node(&p->pushable_dl_tasks, parent, link);
- rb_insert_color(&p->pushable_dl_tasks, &dl_rq->pushable_dl_tasks_root);
+ rb_insert_color_cached(&p->pushable_dl_tasks,
+ &dl_rq->pushable_dl_tasks_root, leftmost);
}
static void dequeue_pushable_dl_task(struct rq *rq, struct task_struct *p)
@@ -445,24 +444,23 @@ static void dequeue_pushable_dl_task(struct rq *rq, struct task_struct *p)
if (RB_EMPTY_NODE(&p->pushable_dl_tasks))
return;
- if (dl_rq->pushable_dl_tasks_leftmost == &p->pushable_dl_tasks) {
+ if (dl_rq->pushable_dl_tasks_root.rb_leftmost == &p->pushable_dl_tasks) {
struct rb_node *next_node;
next_node = rb_next(&p->pushable_dl_tasks);
- dl_rq->pushable_dl_tasks_leftmost = next_node;
if (next_node) {
dl_rq->earliest_dl.next = rb_entry(next_node,
struct task_struct, pushable_dl_tasks)->dl.deadline;
}
}
- rb_erase(&p->pushable_dl_tasks, &dl_rq->pushable_dl_tasks_root);
+ rb_erase_cached(&p->pushable_dl_tasks, &dl_rq->pushable_dl_tasks_root);
RB_CLEAR_NODE(&p->pushable_dl_tasks);
}
static inline int has_pushable_dl_tasks(struct rq *rq)
{
- return !RB_EMPTY_ROOT(&rq->dl.pushable_dl_tasks_root);
+ return !RB_EMPTY_ROOT(&rq->dl.pushable_dl_tasks_root.rb_root);
}
static int push_dl_task(struct rq *rq);
@@ -1266,7 +1264,7 @@ static void dec_dl_deadline(struct dl_rq *dl_rq, u64 deadline)
dl_rq->earliest_dl.next = 0;
cpudl_clear(&rq->rd->cpudl, rq->cpu);
} else {
- struct rb_node *leftmost = dl_rq->rb_leftmost;
+ struct rb_node *leftmost = dl_rq->root.rb_leftmost;
struct sched_dl_entity *entry;
entry = rb_entry(leftmost, struct sched_dl_entity, rb_node);
@@ -1313,7 +1311,7 @@ void dec_dl_tasks(struct sched_dl_entity *dl_se, struct dl_rq *dl_rq)
static void __enqueue_dl_entity(struct sched_dl_entity *dl_se)
{
struct dl_rq *dl_rq = dl_rq_of_se(dl_se);
- struct rb_node **link = &dl_rq->rb_root.rb_node;
+ struct rb_node **link = &dl_rq->root.rb_root.rb_node;
struct rb_node *parent = NULL;
struct sched_dl_entity *entry;
int leftmost = 1;
@@ -1331,11 +1329,8 @@ static void __enqueue_dl_entity(struct sched_dl_entity *dl_se)
}
}
- if (leftmost)
- dl_rq->rb_leftmost = &dl_se->rb_node;
-
rb_link_node(&dl_se->rb_node, parent, link);
- rb_insert_color(&dl_se->rb_node, &dl_rq->rb_root);
+ rb_insert_color_cached(&dl_se->rb_node, &dl_rq->root, leftmost);
inc_dl_tasks(dl_se, dl_rq);
}
@@ -1347,14 +1342,7 @@ static void __dequeue_dl_entity(struct sched_dl_entity *dl_se)
if (RB_EMPTY_NODE(&dl_se->rb_node))
return;
- if (dl_rq->rb_leftmost == &dl_se->rb_node) {
- struct rb_node *next_node;
-
- next_node = rb_next(&dl_se->rb_node);
- dl_rq->rb_leftmost = next_node;
- }
-
- rb_erase(&dl_se->rb_node, &dl_rq->rb_root);
+ rb_erase_cached(&dl_se->rb_node, &dl_rq->root);
RB_CLEAR_NODE(&dl_se->rb_node);
dec_dl_tasks(dl_se, dl_rq);
@@ -1645,7 +1633,7 @@ static void start_hrtick_dl(struct rq *rq, struct task_struct *p)
static struct sched_dl_entity *pick_next_dl_entity(struct rq *rq,
struct dl_rq *dl_rq)
{
- struct rb_node *left = dl_rq->rb_leftmost;
+ struct rb_node *left = rb_first_cached(&dl_rq->root);
if (!left)
return NULL;
@@ -1769,7 +1757,7 @@ static int pick_dl_task(struct rq *rq, struct task_struct *p, int cpu)
*/
static struct task_struct *pick_earliest_pushable_dl_task(struct rq *rq, int cpu)
{
- struct rb_node *next_node = rq->dl.pushable_dl_tasks_leftmost;
+ struct rb_node *next_node = rq->dl.pushable_dl_tasks_root.rb_leftmost;
struct task_struct *p = NULL;
if (!has_pushable_dl_tasks(rq))
@@ -1942,7 +1930,7 @@ static struct task_struct *pick_next_pushable_dl_task(struct rq *rq)
if (!has_pushable_dl_tasks(rq))
return NULL;
- p = rb_entry(rq->dl.pushable_dl_tasks_leftmost,
+ p = rb_entry(rq->dl.pushable_dl_tasks_root.rb_leftmost,
struct task_struct, pushable_dl_tasks);
BUG_ON(rq->cpu != task_cpu(p));
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index bf4d3f7d29c7..d34d1a0dd563 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -549,8 +549,7 @@ struct rt_rq {
/* Deadline class' related fields in a runqueue */
struct dl_rq {
/* runqueue is an rbtree, ordered by deadline */
- struct rb_root rb_root;
- struct rb_node *rb_leftmost;
+ struct rb_root_cached root;
unsigned long dl_nr_running;
@@ -574,8 +573,7 @@ struct dl_rq {
* an rb-tree, ordered by tasks' deadlines, with caching
* of the leftmost (earliest deadline) element.
*/
- struct rb_root pushable_dl_tasks_root;
- struct rb_node *pushable_dl_tasks_leftmost;
+ struct rb_root_cached pushable_dl_tasks_root;
#else
struct dl_bw dl_bw;
#endif
--
2.12.0
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH 4/9] locking/rtmutex: Replace top-waiter and pi_waiters leftmost caching
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
` (2 preceding siblings ...)
2017-06-29 17:15 ` [PATCH 3/9] sched/deadline: Replace earliest dl and rq leftmost caching Davidlohr Bueso
@ 2017-06-29 17:15 ` Davidlohr Bueso
2017-06-29 17:15 ` [PATCH 5/9] block/cfq: Replace cfq_rb_root " Davidlohr Bueso
` (6 subsequent siblings)
10 siblings, 0 replies; 15+ messages in thread
From: Davidlohr Bueso @ 2017-06-29 17:15 UTC (permalink / raw)
To: mingo, peterz, akpm
Cc: torvalds, jack, kirill.shutemov, ldufour, mhocko, mgorman, dave,
linux-kernel, Davidlohr Bueso
... with the generic rbtree flavor instead. No changes
in semantics whatsoever.
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
include/linux/init_task.h | 5 ++---
include/linux/rtmutex.h | 11 +++++------
include/linux/sched.h | 3 +--
kernel/fork.c | 3 +--
kernel/locking/rtmutex-debug.c | 2 +-
kernel/locking/rtmutex.c | 35 +++++++++++------------------------
kernel/locking/rtmutex_common.h | 12 ++++++------
7 files changed, 27 insertions(+), 44 deletions(-)
diff --git a/include/linux/init_task.h b/include/linux/init_task.h
index 9fa5aae21c00..d0f23d50eb20 100644
--- a/include/linux/init_task.h
+++ b/include/linux/init_task.h
@@ -175,9 +175,8 @@ extern struct cred init_cred;
#ifdef CONFIG_RT_MUTEXES
# define INIT_RT_MUTEXES(tsk) \
- .pi_waiters = RB_ROOT, \
- .pi_top_task = NULL, \
- .pi_waiters_leftmost = NULL,
+ .pi_waiters = RB_ROOT_CACHED, \
+ .pi_top_task = NULL,
#else
# define INIT_RT_MUTEXES(tsk)
#endif
diff --git a/include/linux/rtmutex.h b/include/linux/rtmutex.h
index 44fd002f7cd5..53fcbe9de7fd 100644
--- a/include/linux/rtmutex.h
+++ b/include/linux/rtmutex.h
@@ -22,18 +22,17 @@ extern int max_lock_depth; /* for sysctl */
* The rt_mutex structure
*
* @wait_lock: spinlock to protect the structure
- * @waiters: rbtree root to enqueue waiters in priority order
- * @waiters_leftmost: top waiter
+ * @waiters: rbtree root to enqueue waiters in priority order;
+ * caches top-waiter (leftmost node).
* @owner: the mutex owner
*/
struct rt_mutex {
raw_spinlock_t wait_lock;
- struct rb_root waiters;
- struct rb_node *waiters_leftmost;
+ struct rb_root_cached waiters;
struct task_struct *owner;
#ifdef CONFIG_DEBUG_RT_MUTEXES
int save_state;
- const char *name, *file;
+ const char *name, *file;
int line;
void *magic;
#endif
@@ -84,7 +83,7 @@ do { \
#define __RT_MUTEX_INITIALIZER(mutexname) \
{ .wait_lock = __RAW_SPIN_LOCK_UNLOCKED(mutexname.wait_lock) \
- , .waiters = RB_ROOT \
+ , .waiters = RB_ROOT_CACHED \
, .owner = NULL \
__DEBUG_RT_MUTEX_INITIALIZER(mutexname) \
__DEP_MAP_RT_MUTEX_INITIALIZER(mutexname)}
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 4e933f368cc0..5cb5d2e31c02 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -803,8 +803,7 @@ struct task_struct {
#ifdef CONFIG_RT_MUTEXES
/* PI waiters blocked on a rt_mutex held by this task: */
- struct rb_root pi_waiters;
- struct rb_node *pi_waiters_leftmost;
+ struct rb_root_cached pi_waiters;
/* Updated under owner's pi_lock and rq lock */
struct task_struct *pi_top_task;
/* Deadlock detection and priority inheritance handling: */
diff --git a/kernel/fork.c b/kernel/fork.c
index 05a4984fc044..8dcc63e8d9e6 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1458,8 +1458,7 @@ static void rt_mutex_init_task(struct task_struct *p)
{
raw_spin_lock_init(&p->pi_lock);
#ifdef CONFIG_RT_MUTEXES
- p->pi_waiters = RB_ROOT;
- p->pi_waiters_leftmost = NULL;
+ p->pi_waiters = RB_ROOT_CACHED;
p->pi_top_task = NULL;
p->pi_blocked_on = NULL;
#endif
diff --git a/kernel/locking/rtmutex-debug.c b/kernel/locking/rtmutex-debug.c
index ac35e648b0e5..f4a74e78d467 100644
--- a/kernel/locking/rtmutex-debug.c
+++ b/kernel/locking/rtmutex-debug.c
@@ -58,7 +58,7 @@ static void printk_lock(struct rt_mutex *lock, int print_owner)
void rt_mutex_debug_task_free(struct task_struct *task)
{
- DEBUG_LOCKS_WARN_ON(!RB_EMPTY_ROOT(&task->pi_waiters));
+ DEBUG_LOCKS_WARN_ON(!RB_EMPTY_ROOT(&task->pi_waiters.rb_root));
DEBUG_LOCKS_WARN_ON(task->pi_blocked_on);
}
diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c
index 78069895032a..e40eee33174f 100644
--- a/kernel/locking/rtmutex.c
+++ b/kernel/locking/rtmutex.c
@@ -271,10 +271,10 @@ rt_mutex_waiter_equal(struct rt_mutex_waiter *left,
static void
rt_mutex_enqueue(struct rt_mutex *lock, struct rt_mutex_waiter *waiter)
{
- struct rb_node **link = &lock->waiters.rb_node;
+ struct rb_node **link = &lock->waiters.rb_root.rb_node;
struct rb_node *parent = NULL;
struct rt_mutex_waiter *entry;
- int leftmost = 1;
+ bool leftmost = true;
while (*link) {
parent = *link;
@@ -283,15 +283,12 @@ rt_mutex_enqueue(struct rt_mutex *lock, struct rt_mutex_waiter *waiter)
link = &parent->rb_left;
} else {
link = &parent->rb_right;
- leftmost = 0;
+ leftmost = false;
}
}
- if (leftmost)
- lock->waiters_leftmost = &waiter->tree_entry;
-
rb_link_node(&waiter->tree_entry, parent, link);
- rb_insert_color(&waiter->tree_entry, &lock->waiters);
+ rb_insert_color_cached(&waiter->tree_entry, &lock->waiters, leftmost);
}
static void
@@ -300,20 +297,17 @@ rt_mutex_dequeue(struct rt_mutex *lock, struct rt_mutex_waiter *waiter)
if (RB_EMPTY_NODE(&waiter->tree_entry))
return;
- if (lock->waiters_leftmost == &waiter->tree_entry)
- lock->waiters_leftmost = rb_next(&waiter->tree_entry);
-
- rb_erase(&waiter->tree_entry, &lock->waiters);
+ rb_erase_cached(&waiter->tree_entry, &lock->waiters);
RB_CLEAR_NODE(&waiter->tree_entry);
}
static void
rt_mutex_enqueue_pi(struct task_struct *task, struct rt_mutex_waiter *waiter)
{
- struct rb_node **link = &task->pi_waiters.rb_node;
+ struct rb_node **link = &task->pi_waiters.rb_root.rb_node;
struct rb_node *parent = NULL;
struct rt_mutex_waiter *entry;
- int leftmost = 1;
+ bool leftmost = true;
while (*link) {
parent = *link;
@@ -322,15 +316,12 @@ rt_mutex_enqueue_pi(struct task_struct *task, struct rt_mutex_waiter *waiter)
link = &parent->rb_left;
} else {
link = &parent->rb_right;
- leftmost = 0;
+ leftmost = false;
}
}
- if (leftmost)
- task->pi_waiters_leftmost = &waiter->pi_tree_entry;
-
rb_link_node(&waiter->pi_tree_entry, parent, link);
- rb_insert_color(&waiter->pi_tree_entry, &task->pi_waiters);
+ rb_insert_color_cached(&waiter->pi_tree_entry, &task->pi_waiters, leftmost);
}
static void
@@ -339,10 +330,7 @@ rt_mutex_dequeue_pi(struct task_struct *task, struct rt_mutex_waiter *waiter)
if (RB_EMPTY_NODE(&waiter->pi_tree_entry))
return;
- if (task->pi_waiters_leftmost == &waiter->pi_tree_entry)
- task->pi_waiters_leftmost = rb_next(&waiter->pi_tree_entry);
-
- rb_erase(&waiter->pi_tree_entry, &task->pi_waiters);
+ rb_erase_cached(&waiter->pi_tree_entry, &task->pi_waiters);
RB_CLEAR_NODE(&waiter->pi_tree_entry);
}
@@ -1658,8 +1646,7 @@ void __rt_mutex_init(struct rt_mutex *lock, const char *name,
{
lock->owner = NULL;
raw_spin_lock_init(&lock->wait_lock);
- lock->waiters = RB_ROOT;
- lock->waiters_leftmost = NULL;
+ lock->waiters = RB_ROOT_CACHED;
if (name && key)
debug_rt_mutex_init(lock, name, key);
diff --git a/kernel/locking/rtmutex_common.h b/kernel/locking/rtmutex_common.h
index 72ad45a9a794..524beeee24b0 100644
--- a/kernel/locking/rtmutex_common.h
+++ b/kernel/locking/rtmutex_common.h
@@ -42,7 +42,7 @@ struct rt_mutex_waiter {
*/
static inline int rt_mutex_has_waiters(struct rt_mutex *lock)
{
- return !RB_EMPTY_ROOT(&lock->waiters);
+ return !RB_EMPTY_ROOT(&lock->waiters.rb_root);
}
static inline struct rt_mutex_waiter *
@@ -50,8 +50,8 @@ rt_mutex_top_waiter(struct rt_mutex *lock)
{
struct rt_mutex_waiter *w;
- w = rb_entry(lock->waiters_leftmost, struct rt_mutex_waiter,
- tree_entry);
+ w = rb_entry(lock->waiters.rb_leftmost,
+ struct rt_mutex_waiter, tree_entry);
BUG_ON(w->lock != lock);
return w;
@@ -59,14 +59,14 @@ rt_mutex_top_waiter(struct rt_mutex *lock)
static inline int task_has_pi_waiters(struct task_struct *p)
{
- return !RB_EMPTY_ROOT(&p->pi_waiters);
+ return !RB_EMPTY_ROOT(&p->pi_waiters.rb_root);
}
static inline struct rt_mutex_waiter *
task_top_pi_waiter(struct task_struct *p)
{
- return rb_entry(p->pi_waiters_leftmost, struct rt_mutex_waiter,
- pi_tree_entry);
+ return rb_entry(p->pi_waiters.rb_leftmost,
+ struct rt_mutex_waiter, pi_tree_entry);
}
/*
--
2.12.0
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH 5/9] block/cfq: Replace cfq_rb_root leftmost caching
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
` (3 preceding siblings ...)
2017-06-29 17:15 ` [PATCH 4/9] locking/rtmutex: Replace top-waiter and pi_waiters " Davidlohr Bueso
@ 2017-06-29 17:15 ` Davidlohr Bueso
2017-06-29 17:15 ` [PATCH 6/9] lib/interval_tree: Fast overlap detection Davidlohr Bueso
` (5 subsequent siblings)
10 siblings, 0 replies; 15+ messages in thread
From: Davidlohr Bueso @ 2017-06-29 17:15 UTC (permalink / raw)
To: mingo, peterz, akpm
Cc: torvalds, jack, kirill.shutemov, ldufour, mhocko, mgorman, dave,
linux-kernel, axboe, linux-block, Davidlohr Bueso
... with the generic rbtree flavor instead. No changes
in semantics whatsoever.
Cc: axboe@fb.com
Cc: linux-block@vger.kernel.org
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
This is part of the rbtree internal caching series:
https://lkml.org/lkml/2017/6/8/857
block/cfq-iosched.c | 70 +++++++++++++++--------------------------------------
1 file changed, 20 insertions(+), 50 deletions(-)
diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index 3d5c28945719..92c31683a2bb 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -93,13 +93,12 @@ struct cfq_ttime {
* move this into the elevator for the rq sorting as well.
*/
struct cfq_rb_root {
- struct rb_root rb;
- struct rb_node *left;
+ struct rb_root_cached rb;
unsigned count;
u64 min_vdisktime;
struct cfq_ttime ttime;
};
-#define CFQ_RB_ROOT (struct cfq_rb_root) { .rb = RB_ROOT, \
+#define CFQ_RB_ROOT (struct cfq_rb_root) { .rb = RB_ROOT_CACHED, \
.ttime = {.last_end_request = ktime_get_ns(),},}
/*
@@ -984,10 +983,9 @@ static inline u64 max_vdisktime(u64 min_vdisktime, u64 vdisktime)
static void update_min_vdisktime(struct cfq_rb_root *st)
{
- struct cfq_group *cfqg;
+ if (!RB_EMPTY_ROOT(&st->rb.rb_root)) {
+ struct cfq_group *cfqg = rb_entry_cfqg(st->rb.rb_leftmost);
- if (st->left) {
- cfqg = rb_entry_cfqg(st->left);
st->min_vdisktime = max_vdisktime(st->min_vdisktime,
cfqg->vdisktime);
}
@@ -1169,46 +1167,25 @@ cfq_choose_req(struct cfq_data *cfqd, struct request *rq1, struct request *rq2,
}
}
-/*
- * The below is leftmost cache rbtree addon
- */
static struct cfq_queue *cfq_rb_first(struct cfq_rb_root *root)
{
/* Service tree is empty */
if (!root->count)
return NULL;
- if (!root->left)
- root->left = rb_first(&root->rb);
-
- if (root->left)
- return rb_entry(root->left, struct cfq_queue, rb_node);
-
- return NULL;
+ return rb_entry(rb_first_cached(&root->rb), struct cfq_queue, rb_node);
}
static struct cfq_group *cfq_rb_first_group(struct cfq_rb_root *root)
{
- if (!root->left)
- root->left = rb_first(&root->rb);
-
- if (root->left)
- return rb_entry_cfqg(root->left);
-
- return NULL;
+ return rb_entry_cfqg(rb_first_cached(&root->rb));
}
-static void rb_erase_init(struct rb_node *n, struct rb_root *root)
+static void cfq_rb_erase(struct rb_node *n, struct cfq_rb_root *root)
{
- rb_erase(n, root);
+ rb_erase_cached(n, &root->rb);
RB_CLEAR_NODE(n);
-}
-static void cfq_rb_erase(struct rb_node *n, struct cfq_rb_root *root)
-{
- if (root->left == n)
- root->left = NULL;
- rb_erase_init(n, &root->rb);
--root->count;
}
@@ -1258,11 +1235,11 @@ cfqg_key(struct cfq_rb_root *st, struct cfq_group *cfqg)
static void
__cfq_group_service_tree_add(struct cfq_rb_root *st, struct cfq_group *cfqg)
{
- struct rb_node **node = &st->rb.rb_node;
+ struct rb_node **node = &st->rb.rb_root.rb_node;
struct rb_node *parent = NULL;
struct cfq_group *__cfqg;
s64 key = cfqg_key(st, cfqg);
- int left = 1;
+ bool leftmost = true;
while (*node != NULL) {
parent = *node;
@@ -1272,15 +1249,12 @@ __cfq_group_service_tree_add(struct cfq_rb_root *st, struct cfq_group *cfqg)
node = &parent->rb_left;
else {
node = &parent->rb_right;
- left = 0;
+ leftmost = false;
}
}
- if (left)
- st->left = &cfqg->rb_node;
-
rb_link_node(&cfqg->rb_node, parent, node);
- rb_insert_color(&cfqg->rb_node, &st->rb);
+ rb_insert_color_cached(&cfqg->rb_node, &st->rb, leftmost);
}
/*
@@ -1381,7 +1355,7 @@ cfq_group_notify_queue_add(struct cfq_data *cfqd, struct cfq_group *cfqg)
* so that groups get lesser vtime based on their weights, so that
* if group does not loose all if it was not continuously backlogged.
*/
- n = rb_last(&st->rb);
+ n = rb_last(&st->rb.rb_root);
if (n) {
__cfqg = rb_entry_cfqg(n);
cfqg->vdisktime = __cfqg->vdisktime +
@@ -2223,14 +2197,14 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq,
struct cfq_queue *__cfqq;
u64 rb_key;
struct cfq_rb_root *st;
- int left;
+ bool leftmost = true;
int new_cfqq = 1;
u64 now = ktime_get_ns();
st = st_for(cfqq->cfqg, cfqq_class(cfqq), cfqq_type(cfqq));
if (cfq_class_idle(cfqq)) {
rb_key = CFQ_IDLE_DELAY;
- parent = rb_last(&st->rb);
+ parent = rb_last(&st->rb.rb_root);
if (parent && parent != &cfqq->rb_node) {
__cfqq = rb_entry(parent, struct cfq_queue, rb_node);
rb_key += __cfqq->rb_key;
@@ -2264,10 +2238,9 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq,
cfqq->service_tree = NULL;
}
- left = 1;
parent = NULL;
cfqq->service_tree = st;
- p = &st->rb.rb_node;
+ p = &st->rb.rb_root.rb_node;
while (*p) {
parent = *p;
__cfqq = rb_entry(parent, struct cfq_queue, rb_node);
@@ -2279,16 +2252,13 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq,
p = &parent->rb_left;
else {
p = &parent->rb_right;
- left = 0;
+ leftmost = false;
}
}
- if (left)
- st->left = &cfqq->rb_node;
-
cfqq->rb_key = rb_key;
rb_link_node(&cfqq->rb_node, parent, p);
- rb_insert_color(&cfqq->rb_node, &st->rb);
+ rb_insert_color_cached(&cfqq->rb_node, &st->rb, leftmost);
st->count++;
if (add_front || !new_cfqq)
return;
@@ -2735,7 +2705,7 @@ static struct cfq_queue *cfq_get_next_queue(struct cfq_data *cfqd)
/* There is nothing to dispatch */
if (!st)
return NULL;
- if (RB_EMPTY_ROOT(&st->rb))
+ if (RB_EMPTY_ROOT(&st->rb.rb_root))
return NULL;
return cfq_rb_first(st);
}
@@ -3221,7 +3191,7 @@ static struct cfq_group *cfq_get_next_cfqg(struct cfq_data *cfqd)
struct cfq_rb_root *st = &cfqd->grp_service_tree;
struct cfq_group *cfqg;
- if (RB_EMPTY_ROOT(&st->rb))
+ if (RB_EMPTY_ROOT(&st->rb.rb_root))
return NULL;
cfqg = cfq_rb_first_group(st);
update_min_vdisktime(st);
--
2.12.0
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH 6/9] lib/interval_tree: Fast overlap detection
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
` (4 preceding siblings ...)
2017-06-29 17:15 ` [PATCH 5/9] block/cfq: Replace cfq_rb_root " Davidlohr Bueso
@ 2017-06-29 17:15 ` Davidlohr Bueso
2017-06-29 17:15 ` [PATCH 7/9] lib/interval-tree: Correct comment wrt generic flavor Davidlohr Bueso
` (4 subsequent siblings)
10 siblings, 0 replies; 15+ messages in thread
From: Davidlohr Bueso @ 2017-06-29 17:15 UTC (permalink / raw)
To: mingo, peterz, akpm
Cc: torvalds, jack, kirill.shutemov, ldufour, mhocko, mgorman, dave,
linux-kernel, David Airlie, dri-devel, Michael S. Tsirkin,
Jason Wang, Doug Ledford, Christian Benvenuti, linux-rdma,
Davidlohr Bueso
Allow interval trees to quickly check for overlaps to avoid
unnecesary tree lookups in interval_tree_iter_first().
As of this patch, all interval tree flavors will require
using a 'rb_root_cached' such that we can have the leftmost
node easily available. While most users will make use of this
feature, those with special functions (in addition to the generic
insert, delete, search calls) will avoid using the cached
option as they can do funky things with insertions -- for example,
vma_interval_tree_insert_after().
Cc: David Airlie <airlied@linux.ie>
Cc: dri-devel@lists.freedesktop.org
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Doug Ledford <dledford@redhat.com>
Cc: Christian Benvenuti <benve@cisco.com>
Cc: linux-rdma@vger.kernel.org
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
This is part of the rbtree internal caching series:
https://lkml.org/lkml/2017/6/8/857
drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c | 8 ++--
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 7 ++--
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 2 +-
drivers/gpu/drm/drm_mm.c | 19 +++++----
drivers/gpu/drm/drm_vma_manager.c | 2 +-
drivers/gpu/drm/i915/i915_gem_userptr.c | 6 +--
drivers/gpu/drm/radeon/radeon.h | 2 +-
drivers/gpu/drm/radeon/radeon_mn.c | 8 ++--
drivers/gpu/drm/radeon/radeon_vm.c | 7 ++--
drivers/infiniband/core/umem_rbtree.c | 4 +-
drivers/infiniband/core/uverbs_cmd.c | 2 +-
drivers/infiniband/hw/hfi1/mmu_rb.c | 10 ++---
drivers/infiniband/hw/usnic/usnic_uiom.c | 6 +--
drivers/infiniband/hw/usnic/usnic_uiom.h | 2 +-
.../infiniband/hw/usnic/usnic_uiom_interval_tree.c | 15 +++----
.../infiniband/hw/usnic/usnic_uiom_interval_tree.h | 12 +++---
drivers/vhost/vhost.c | 2 +-
drivers/vhost/vhost.h | 2 +-
fs/hugetlbfs/inode.c | 6 +--
fs/inode.c | 2 +-
include/drm/drm_mm.h | 2 +-
include/linux/fs.h | 4 +-
include/linux/interval_tree.h | 8 ++--
include/linux/interval_tree_generic.h | 46 +++++++++++++++++-----
include/linux/mm.h | 17 ++++----
include/linux/rmap.h | 4 +-
include/rdma/ib_umem_odp.h | 11 ++++--
include/rdma/ib_verbs.h | 2 +-
lib/interval_tree_test.c | 4 +-
mm/interval_tree.c | 10 ++---
mm/memory.c | 4 +-
mm/mmap.c | 10 ++---
mm/rmap.c | 4 +-
33 files changed, 145 insertions(+), 105 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c
index 38f739fb727b..3f8aef21b9a6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c
@@ -51,7 +51,7 @@ struct amdgpu_mn {
/* objects protected by lock */
struct mutex lock;
- struct rb_root objects;
+ struct rb_root_cached objects;
};
struct amdgpu_mn_node {
@@ -76,8 +76,8 @@ static void amdgpu_mn_destroy(struct work_struct *work)
mutex_lock(&adev->mn_lock);
mutex_lock(&rmn->lock);
hash_del(&rmn->node);
- rbtree_postorder_for_each_entry_safe(node, next_node, &rmn->objects,
- it.rb) {
+ rbtree_postorder_for_each_entry_safe(node, next_node,
+ &rmn->objects.rb_root, it.rb) {
list_for_each_entry_safe(bo, next_bo, &node->bos, mn_list) {
bo->mn = NULL;
list_del_init(&bo->mn_list);
@@ -252,7 +252,7 @@ static struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev)
rmn->mm = mm;
rmn->mn.ops = &amdgpu_mn_ops;
mutex_init(&rmn->lock);
- rmn->objects = RB_ROOT;
+ rmn->objects = RB_ROOT_CACHED;
r = __mmu_notifier_register(&rmn->mn, mm);
if (r)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 5795f81369f0..f872e2179bbd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -2405,7 +2405,7 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
int r, i;
u64 flags;
- vm->va = RB_ROOT;
+ vm->va = RB_ROOT_CACHED;
vm->client_id = atomic64_inc_return(&adev->vm_manager.client_counter);
for (i = 0; i < AMDGPU_MAX_VMHUBS; i++)
vm->reserved_vmid[i] = NULL;
@@ -2512,10 +2512,11 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
amd_sched_entity_fini(vm->entity.sched, &vm->entity);
- if (!RB_EMPTY_ROOT(&vm->va)) {
+ if (!RB_EMPTY_ROOT(&vm->va.rb_root)) {
dev_err(adev->dev, "still active bo inside vm\n");
}
- rbtree_postorder_for_each_entry_safe(mapping, tmp, &vm->va, rb) {
+ rbtree_postorder_for_each_entry_safe(mapping, tmp,
+ &vm->va.rb_root, rb) {
list_del(&mapping->list);
amdgpu_vm_it_remove(mapping, &vm->va);
kfree(mapping);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 936f158bc5ec..ebffc1253f85 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -106,7 +106,7 @@ struct amdgpu_vm_pt {
struct amdgpu_vm {
/* tree of virtual addresses mapped */
- struct rb_root va;
+ struct rb_root_cached va;
/* protecting invalidated */
spinlock_t status_lock;
diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c
index f794089d30ac..61a1c8ea74bc 100644
--- a/drivers/gpu/drm/drm_mm.c
+++ b/drivers/gpu/drm/drm_mm.c
@@ -169,7 +169,7 @@ INTERVAL_TREE_DEFINE(struct drm_mm_node, rb,
struct drm_mm_node *
__drm_mm_interval_first(const struct drm_mm *mm, u64 start, u64 last)
{
- return drm_mm_interval_tree_iter_first((struct rb_root *)&mm->interval_tree,
+ return drm_mm_interval_tree_iter_first((struct rb_root_cached *)&mm->interval_tree,
start, last) ?: (struct drm_mm_node *)&mm->head_node;
}
EXPORT_SYMBOL(__drm_mm_interval_first);
@@ -180,6 +180,7 @@ static void drm_mm_interval_tree_add_node(struct drm_mm_node *hole_node,
struct drm_mm *mm = hole_node->mm;
struct rb_node **link, *rb;
struct drm_mm_node *parent;
+ bool leftmost = true;
node->__subtree_last = LAST(node);
@@ -196,9 +197,10 @@ static void drm_mm_interval_tree_add_node(struct drm_mm_node *hole_node,
rb = &hole_node->rb;
link = &hole_node->rb.rb_right;
+ leftmost = false;
} else {
rb = NULL;
- link = &mm->interval_tree.rb_node;
+ link = &mm->interval_tree.rb_root.rb_node;
}
while (*link) {
@@ -208,14 +210,15 @@ static void drm_mm_interval_tree_add_node(struct drm_mm_node *hole_node,
parent->__subtree_last = node->__subtree_last;
if (node->start < parent->start)
link = &parent->rb.rb_left;
- else
+ else {
link = &parent->rb.rb_right;
+ leftmost = true;
+ }
}
rb_link_node(&node->rb, rb, link);
- rb_insert_augmented(&node->rb,
- &mm->interval_tree,
- &drm_mm_interval_tree_augment);
+ rb_insert_augmented_cached(&node->rb, &mm->interval_tree, leftmost,
+ &drm_mm_interval_tree_augment);
}
#define RB_INSERT(root, member, expr) do { \
@@ -577,7 +580,7 @@ void drm_mm_replace_node(struct drm_mm_node *old, struct drm_mm_node *new)
*new = *old;
list_replace(&old->node_list, &new->node_list);
- rb_replace_node(&old->rb, &new->rb, &old->mm->interval_tree);
+ rb_replace_node(&old->rb, &new->rb, &old->mm->interval_tree.rb_root);
if (drm_mm_hole_follows(old)) {
list_replace(&old->hole_stack, &new->hole_stack);
@@ -863,7 +866,7 @@ void drm_mm_init(struct drm_mm *mm, u64 start, u64 size)
mm->color_adjust = NULL;
INIT_LIST_HEAD(&mm->hole_stack);
- mm->interval_tree = RB_ROOT;
+ mm->interval_tree = RB_ROOT_CACHED;
mm->holes_size = RB_ROOT;
mm->holes_addr = RB_ROOT;
diff --git a/drivers/gpu/drm/drm_vma_manager.c b/drivers/gpu/drm/drm_vma_manager.c
index d9100b565198..28f1226576f8 100644
--- a/drivers/gpu/drm/drm_vma_manager.c
+++ b/drivers/gpu/drm/drm_vma_manager.c
@@ -147,7 +147,7 @@ struct drm_vma_offset_node *drm_vma_offset_lookup_locked(struct drm_vma_offset_m
struct rb_node *iter;
unsigned long offset;
- iter = mgr->vm_addr_space_mm.interval_tree.rb_node;
+ iter = mgr->vm_addr_space_mm.interval_tree.rb_root.rb_node;
best = NULL;
while (likely(iter)) {
diff --git a/drivers/gpu/drm/i915/i915_gem_userptr.c b/drivers/gpu/drm/i915/i915_gem_userptr.c
index ccd09e8419f5..71dddf66baaa 100644
--- a/drivers/gpu/drm/i915/i915_gem_userptr.c
+++ b/drivers/gpu/drm/i915/i915_gem_userptr.c
@@ -49,7 +49,7 @@ struct i915_mmu_notifier {
spinlock_t lock;
struct hlist_node node;
struct mmu_notifier mn;
- struct rb_root objects;
+ struct rb_root_cached objects;
struct workqueue_struct *wq;
};
@@ -123,7 +123,7 @@ static void i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn,
struct interval_tree_node *it;
LIST_HEAD(cancelled);
- if (RB_EMPTY_ROOT(&mn->objects))
+ if (RB_EMPTY_ROOT(&mn->objects.rb_root))
return;
/* interval ranges are inclusive, but invalidate range is exclusive */
@@ -172,7 +172,7 @@ i915_mmu_notifier_create(struct mm_struct *mm)
spin_lock_init(&mn->lock);
mn->mn.ops = &i915_gem_userptr_notifier;
- mn->objects = RB_ROOT;
+ mn->objects = RB_ROOT_CACHED;
mn->wq = alloc_workqueue("i915-userptr-release", WQ_UNBOUND, 0);
if (mn->wq == NULL) {
kfree(mn);
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 5008f3d4cccc..10d0dd146808 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -924,7 +924,7 @@ struct radeon_vm_id {
struct radeon_vm {
struct mutex mutex;
- struct rb_root va;
+ struct rb_root_cached va;
/* protecting invalidated and freed */
spinlock_t status_lock;
diff --git a/drivers/gpu/drm/radeon/radeon_mn.c b/drivers/gpu/drm/radeon/radeon_mn.c
index 896f2cf51e4e..1d62288b7ee3 100644
--- a/drivers/gpu/drm/radeon/radeon_mn.c
+++ b/drivers/gpu/drm/radeon/radeon_mn.c
@@ -50,7 +50,7 @@ struct radeon_mn {
/* objects protected by lock */
struct mutex lock;
- struct rb_root objects;
+ struct rb_root_cached objects;
};
struct radeon_mn_node {
@@ -75,8 +75,8 @@ static void radeon_mn_destroy(struct work_struct *work)
mutex_lock(&rdev->mn_lock);
mutex_lock(&rmn->lock);
hash_del(&rmn->node);
- rbtree_postorder_for_each_entry_safe(node, next_node, &rmn->objects,
- it.rb) {
+ rbtree_postorder_for_each_entry_safe(node, next_node,
+ &rmn->objects.rb_root, it.rb) {
interval_tree_remove(&node->it, &rmn->objects);
list_for_each_entry_safe(bo, next_bo, &node->bos, mn_list) {
@@ -205,7 +205,7 @@ static struct radeon_mn *radeon_mn_get(struct radeon_device *rdev)
rmn->mm = mm;
rmn->mn.ops = &radeon_mn_ops;
mutex_init(&rmn->lock);
- rmn->objects = RB_ROOT;
+ rmn->objects = RB_ROOT_CACHED;
r = __mmu_notifier_register(&rmn->mn, mm);
if (r)
diff --git a/drivers/gpu/drm/radeon/radeon_vm.c b/drivers/gpu/drm/radeon/radeon_vm.c
index 5f68245579a3..f44777a6c2e8 100644
--- a/drivers/gpu/drm/radeon/radeon_vm.c
+++ b/drivers/gpu/drm/radeon/radeon_vm.c
@@ -1185,7 +1185,7 @@ int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
vm->ids[i].last_id_use = NULL;
}
mutex_init(&vm->mutex);
- vm->va = RB_ROOT;
+ vm->va = RB_ROOT_CACHED;
spin_lock_init(&vm->status_lock);
INIT_LIST_HEAD(&vm->invalidated);
INIT_LIST_HEAD(&vm->freed);
@@ -1232,10 +1232,11 @@ void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
struct radeon_bo_va *bo_va, *tmp;
int i, r;
- if (!RB_EMPTY_ROOT(&vm->va)) {
+ if (!RB_EMPTY_ROOT(&vm->va.rb_root)) {
dev_err(rdev->dev, "still active bo inside vm\n");
}
- rbtree_postorder_for_each_entry_safe(bo_va, tmp, &vm->va, it.rb) {
+ rbtree_postorder_for_each_entry_safe(bo_va, tmp,
+ &vm->va.rb_root, it.rb) {
interval_tree_remove(&bo_va->it, &vm->va);
r = radeon_bo_reserve(bo_va->bo, false);
if (!r) {
diff --git a/drivers/infiniband/core/umem_rbtree.c b/drivers/infiniband/core/umem_rbtree.c
index d176597b4d78..fc801920e341 100644
--- a/drivers/infiniband/core/umem_rbtree.c
+++ b/drivers/infiniband/core/umem_rbtree.c
@@ -72,7 +72,7 @@ INTERVAL_TREE_DEFINE(struct umem_odp_node, rb, u64, __subtree_last,
/* @last is not a part of the interval. See comment for function
* node_last.
*/
-int rbt_ib_umem_for_each_in_range(struct rb_root *root,
+int rbt_ib_umem_for_each_in_range(struct rb_root_cached *root,
u64 start, u64 last,
umem_call_back cb,
void *cookie)
@@ -95,7 +95,7 @@ int rbt_ib_umem_for_each_in_range(struct rb_root *root,
}
EXPORT_SYMBOL(rbt_ib_umem_for_each_in_range);
-struct ib_umem_odp *rbt_ib_umem_lookup(struct rb_root *root,
+struct ib_umem_odp *rbt_ib_umem_lookup(struct rb_root_cached *root,
u64 addr, u64 length)
{
struct umem_odp_node *node;
diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c
index 0ad3b05405d8..f73d4153dbd0 100644
--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -117,7 +117,7 @@ ssize_t ib_uverbs_get_context(struct ib_uverbs_file *file,
ucontext->closing = 0;
#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
- ucontext->umem_tree = RB_ROOT;
+ ucontext->umem_tree = RB_ROOT_CACHED;
init_rwsem(&ucontext->umem_rwsem);
ucontext->odp_mrs_count = 0;
INIT_LIST_HEAD(&ucontext->no_private_counters);
diff --git a/drivers/infiniband/hw/hfi1/mmu_rb.c b/drivers/infiniband/hw/hfi1/mmu_rb.c
index ccbf52c8ff6f..1835447dcd73 100644
--- a/drivers/infiniband/hw/hfi1/mmu_rb.c
+++ b/drivers/infiniband/hw/hfi1/mmu_rb.c
@@ -54,7 +54,7 @@
struct mmu_rb_handler {
struct mmu_notifier mn;
- struct rb_root root;
+ struct rb_root_cached root;
void *ops_arg;
spinlock_t lock; /* protect the RB tree */
struct mmu_rb_ops *ops;
@@ -111,7 +111,7 @@ int hfi1_mmu_rb_register(void *ops_arg, struct mm_struct *mm,
if (!handlr)
return -ENOMEM;
- handlr->root = RB_ROOT;
+ handlr->root = RB_ROOT_CACHED;
handlr->ops = ops;
handlr->ops_arg = ops_arg;
INIT_HLIST_NODE(&handlr->mn.hlist);
@@ -152,9 +152,9 @@ void hfi1_mmu_rb_unregister(struct mmu_rb_handler *handler)
INIT_LIST_HEAD(&del_list);
spin_lock_irqsave(&handler->lock, flags);
- while ((node = rb_first(&handler->root))) {
+ while ((node = rb_first_cached(&handler->root))) {
rbnode = rb_entry(node, struct mmu_rb_node, node);
- rb_erase(node, &handler->root);
+ rb_erase_cached(node, &handler->root);
/* move from LRU list to delete list */
list_move(&rbnode->list, &del_list);
}
@@ -305,7 +305,7 @@ static void mmu_notifier_mem_invalidate(struct mmu_notifier *mn,
{
struct mmu_rb_handler *handler =
container_of(mn, struct mmu_rb_handler, mn);
- struct rb_root *root = &handler->root;
+ struct rb_root_cached *root = &handler->root;
struct mmu_rb_node *node, *ptr = NULL;
unsigned long flags;
bool added = false;
diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c
index c49db7c33979..4381c0a9a873 100644
--- a/drivers/infiniband/hw/usnic/usnic_uiom.c
+++ b/drivers/infiniband/hw/usnic/usnic_uiom.c
@@ -227,7 +227,7 @@ static void __usnic_uiom_reg_release(struct usnic_uiom_pd *pd,
vpn_last = vpn_start + npages - 1;
spin_lock(&pd->lock);
- usnic_uiom_remove_interval(&pd->rb_root, vpn_start,
+ usnic_uiom_remove_interval(&pd->root, vpn_start,
vpn_last, &rm_intervals);
usnic_uiom_unmap_sorted_intervals(&rm_intervals, pd);
@@ -379,7 +379,7 @@ struct usnic_uiom_reg *usnic_uiom_reg_get(struct usnic_uiom_pd *pd,
err = usnic_uiom_get_intervals_diff(vpn_start, vpn_last,
(writable) ? IOMMU_WRITE : 0,
IOMMU_WRITE,
- &pd->rb_root,
+ &pd->root,
&sorted_diff_intervals);
if (err) {
usnic_err("Failed disjoint interval vpn [0x%lx,0x%lx] err %d\n",
@@ -395,7 +395,7 @@ struct usnic_uiom_reg *usnic_uiom_reg_get(struct usnic_uiom_pd *pd,
}
- err = usnic_uiom_insert_interval(&pd->rb_root, vpn_start, vpn_last,
+ err = usnic_uiom_insert_interval(&pd->root, vpn_start, vpn_last,
(writable) ? IOMMU_WRITE : 0);
if (err) {
usnic_err("Failed insert interval vpn [0x%lx,0x%lx] err %d\n",
diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.h b/drivers/infiniband/hw/usnic/usnic_uiom.h
index 45ca7c1613a7..431efe4143f4 100644
--- a/drivers/infiniband/hw/usnic/usnic_uiom.h
+++ b/drivers/infiniband/hw/usnic/usnic_uiom.h
@@ -55,7 +55,7 @@ struct usnic_uiom_dev {
struct usnic_uiom_pd {
struct iommu_domain *domain;
spinlock_t lock;
- struct rb_root rb_root;
+ struct rb_root_cached root;
struct list_head devs;
int dev_cnt;
};
diff --git a/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.c b/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.c
index 42b4b4c4e452..d399523206c7 100644
--- a/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.c
+++ b/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.c
@@ -100,9 +100,9 @@ static int interval_cmp(void *priv, struct list_head *a, struct list_head *b)
}
static void
-find_intervals_intersection_sorted(struct rb_root *root, unsigned long start,
- unsigned long last,
- struct list_head *list)
+find_intervals_intersection_sorted(struct rb_root_cached *root,
+ unsigned long start, unsigned long last,
+ struct list_head *list)
{
struct usnic_uiom_interval_node *node;
@@ -118,7 +118,7 @@ find_intervals_intersection_sorted(struct rb_root *root, unsigned long start,
int usnic_uiom_get_intervals_diff(unsigned long start, unsigned long last,
int flags, int flag_mask,
- struct rb_root *root,
+ struct rb_root_cached *root,
struct list_head *diff_set)
{
struct usnic_uiom_interval_node *interval, *tmp;
@@ -175,7 +175,7 @@ void usnic_uiom_put_interval_set(struct list_head *intervals)
kfree(interval);
}
-int usnic_uiom_insert_interval(struct rb_root *root, unsigned long start,
+int usnic_uiom_insert_interval(struct rb_root_cached *root, unsigned long start,
unsigned long last, int flags)
{
struct usnic_uiom_interval_node *interval, *tmp;
@@ -246,8 +246,9 @@ int usnic_uiom_insert_interval(struct rb_root *root, unsigned long start,
return err;
}
-void usnic_uiom_remove_interval(struct rb_root *root, unsigned long start,
- unsigned long last, struct list_head *removed)
+void usnic_uiom_remove_interval(struct rb_root_cached *root,
+ unsigned long start, unsigned long last,
+ struct list_head *removed)
{
struct usnic_uiom_interval_node *interval;
diff --git a/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.h b/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.h
index c0b0b876ab90..1d7fc3226bca 100644
--- a/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.h
+++ b/drivers/infiniband/hw/usnic/usnic_uiom_interval_tree.h
@@ -48,12 +48,12 @@ struct usnic_uiom_interval_node {
extern void
usnic_uiom_interval_tree_insert(struct usnic_uiom_interval_node *node,
- struct rb_root *root);
+ struct rb_root_cached *root);
extern void
usnic_uiom_interval_tree_remove(struct usnic_uiom_interval_node *node,
- struct rb_root *root);
+ struct rb_root_cached *root);
extern struct usnic_uiom_interval_node *
-usnic_uiom_interval_tree_iter_first(struct rb_root *root,
+usnic_uiom_interval_tree_iter_first(struct rb_root_cached *root,
unsigned long start,
unsigned long last);
extern struct usnic_uiom_interval_node *
@@ -63,7 +63,7 @@ usnic_uiom_interval_tree_iter_next(struct usnic_uiom_interval_node *node,
* Inserts {start...last} into {root}. If there are overlaps,
* nodes will be broken up and merged
*/
-int usnic_uiom_insert_interval(struct rb_root *root,
+int usnic_uiom_insert_interval(struct rb_root_cached *root,
unsigned long start, unsigned long last,
int flags);
/*
@@ -71,7 +71,7 @@ int usnic_uiom_insert_interval(struct rb_root *root,
* 'removed.' The caller is responsibile for freeing memory of nodes in
* 'removed.'
*/
-void usnic_uiom_remove_interval(struct rb_root *root,
+void usnic_uiom_remove_interval(struct rb_root_cached *root,
unsigned long start, unsigned long last,
struct list_head *removed);
/*
@@ -81,7 +81,7 @@ void usnic_uiom_remove_interval(struct rb_root *root,
int usnic_uiom_get_intervals_diff(unsigned long start,
unsigned long last, int flags,
int flag_mask,
- struct rb_root *root,
+ struct rb_root_cached *root,
struct list_head *diff_set);
/* Call this to free diff_set returned by usnic_uiom_get_intervals_diff */
void usnic_uiom_put_interval_set(struct list_head *intervals);
diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c
index e4613a3c362d..88dc214de068 100644
--- a/drivers/vhost/vhost.c
+++ b/drivers/vhost/vhost.c
@@ -1272,7 +1272,7 @@ static struct vhost_umem *vhost_umem_alloc(void)
if (!umem)
return NULL;
- umem->umem_tree = RB_ROOT;
+ umem->umem_tree = RB_ROOT_CACHED;
umem->numem = 0;
INIT_LIST_HEAD(&umem->umem_list);
diff --git a/drivers/vhost/vhost.h b/drivers/vhost/vhost.h
index f72095868b93..a0278ba6a8b4 100644
--- a/drivers/vhost/vhost.h
+++ b/drivers/vhost/vhost.h
@@ -71,7 +71,7 @@ struct vhost_umem_node {
};
struct vhost_umem {
- struct rb_root umem_tree;
+ struct rb_root_cached umem_tree;
struct list_head umem_list;
int numem;
};
diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index 52388611635e..9e755d37c575 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -334,7 +334,7 @@ static void remove_huge_page(struct page *page)
}
static void
-hugetlb_vmdelete_list(struct rb_root *root, pgoff_t start, pgoff_t end)
+hugetlb_vmdelete_list(struct rb_root_cached *root, pgoff_t start, pgoff_t end)
{
struct vm_area_struct *vma;
@@ -514,7 +514,7 @@ static int hugetlb_vmtruncate(struct inode *inode, loff_t offset)
i_size_write(inode, offset);
i_mmap_lock_write(mapping);
- if (!RB_EMPTY_ROOT(&mapping->i_mmap))
+ if (!RB_EMPTY_ROOT(&mapping->i_mmap.rb_root))
hugetlb_vmdelete_list(&mapping->i_mmap, pgoff, 0);
i_mmap_unlock_write(mapping);
remove_inode_hugepages(inode, offset, LLONG_MAX);
@@ -539,7 +539,7 @@ static long hugetlbfs_punch_hole(struct inode *inode, loff_t offset, loff_t len)
inode_lock(inode);
i_mmap_lock_write(mapping);
- if (!RB_EMPTY_ROOT(&mapping->i_mmap))
+ if (!RB_EMPTY_ROOT(&mapping->i_mmap.rb_root))
hugetlb_vmdelete_list(&mapping->i_mmap,
hole_start >> PAGE_SHIFT,
hole_end >> PAGE_SHIFT);
diff --git a/fs/inode.c b/fs/inode.c
index 5cbc8e6e9390..419a5c83b689 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -353,7 +353,7 @@ void address_space_init_once(struct address_space *mapping)
init_rwsem(&mapping->i_mmap_rwsem);
INIT_LIST_HEAD(&mapping->private_list);
spin_lock_init(&mapping->private_lock);
- mapping->i_mmap = RB_ROOT;
+ mapping->i_mmap = RB_ROOT_CACHED;
}
EXPORT_SYMBOL(address_space_init_once);
diff --git a/include/drm/drm_mm.h b/include/drm/drm_mm.h
index 49b292e98fec..8d10fc97801c 100644
--- a/include/drm/drm_mm.h
+++ b/include/drm/drm_mm.h
@@ -172,7 +172,7 @@ struct drm_mm {
* according to the (increasing) start address of the memory node. */
struct drm_mm_node head_node;
/* Keep an interval_tree for fast lookup of drm_mm_nodes by address. */
- struct rb_root interval_tree;
+ struct rb_root_cached interval_tree;
struct rb_root holes_size;
struct rb_root holes_addr;
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 5d11d4335125..57d7baa734cb 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -389,7 +389,7 @@ struct address_space {
struct radix_tree_root page_tree; /* radix tree of all pages */
spinlock_t tree_lock; /* and lock protecting it */
atomic_t i_mmap_writable;/* count VM_SHARED mappings */
- struct rb_root i_mmap; /* tree of private and shared mappings */
+ struct rb_root_cached i_mmap; /* tree of private and shared mappings */
struct rw_semaphore i_mmap_rwsem; /* protect tree, count, list */
/* Protected by tree_lock together with the radix tree */
unsigned long nrpages; /* number of total pages */
@@ -483,7 +483,7 @@ static inline void i_mmap_unlock_read(struct address_space *mapping)
*/
static inline int mapping_mapped(struct address_space *mapping)
{
- return !RB_EMPTY_ROOT(&mapping->i_mmap);
+ return !RB_EMPTY_ROOT(&mapping->i_mmap.rb_root);
}
/*
diff --git a/include/linux/interval_tree.h b/include/linux/interval_tree.h
index 724556aa3c95..202ee1283f4b 100644
--- a/include/linux/interval_tree.h
+++ b/include/linux/interval_tree.h
@@ -11,13 +11,15 @@ struct interval_tree_node {
};
extern void
-interval_tree_insert(struct interval_tree_node *node, struct rb_root *root);
+interval_tree_insert(struct interval_tree_node *node,
+ struct rb_root_cached *root);
extern void
-interval_tree_remove(struct interval_tree_node *node, struct rb_root *root);
+interval_tree_remove(struct interval_tree_node *node,
+ struct rb_root_cached *root);
extern struct interval_tree_node *
-interval_tree_iter_first(struct rb_root *root,
+interval_tree_iter_first(struct rb_root_cached *root,
unsigned long start, unsigned long last);
extern struct interval_tree_node *
diff --git a/include/linux/interval_tree_generic.h b/include/linux/interval_tree_generic.h
index 58370e1862ad..f096423c8cbd 100644
--- a/include/linux/interval_tree_generic.h
+++ b/include/linux/interval_tree_generic.h
@@ -65,11 +65,13 @@ RB_DECLARE_CALLBACKS(static, ITPREFIX ## _augment, ITSTRUCT, ITRB, \
\
/* Insert / remove interval nodes from the tree */ \
\
-ITSTATIC void ITPREFIX ## _insert(ITSTRUCT *node, struct rb_root *root) \
+ITSTATIC void ITPREFIX ## _insert(ITSTRUCT *node, \
+ struct rb_root_cached *root) \
{ \
- struct rb_node **link = &root->rb_node, *rb_parent = NULL; \
+ struct rb_node **link = &root->rb_root.rb_node, *rb_parent = NULL; \
ITTYPE start = ITSTART(node), last = ITLAST(node); \
ITSTRUCT *parent; \
+ bool leftmost = true; \
\
while (*link) { \
rb_parent = *link; \
@@ -78,18 +80,22 @@ ITSTATIC void ITPREFIX ## _insert(ITSTRUCT *node, struct rb_root *root) \
parent->ITSUBTREE = last; \
if (start < ITSTART(parent)) \
link = &parent->ITRB.rb_left; \
- else \
+ else { \
link = &parent->ITRB.rb_right; \
+ leftmost = false; \
+ } \
} \
\
node->ITSUBTREE = last; \
rb_link_node(&node->ITRB, rb_parent, link); \
- rb_insert_augmented(&node->ITRB, root, &ITPREFIX ## _augment); \
+ rb_insert_augmented_cached(&node->ITRB, root, \
+ leftmost, &ITPREFIX ## _augment); \
} \
\
-ITSTATIC void ITPREFIX ## _remove(ITSTRUCT *node, struct rb_root *root) \
+ITSTATIC void ITPREFIX ## _remove(ITSTRUCT *node, \
+ struct rb_root_cached *root) \
{ \
- rb_erase_augmented(&node->ITRB, root, &ITPREFIX ## _augment); \
+ rb_erase_augmented_cached(&node->ITRB, root, &ITPREFIX ## _augment); \
} \
\
/* \
@@ -140,15 +146,35 @@ ITPREFIX ## _subtree_search(ITSTRUCT *node, ITTYPE start, ITTYPE last) \
} \
\
ITSTATIC ITSTRUCT * \
-ITPREFIX ## _iter_first(struct rb_root *root, ITTYPE start, ITTYPE last) \
+ITPREFIX ## _iter_first(struct rb_root_cached *root, \
+ ITTYPE start, ITTYPE last) \
{ \
- ITSTRUCT *node; \
+ ITSTRUCT *node, *leftmost; \
\
- if (!root->rb_node) \
+ if (!root->rb_root.rb_node) \
return NULL; \
- node = rb_entry(root->rb_node, ITSTRUCT, ITRB); \
+ \
+ /* \
+ * Fastpath range intersection/overlap between A: [a0, a1] and \
+ * B: [b0, b1] is given by: \
+ * \
+ * a0 <= b1 && b0 <= a1 \
+ * \
+ * ... where A holds the lock range and B holds the smallest \
+ * 'start' and largest 'last' in the tree. For the later, we \
+ * rely on the root node, which by augmented interval tree \
+ * property, holds the largest value in its last-in-subtree. \
+ * This allows mitigating some of the tree walk overhead for \
+ * for non-intersecting ranges, maintained and consulted in O(1). \
+ */ \
+ node = rb_entry(root->rb_root.rb_node, ITSTRUCT, ITRB); \
if (node->ITSUBTREE < start) \
return NULL; \
+ \
+ leftmost = rb_entry(root->rb_leftmost, ITSTRUCT, ITRB); \
+ if (ITSTART(leftmost) > last) \
+ return NULL; \
+ \
return ITPREFIX ## _subtree_search(node, start, last); \
} \
\
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 46b9ac5e8569..3a2652efbbfb 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1992,13 +1992,13 @@ extern int nommu_shrink_inode_mappings(struct inode *, size_t, size_t);
/* interval_tree.c */
void vma_interval_tree_insert(struct vm_area_struct *node,
- struct rb_root *root);
+ struct rb_root_cached *root);
void vma_interval_tree_insert_after(struct vm_area_struct *node,
struct vm_area_struct *prev,
- struct rb_root *root);
+ struct rb_root_cached *root);
void vma_interval_tree_remove(struct vm_area_struct *node,
- struct rb_root *root);
-struct vm_area_struct *vma_interval_tree_iter_first(struct rb_root *root,
+ struct rb_root_cached *root);
+struct vm_area_struct *vma_interval_tree_iter_first(struct rb_root_cached *root,
unsigned long start, unsigned long last);
struct vm_area_struct *vma_interval_tree_iter_next(struct vm_area_struct *node,
unsigned long start, unsigned long last);
@@ -2008,11 +2008,12 @@ struct vm_area_struct *vma_interval_tree_iter_next(struct vm_area_struct *node,
vma; vma = vma_interval_tree_iter_next(vma, start, last))
void anon_vma_interval_tree_insert(struct anon_vma_chain *node,
- struct rb_root *root);
+ struct rb_root_cached *root);
void anon_vma_interval_tree_remove(struct anon_vma_chain *node,
- struct rb_root *root);
-struct anon_vma_chain *anon_vma_interval_tree_iter_first(
- struct rb_root *root, unsigned long start, unsigned long last);
+ struct rb_root_cached *root);
+struct anon_vma_chain *
+anon_vma_interval_tree_iter_first(struct rb_root_cached *root,
+ unsigned long start, unsigned long last);
struct anon_vma_chain *anon_vma_interval_tree_iter_next(
struct anon_vma_chain *node, unsigned long start, unsigned long last);
#ifdef CONFIG_DEBUG_VM_RB
diff --git a/include/linux/rmap.h b/include/linux/rmap.h
index 43ef2c30cb0f..22c298c6cc26 100644
--- a/include/linux/rmap.h
+++ b/include/linux/rmap.h
@@ -55,7 +55,9 @@ struct anon_vma {
* is serialized by a system wide lock only visible to
* mm_take_all_locks() (mm_all_locks_mutex).
*/
- struct rb_root rb_root; /* Interval tree of private "related" vmas */
+
+ /* Interval tree of private "related" vmas */
+ struct rb_root_cached rb_root;
};
/*
diff --git a/include/rdma/ib_umem_odp.h b/include/rdma/ib_umem_odp.h
index fb67554aabd6..5eb7f5bc8248 100644
--- a/include/rdma/ib_umem_odp.h
+++ b/include/rdma/ib_umem_odp.h
@@ -111,22 +111,25 @@ int ib_umem_odp_map_dma_pages(struct ib_umem *umem, u64 start_offset, u64 bcnt,
void ib_umem_odp_unmap_dma_pages(struct ib_umem *umem, u64 start_offset,
u64 bound);
-void rbt_ib_umem_insert(struct umem_odp_node *node, struct rb_root *root);
-void rbt_ib_umem_remove(struct umem_odp_node *node, struct rb_root *root);
+void rbt_ib_umem_insert(struct umem_odp_node *node,
+ struct rb_root_cached *root);
+void rbt_ib_umem_remove(struct umem_odp_node *node,
+ struct rb_root_cached *root);
typedef int (*umem_call_back)(struct ib_umem *item, u64 start, u64 end,
void *cookie);
/*
* Call the callback on each ib_umem in the range. Returns the logical or of
* the return values of the functions called.
*/
-int rbt_ib_umem_for_each_in_range(struct rb_root *root, u64 start, u64 end,
+int rbt_ib_umem_for_each_in_range(struct rb_root_cached *root,
+ u64 start, u64 end,
umem_call_back cb, void *cookie);
/*
* Find first region intersecting with address range.
* Return NULL if not found
*/
-struct ib_umem_odp *rbt_ib_umem_lookup(struct rb_root *root,
+struct ib_umem_odp *rbt_ib_umem_lookup(struct rb_root_cached *root,
u64 addr, u64 length);
static inline int ib_umem_mmu_notifier_retry(struct ib_umem *item,
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 0e480a5630d4..3b54b19a8eac 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -1417,7 +1417,7 @@ struct ib_ucontext {
struct pid *tgid;
#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
- struct rb_root umem_tree;
+ struct rb_root_cached umem_tree;
/*
* Protects .umem_rbroot and tree, as well as odp_mrs_count and
* mmu notifiers registration.
diff --git a/lib/interval_tree_test.c b/lib/interval_tree_test.c
index df495fe81421..0e343fd29570 100644
--- a/lib/interval_tree_test.c
+++ b/lib/interval_tree_test.c
@@ -19,14 +19,14 @@ __param(bool, search_all, false, "Searches will iterate all nodes in the tree");
__param(uint, max_endpoint, ~0, "Largest value for the interval's endpoint");
-static struct rb_root root = RB_ROOT;
+static struct rb_root_cached root = RB_ROOT_CACHED;
static struct interval_tree_node *nodes = NULL;
static u32 *queries = NULL;
static struct rnd_state rnd;
static inline unsigned long
-search(struct rb_root *root, unsigned long start, unsigned long last)
+search(struct rb_root_cached *root, unsigned long start, unsigned long last)
{
struct interval_tree_node *node;
unsigned long results = 0;
diff --git a/mm/interval_tree.c b/mm/interval_tree.c
index f2c2492681bf..b47664358796 100644
--- a/mm/interval_tree.c
+++ b/mm/interval_tree.c
@@ -28,7 +28,7 @@ INTERVAL_TREE_DEFINE(struct vm_area_struct, shared.rb,
/* Insert node immediately after prev in the interval tree */
void vma_interval_tree_insert_after(struct vm_area_struct *node,
struct vm_area_struct *prev,
- struct rb_root *root)
+ struct rb_root_cached *root)
{
struct rb_node **link;
struct vm_area_struct *parent;
@@ -55,7 +55,7 @@ void vma_interval_tree_insert_after(struct vm_area_struct *node,
node->shared.rb_subtree_last = last;
rb_link_node(&node->shared.rb, &parent->shared.rb, link);
- rb_insert_augmented(&node->shared.rb, root,
+ rb_insert_augmented(&node->shared.rb, &root->rb_root,
&vma_interval_tree_augment);
}
@@ -74,7 +74,7 @@ INTERVAL_TREE_DEFINE(struct anon_vma_chain, rb, unsigned long, rb_subtree_last,
static inline, __anon_vma_interval_tree)
void anon_vma_interval_tree_insert(struct anon_vma_chain *node,
- struct rb_root *root)
+ struct rb_root_cached *root)
{
#ifdef CONFIG_DEBUG_VM_RB
node->cached_vma_start = avc_start_pgoff(node);
@@ -84,13 +84,13 @@ void anon_vma_interval_tree_insert(struct anon_vma_chain *node,
}
void anon_vma_interval_tree_remove(struct anon_vma_chain *node,
- struct rb_root *root)
+ struct rb_root_cached *root)
{
__anon_vma_interval_tree_remove(node, root);
}
struct anon_vma_chain *
-anon_vma_interval_tree_iter_first(struct rb_root *root,
+anon_vma_interval_tree_iter_first(struct rb_root_cached *root,
unsigned long first, unsigned long last)
{
return __anon_vma_interval_tree_iter_first(root, first, last);
diff --git a/mm/memory.c b/mm/memory.c
index cbb57194687e..d533a8913535 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -2593,7 +2593,7 @@ static void unmap_mapping_range_vma(struct vm_area_struct *vma,
zap_page_range_single(vma, start_addr, end_addr - start_addr, details);
}
-static inline void unmap_mapping_range_tree(struct rb_root *root,
+static inline void unmap_mapping_range_tree(struct rb_root_cached *root,
struct zap_details *details)
{
struct vm_area_struct *vma;
@@ -2657,7 +2657,7 @@ void unmap_mapping_range(struct address_space *mapping,
details.last_index = ULONG_MAX;
i_mmap_lock_write(mapping);
- if (unlikely(!RB_EMPTY_ROOT(&mapping->i_mmap)))
+ if (unlikely(!RB_EMPTY_ROOT(&mapping->i_mmap.rb_root)))
unmap_mapping_range_tree(&mapping->i_mmap, &details);
i_mmap_unlock_write(mapping);
}
diff --git a/mm/mmap.c b/mm/mmap.c
index 7f8cfe9d9b4d..81d731030a6f 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c
@@ -684,7 +684,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
struct mm_struct *mm = vma->vm_mm;
struct vm_area_struct *next = vma->vm_next, *orig_vma = vma;
struct address_space *mapping = NULL;
- struct rb_root *root = NULL;
+ struct rb_root_cached *root = NULL;
struct anon_vma *anon_vma = NULL;
struct file *file = vma->vm_file;
bool start_changed = false, end_changed = false;
@@ -3317,7 +3317,7 @@ static DEFINE_MUTEX(mm_all_locks_mutex);
static void vm_lock_anon_vma(struct mm_struct *mm, struct anon_vma *anon_vma)
{
- if (!test_bit(0, (unsigned long *) &anon_vma->root->rb_root.rb_node)) {
+ if (!test_bit(0, (unsigned long *) &anon_vma->rb_root.rb_root.rb_node)) {
/*
* The LSB of head.next can't change from under us
* because we hold the mm_all_locks_mutex.
@@ -3333,7 +3333,7 @@ static void vm_lock_anon_vma(struct mm_struct *mm, struct anon_vma *anon_vma)
* anon_vma->root->rwsem.
*/
if (__test_and_set_bit(0, (unsigned long *)
- &anon_vma->root->rb_root.rb_node))
+ &anon_vma->root->rb_root.rb_root.rb_node))
BUG();
}
}
@@ -3435,7 +3435,7 @@ int mm_take_all_locks(struct mm_struct *mm)
static void vm_unlock_anon_vma(struct anon_vma *anon_vma)
{
- if (test_bit(0, (unsigned long *) &anon_vma->root->rb_root.rb_node)) {
+ if (test_bit(0, (unsigned long *) &anon_vma->root->rb_root.rb_root.rb_node)) {
/*
* The LSB of head.next can't change to 0 from under
* us because we hold the mm_all_locks_mutex.
@@ -3449,7 +3449,7 @@ static void vm_unlock_anon_vma(struct anon_vma *anon_vma)
* anon_vma->root->rwsem.
*/
if (!__test_and_clear_bit(0, (unsigned long *)
- &anon_vma->root->rb_root.rb_node))
+ &anon_vma->root->rb_root.rb_root.rb_node))
BUG();
anon_vma_unlock_write(anon_vma);
}
diff --git a/mm/rmap.c b/mm/rmap.c
index ced14f1af6dc..ad479e5e081d 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -390,7 +390,7 @@ void unlink_anon_vmas(struct vm_area_struct *vma)
* Leave empty anon_vmas on the list - we'll need
* to free them outside the lock.
*/
- if (RB_EMPTY_ROOT(&anon_vma->rb_root)) {
+ if (RB_EMPTY_ROOT(&anon_vma->rb_root.rb_root)) {
anon_vma->parent->degree--;
continue;
}
@@ -424,7 +424,7 @@ static void anon_vma_ctor(void *data)
init_rwsem(&anon_vma->rwsem);
atomic_set(&anon_vma->refcount, 0);
- anon_vma->rb_root = RB_ROOT;
+ anon_vma->rb_root = RB_ROOT_CACHED;
}
void __init anon_vma_init(void)
--
2.12.0
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH 7/9] lib/interval-tree: Correct comment wrt generic flavor
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
` (5 preceding siblings ...)
2017-06-29 17:15 ` [PATCH 6/9] lib/interval_tree: Fast overlap detection Davidlohr Bueso
@ 2017-06-29 17:15 ` Davidlohr Bueso
2017-06-29 17:15 ` [PATCH 8/9] procfs: Use faster rb_first_cached() Davidlohr Bueso
` (3 subsequent siblings)
10 siblings, 0 replies; 15+ messages in thread
From: Davidlohr Bueso @ 2017-06-29 17:15 UTC (permalink / raw)
To: mingo, peterz, akpm
Cc: torvalds, jack, kirill.shutemov, ldufour, mhocko, mgorman, dave,
linux-kernel, Davidlohr Bueso
interval_tree.h _is_ the generic flavor.
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
include/linux/interval_tree_generic.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/linux/interval_tree_generic.h b/include/linux/interval_tree_generic.h
index f096423c8cbd..1f97ce26cccc 100644
--- a/include/linux/interval_tree_generic.h
+++ b/include/linux/interval_tree_generic.h
@@ -33,7 +33,7 @@
* ITSTATIC: 'static' or empty
* ITPREFIX: prefix to use for the inline tree definitions
*
- * Note - before using this, please consider if non-generic version
+ * Note - before using this, please consider if generic version
* (interval_tree.h) would work for you...
*/
--
2.12.0
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH 8/9] procfs: Use faster rb_first_cached()
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
` (6 preceding siblings ...)
2017-06-29 17:15 ` [PATCH 7/9] lib/interval-tree: Correct comment wrt generic flavor Davidlohr Bueso
@ 2017-06-29 17:15 ` Davidlohr Bueso
2017-06-29 17:15 ` [PATCH 9/9] fs/epoll: " Davidlohr Bueso
` (2 subsequent siblings)
10 siblings, 0 replies; 15+ messages in thread
From: Davidlohr Bueso @ 2017-06-29 17:15 UTC (permalink / raw)
To: mingo, peterz, akpm
Cc: torvalds, jack, kirill.shutemov, ldufour, mhocko, mgorman, dave,
linux-kernel, Davidlohr Bueso
... such that we can avoid the tree walks to get the
node with the smallest key. Semantically the same,
as the previously used rb_first(), but O(1). The
main overhead is the extra footprint for the cached
rb_node pointer, which should not matter for procfs.
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
fs/proc/generic.c | 26 ++++++++++++++------------
fs/proc/internal.h | 2 +-
fs/proc/proc_net.c | 2 +-
fs/proc/root.c | 2 +-
4 files changed, 17 insertions(+), 15 deletions(-)
diff --git a/fs/proc/generic.c b/fs/proc/generic.c
index e3cda0b5968f..ab6496356dc2 100644
--- a/fs/proc/generic.c
+++ b/fs/proc/generic.c
@@ -40,8 +40,8 @@ static int proc_match(unsigned int len, const char *name, struct proc_dir_entry
static struct proc_dir_entry *pde_subdir_first(struct proc_dir_entry *dir)
{
- return rb_entry_safe(rb_first(&dir->subdir), struct proc_dir_entry,
- subdir_node);
+ return rb_entry_safe(rb_first_cached(&dir->subdir),
+ struct proc_dir_entry, subdir_node);
}
static struct proc_dir_entry *pde_subdir_next(struct proc_dir_entry *dir)
@@ -54,7 +54,7 @@ static struct proc_dir_entry *pde_subdir_find(struct proc_dir_entry *dir,
const char *name,
unsigned int len)
{
- struct rb_node *node = dir->subdir.rb_node;
+ struct rb_node *node = dir->subdir.rb_root.rb_node;
while (node) {
struct proc_dir_entry *de = rb_entry(node,
@@ -75,8 +75,9 @@ static struct proc_dir_entry *pde_subdir_find(struct proc_dir_entry *dir,
static bool pde_subdir_insert(struct proc_dir_entry *dir,
struct proc_dir_entry *de)
{
- struct rb_root *root = &dir->subdir;
- struct rb_node **new = &root->rb_node, *parent = NULL;
+ struct rb_root_cached *root = &dir->subdir;
+ struct rb_node **new = &root->rb_root.rb_node, *parent = NULL;
+ bool leftmost = true;
/* Figure out where to put new node */
while (*new) {
@@ -88,15 +89,16 @@ static bool pde_subdir_insert(struct proc_dir_entry *dir,
parent = *new;
if (result < 0)
new = &(*new)->rb_left;
- else if (result > 0)
+ else if (result > 0) {
new = &(*new)->rb_right;
- else
+ leftmost = false;
+ } else
return false;
}
/* Add new node and rebalance tree. */
rb_link_node(&de->subdir_node, parent, new);
- rb_insert_color(&de->subdir_node, root);
+ rb_insert_color_cached(&de->subdir_node, root, leftmost);
return true;
}
@@ -369,7 +371,7 @@ static struct proc_dir_entry *__proc_create(struct proc_dir_entry **parent,
ent->namelen = qstr.len;
ent->mode = mode;
ent->nlink = nlink;
- ent->subdir = RB_ROOT;
+ ent->subdir = RB_ROOT_CACHED;
atomic_set(&ent->count, 1);
spin_lock_init(&ent->pde_unload_lock);
INIT_LIST_HEAD(&ent->pde_openers);
@@ -545,7 +547,7 @@ void remove_proc_entry(const char *name, struct proc_dir_entry *parent)
de = pde_subdir_find(parent, fn, len);
if (de)
- rb_erase(&de->subdir_node, &parent->subdir);
+ rb_erase_cached(&de->subdir_node, &parent->subdir);
write_unlock(&proc_subdir_lock);
if (!de) {
WARN(1, "name '%s'\n", name);
@@ -582,13 +584,13 @@ int remove_proc_subtree(const char *name, struct proc_dir_entry *parent)
write_unlock(&proc_subdir_lock);
return -ENOENT;
}
- rb_erase(&root->subdir_node, &parent->subdir);
+ rb_erase_cached(&root->subdir_node, &parent->subdir);
de = root;
while (1) {
next = pde_subdir_first(de);
if (next) {
- rb_erase(&next->subdir_node, &de->subdir);
+ rb_erase_cached(&next->subdir_node, &de->subdir);
de = next;
continue;
}
diff --git a/fs/proc/internal.h b/fs/proc/internal.h
index 07b16318223f..4db394edb6d9 100644
--- a/fs/proc/internal.h
+++ b/fs/proc/internal.h
@@ -40,7 +40,7 @@ struct proc_dir_entry {
const struct inode_operations *proc_iops;
const struct file_operations *proc_fops;
struct proc_dir_entry *parent;
- struct rb_root subdir;
+ struct rb_root_cached subdir;
struct rb_node subdir_node;
void *data;
atomic_t count; /* use count */
diff --git a/fs/proc/proc_net.c b/fs/proc/proc_net.c
index d72fc40241d9..a2bf369c923d 100644
--- a/fs/proc/proc_net.c
+++ b/fs/proc/proc_net.c
@@ -196,7 +196,7 @@ static __net_init int proc_net_ns_init(struct net *net)
if (!netd)
goto out;
- netd->subdir = RB_ROOT;
+ netd->subdir = RB_ROOT_CACHED;
netd->data = net;
netd->nlink = 2;
netd->namelen = 3;
diff --git a/fs/proc/root.c b/fs/proc/root.c
index deecb397daa3..926fb27f4ca2 100644
--- a/fs/proc/root.c
+++ b/fs/proc/root.c
@@ -210,7 +210,7 @@ struct proc_dir_entry proc_root = {
.proc_iops = &proc_root_inode_operations,
.proc_fops = &proc_root_operations,
.parent = &proc_root,
- .subdir = RB_ROOT,
+ .subdir = RB_ROOT_CACHED,
.name = "/proc",
};
--
2.12.0
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH 9/9] fs/epoll: Use faster rb_first_cached()
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
` (7 preceding siblings ...)
2017-06-29 17:15 ` [PATCH 8/9] procfs: Use faster rb_first_cached() Davidlohr Bueso
@ 2017-06-29 17:15 ` Davidlohr Bueso
2017-07-05 17:47 ` [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Peter Zijlstra
2017-07-19 22:54 ` Andrew Morton
10 siblings, 0 replies; 15+ messages in thread
From: Davidlohr Bueso @ 2017-06-29 17:15 UTC (permalink / raw)
To: mingo, peterz, akpm
Cc: torvalds, jack, kirill.shutemov, ldufour, mhocko, mgorman, dave,
linux-kernel, Davidlohr Bueso
... such that we can avoid the tree walks to get the
node with the smallest key. Semantically the same,
as the previously used rb_first(), but O(1). The
main overhead is the extra footprint for the cached
rb_node pointer, which should not matter for epoll.
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
fs/eventpoll.c | 30 ++++++++++++++++--------------
1 file changed, 16 insertions(+), 14 deletions(-)
diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index e767e4389cb1..c56e842bc9cc 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -205,7 +205,7 @@ struct eventpoll {
struct list_head rdllist;
/* RB tree root used to store monitored fd structs */
- struct rb_root rbr;
+ struct rb_root_cached rbr;
/*
* This is a single linked list that chains all the "struct epitem" that
@@ -791,7 +791,7 @@ static int ep_remove(struct eventpoll *ep, struct epitem *epi)
list_del_rcu(&epi->fllink);
spin_unlock(&file->f_lock);
- rb_erase(&epi->rbn, &ep->rbr);
+ rb_erase_cached(&epi->rbn, &ep->rbr);
spin_lock_irqsave(&ep->lock, flags);
if (ep_is_linked(&epi->rdllink))
@@ -835,7 +835,7 @@ static void ep_free(struct eventpoll *ep)
/*
* Walks through the whole tree by unregistering poll callbacks.
*/
- for (rbp = rb_first(&ep->rbr); rbp; rbp = rb_next(rbp)) {
+ for (rbp = rb_first_cached(&ep->rbr); rbp; rbp = rb_next(rbp)) {
epi = rb_entry(rbp, struct epitem, rbn);
ep_unregister_pollwait(ep, epi);
@@ -851,7 +851,7 @@ static void ep_free(struct eventpoll *ep)
* a lockdep warning.
*/
mutex_lock(&ep->mtx);
- while ((rbp = rb_first(&ep->rbr)) != NULL) {
+ while ((rbp = rb_first_cached(&ep->rbr)) != NULL) {
epi = rb_entry(rbp, struct epitem, rbn);
ep_remove(ep, epi);
cond_resched();
@@ -958,7 +958,7 @@ static void ep_show_fdinfo(struct seq_file *m, struct file *f)
struct rb_node *rbp;
mutex_lock(&ep->mtx);
- for (rbp = rb_first(&ep->rbr); rbp; rbp = rb_next(rbp)) {
+ for (rbp = rb_first_cached(&ep->rbr); rbp; rbp = rb_next(rbp)) {
struct epitem *epi = rb_entry(rbp, struct epitem, rbn);
struct inode *inode = file_inode(epi->ffd.file);
@@ -1035,7 +1035,7 @@ static int ep_alloc(struct eventpoll **pep)
init_waitqueue_head(&ep->wq);
init_waitqueue_head(&ep->poll_wait);
INIT_LIST_HEAD(&ep->rdllist);
- ep->rbr = RB_ROOT;
+ ep->rbr = RB_ROOT_CACHED;
ep->ovflist = EP_UNACTIVE_PTR;
ep->user = user;
@@ -1061,7 +1061,7 @@ static struct epitem *ep_find(struct eventpoll *ep, struct file *file, int fd)
struct epoll_filefd ffd;
ep_set_ffd(&ffd, file, fd);
- for (rbp = ep->rbr.rb_node; rbp; ) {
+ for (rbp = ep->rbr.rb_root.rb_node; rbp; ) {
epi = rb_entry(rbp, struct epitem, rbn);
kcmp = ep_cmp_ffd(&ffd, &epi->ffd);
if (kcmp > 0)
@@ -1083,7 +1083,7 @@ static struct epitem *ep_find_tfd(struct eventpoll *ep, int tfd, unsigned long t
struct rb_node *rbp;
struct epitem *epi;
- for (rbp = rb_first(&ep->rbr); rbp; rbp = rb_next(rbp)) {
+ for (rbp = rb_first_cached(&ep->rbr); rbp; rbp = rb_next(rbp)) {
epi = rb_entry(rbp, struct epitem, rbn);
if (epi->ffd.fd == tfd) {
if (toff == 0)
@@ -1263,20 +1263,22 @@ static void ep_ptable_queue_proc(struct file *file, wait_queue_head_t *whead,
static void ep_rbtree_insert(struct eventpoll *ep, struct epitem *epi)
{
int kcmp;
- struct rb_node **p = &ep->rbr.rb_node, *parent = NULL;
+ struct rb_node **p = &ep->rbr.rb_root.rb_node, *parent = NULL;
struct epitem *epic;
+ bool leftmost = true;
while (*p) {
parent = *p;
epic = rb_entry(parent, struct epitem, rbn);
kcmp = ep_cmp_ffd(&epi->ffd, &epic->ffd);
- if (kcmp > 0)
+ if (kcmp > 0) {
p = &parent->rb_right;
- else
+ leftmost = false;
+ } else
p = &parent->rb_left;
}
rb_link_node(&epi->rbn, parent, p);
- rb_insert_color(&epi->rbn, &ep->rbr);
+ rb_insert_color_cached(&epi->rbn, &ep->rbr, leftmost);
}
@@ -1520,7 +1522,7 @@ static int ep_insert(struct eventpoll *ep, struct epoll_event *event,
list_del_rcu(&epi->fllink);
spin_unlock(&tfile->f_lock);
- rb_erase(&epi->rbn, &ep->rbr);
+ rb_erase_cached(&epi->rbn, &ep->rbr);
error_unregister:
ep_unregister_pollwait(ep, epi);
@@ -1868,7 +1870,7 @@ static int ep_loop_check_proc(void *priv, void *cookie, int call_nests)
mutex_lock_nested(&ep->mtx, call_nests + 1);
ep->visited = 1;
list_add(&ep->visited_list_link, &visited_list);
- for (rbp = rb_first(&ep->rbr); rbp; rbp = rb_next(rbp)) {
+ for (rbp = rb_first_cached(&ep->rbr); rbp; rbp = rb_next(rbp)) {
epi = rb_entry(rbp, struct epitem, rbn);
if (unlikely(is_file_epoll(epi->ffd.file))) {
ep_tovisit = epi->ffd.file->private_data;
--
2.12.0
^ permalink raw reply related [flat|nested] 15+ messages in thread* Re: [PATCH -next v3 0/9] rbtree: Cache leftmost node internally
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
` (8 preceding siblings ...)
2017-06-29 17:15 ` [PATCH 9/9] fs/epoll: " Davidlohr Bueso
@ 2017-07-05 17:47 ` Peter Zijlstra
2017-07-19 22:54 ` Andrew Morton
10 siblings, 0 replies; 15+ messages in thread
From: Peter Zijlstra @ 2017-07-05 17:47 UTC (permalink / raw)
To: Davidlohr Bueso
Cc: mingo, akpm, torvalds, jack, kirill.shutemov, ldufour, mhocko,
mgorman, linux-kernel
On Thu, Jun 29, 2017 at 10:15:44AM -0700, Davidlohr Bueso wrote:
> Here's a proposal for 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.
>
> Patch 1: Layout the rb machinery.
>
> Patches 2-5: Make use of the internal leftmost node in scheduler and
> rt mutexes and cfq.
>
> Patch 6: Implements fast overlap checks for interval trees.
>
> Patch 7: rocket science.
>
> Patches 8,9: New patches that convert to O(1) rb_first_cached().
>
> The series has survived booting, kernel builds and pistress workloads.
>
> Ingo, I know it's late in the game, but could it be considered for
> v4.13? Given that v2 has been there a while and there are no issues
> currently. Applies on top of today's -next.
IIRC akpm typically collects rb-tree patches.
In any case:
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
^ permalink raw reply [flat|nested] 15+ messages in thread* Re: [PATCH -next v3 0/9] rbtree: Cache leftmost node internally
2017-06-29 17:15 [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Davidlohr Bueso
` (9 preceding siblings ...)
2017-07-05 17:47 ` [PATCH -next v3 0/9] rbtree: Cache leftmost node internally Peter Zijlstra
@ 2017-07-19 22:54 ` Andrew Morton
2017-07-19 23:00 ` Davidlohr Bueso
10 siblings, 1 reply; 15+ messages in thread
From: Andrew Morton @ 2017-07-19 22:54 UTC (permalink / raw)
To: Davidlohr Bueso
Cc: mingo, peterz, torvalds, jack, kirill.shutemov, ldufour, mhocko,
mgorman, linux-kernel
On Thu, 29 Jun 2017 10:15:44 -0700 Davidlohr Bueso <dave@stgolabs.net> wrote:
> Changes from v2 (https://lkml.org/lkml/2017/6/8/857):
> - Fixed 0day reported crash for drm_mm selftest program. We were
> not correctly using the cached version of rbtree with the allocated
> nodes.
> - Added cfq patch to use internal rbtree caching.
> - Added Christian's and Jan's reviews.
>
> Changes from v1 (https://marc.info/?l=linux-kernel&m=149611025616685):
> - No longer rfc.
> - Removed bogus semimcolon in rb_first_cached()
> - Updated missing interval tree user drivers/infiniband/hw/hfi1/
> - Removed redundant @cached arg in when erasing a node.
> - Added more patches that make use of rb_first_cached(), which I
> thought might be worth it: procfs and epoll.
> - Cc more people for patch 5, which touches drivers such as infiniband
> and gpu. The rest of the changes are pretty covered with the current
> Cc'ed maintainers and mm folks.
>
> Hi,
>
> Here's a proposal for 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.
That's nice. Except the series adds more lines than it removes.
> (ii) Optimize all interval tree users.
Was any attempt made to quantify the benefit?
> (iii) Convert at least two new users (epoll and procfs) to the new interface.
>
^ permalink raw reply [flat|nested] 15+ messages in thread* Re: [PATCH -next v3 0/9] rbtree: Cache leftmost node internally
2017-07-19 22:54 ` Andrew Morton
@ 2017-07-19 23:00 ` Davidlohr Bueso
0 siblings, 0 replies; 15+ messages in thread
From: Davidlohr Bueso @ 2017-07-19 23:00 UTC (permalink / raw)
To: Andrew Morton
Cc: mingo, peterz, torvalds, jack, kirill.shutemov, ldufour, mhocko,
mgorman, linux-kernel
On Wed, 19 Jul 2017, Andrew Morton wrote:
>On Thu, 29 Jun 2017 10:15:44 -0700 Davidlohr Bueso <dave@stgolabs.net> wrote:
>
>> Changes from v2 (https://lkml.org/lkml/2017/6/8/857):
>> - Fixed 0day reported crash for drm_mm selftest program. We were
>> not correctly using the cached version of rbtree with the allocated
>> nodes.
>> - Added cfq patch to use internal rbtree caching.
>> - Added Christian's and Jan's reviews.
>>
>> Changes from v1 (https://marc.info/?l=linux-kernel&m=149611025616685):
>> - No longer rfc.
>> - Removed bogus semimcolon in rb_first_cached()
>> - Updated missing interval tree user drivers/infiniband/hw/hfi1/
>> - Removed redundant @cached arg in when erasing a node.
>> - Added more patches that make use of rb_first_cached(), which I
>> thought might be worth it: procfs and epoll.
>> - Cc more people for patch 5, which touches drivers such as infiniband
>> and gpu. The rest of the changes are pretty covered with the current
>> Cc'ed maintainers and mm folks.
>>
>> Hi,
>>
>> Here's a proposal for 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.
>
>That's nice. Except the series adds more lines than it removes.
>
>> (ii) Optimize all interval tree users.
>
>Was any attempt made to quantify the benefit?
Yes, but ultimately it will depend a lot on the workload and size of the tree.
For bare numbers, on a Xeon E5-2450 @ 2.10GHz the cost of a rb_first() was
~60 cycles for 100 nodes, and ~75 cycles with 1000 nodes. fwiw.
Thanks,
Davidlohr
^ permalink raw reply [flat|nested] 15+ messages in thread