summaryrefslogtreecommitdiffstats
path: root/lib/maple_tree.c
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2026-10-02 12:17:24 -0700
committerLinus Torvalds <torvalds@linux-foundation.org>2026-10-02 12:17:24 -0700
commit3f1fe48a36b0b6722dc3fd421d93512bac138e9a (patch)
treeb767d7f6e26bc64334d3f14f8422d187239dc45d /lib/maple_tree.c
downloadlinux-stable-3f1fe48a36b0b6722dc3fd421d93512bac138e9a.tar.gz
linux-stable-3f1fe48a36b0b6722dc3fd421d93512bac138e9a.zip
Merge tag 'io_uring-7.3-20261002' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linuxgrafted
Pull io_uring fixes from Jens Axboe: - Fix a task_work add use-after-free with SQPOLL. The sqpoll thread could pop and complete the last request while io_req_normal_work_add() was still looking at them after the mpscq push. Use the same approach as DEFER_TASKRUN to protect from that, holding an RCU read lock across the add, and have exit wait for an RCU grace period for SQPOLL rings as well. - CQE32 ring fixes: correct the free entry check for 32b CQEs, zero the big_cqe for aux CQEs, and only post the dummy skip CQE on CQE_MIXED rings - Mark the source filter table as COW when cloning bpf filters, so registering another filter on the source doesn't modify the shared table in place - Initialize the task context before running the BPF loop - Requeue zcrx multishot receives stopped by a local resource - End a TX_TIMESTAMP multishot cmd when the CQ is full (lollipopkit) * tag 'io_uring-7.3-20261002' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux: io_uring: fix task_work add use-after-free with SQPOLL io_uring/cmd_net: end TX_TIMESTAMP multishot when the CQ is full io_uring/zcrx: requeue multishot receives stopped by a local resource io_uring: initialize task context before running the BPF loop io_uring: zero big_cqe for aux CQEs on CQE32 rings io_uring: fix free entry check for 32b CQEs on CQE32 rings io_uring: only post the dummy skip CQE on CQE_MIXED rings io_uring/bpf_filter: mark source as COW when cloning filters
Diffstat (limited to 'lib/maple_tree.c')
-rw-r--r--lib/maple_tree.c7204
1 files changed, 7204 insertions, 0 deletions
diff --git a/lib/maple_tree.c b/lib/maple_tree.c
new file mode 100644
index 000000000..1aba6cced
--- /dev/null
+++ b/lib/maple_tree.c
@@ -0,0 +1,7204 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Maple Tree implementation
+ * Copyright (c) 2018-2022 Oracle Corporation
+ * Authors: Liam R. Howlett <liam@infradead.org>
+ * Matthew Wilcox <willy@infradead.org>
+ * Copyright (c) 2023 ByteDance
+ * Author: Peng Zhang <zhangpeng.00@bytedance.com>
+ */
+
+/*
+ * DOC: Interesting implementation details of the Maple Tree
+ *
+ * Each node type has a number of slots for entries and a number of slots for
+ * pivots. In the case of dense nodes, the pivots are implied by the position
+ * and are simply the slot index + the minimum of the node.
+ *
+ * In regular B-Tree terms, pivots are called keys. The term pivot is used to
+ * indicate that the tree is specifying ranges. Pivots may appear in the
+ * subtree with an entry attached to the value whereas keys are unique to a
+ * specific position of a B-tree. Pivot values are inclusive of the slot with
+ * the same index.
+ *
+ *
+ * The following illustrates the layout of a range64 nodes slots and pivots.
+ *
+ *
+ * Slots -> | 0 | 1 | 2 | ... | 12 | 13 | 14 | 15 |
+ * ┬ ┬ ┬ ┬ ┬ ┬ ┬ ┬ ┬
+ * │ │ │ │ │ │ │ │ └─ Implied maximum
+ * │ │ │ │ │ │ │ └─ Pivot 14
+ * │ │ │ │ │ │ └─ Pivot 13
+ * │ │ │ │ │ └─ Pivot 12
+ * │ │ │ │ └─ Pivot 11
+ * │ │ │ └─ Pivot 2
+ * │ │ └─ Pivot 1
+ * │ └─ Pivot 0
+ * └─ Implied minimum
+ *
+ * Slot contents:
+ * Internal (non-leaf) nodes contain pointers to other nodes.
+ * Leaf nodes contain entries.
+ *
+ * The location of interest is often referred to as an offset. All offsets have
+ * a slot, but the last offset has an implied pivot from the node above (or
+ * UINT_MAX for the root node.
+ *
+ * Ranges complicate certain write activities. When modifying any of
+ * the B-tree variants, it is known that one entry will either be added or
+ * deleted. When modifying the Maple Tree, one store operation may overwrite
+ * the entire data set, or one half of the tree, or the middle half of the tree.
+ *
+ */
+
+
+#include <linux/maple_tree.h>
+#include <linux/xarray.h>
+#include <linux/types.h>
+#include <linux/export.h>
+#include <linux/slab.h>
+#include <linux/limits.h>
+#include <asm/barrier.h>
+
+#define CREATE_TRACE_POINTS
+#include <trace/events/maple_tree.h>
+
+#define TP_FCT tracepoint_string(__func__)
+
+/*
+ * Kernel pointer hashing renders much of the maple tree dump useless as tagged
+ * pointers get hashed to arbitrary values.
+ *
+ * If CONFIG_DEBUG_VM_MAPLE_TREE is set we are in a debug mode where it is
+ * permissible to bypass this. Otherwise remain cautious and retain the hashing.
+ *
+ * Userland doesn't know about %px so also use %p there.
+ */
+#if defined(__KERNEL__) && defined(CONFIG_DEBUG_VM_MAPLE_TREE)
+#define PTR_FMT "%px"
+#else
+#define PTR_FMT "%p"
+#endif
+
+#define MA_ROOT_PARENT 1
+
+/*
+ * Maple state flags
+ * * MA_STATE_PREALLOC - Preallocated nodes, WARN_ON allocation
+ */
+#define MA_STATE_PREALLOC 1
+
+#define ma_parent_ptr(x) ((struct maple_pnode *)(x))
+#define mas_tree_parent(x) ((unsigned long)(x->tree) | MA_ROOT_PARENT)
+#define ma_mnode_ptr(x) ((struct maple_node *)(x))
+#define ma_enode_ptr(x) ((struct maple_enode *)(x))
+static struct kmem_cache *maple_node_cache;
+
+#ifdef CONFIG_DEBUG_MAPLE_TREE
+static const unsigned long mt_max[] = {
+ [maple_dense] = MAPLE_NODE_SLOTS,
+ [maple_leaf_64] = ULONG_MAX,
+ [maple_range_64] = ULONG_MAX,
+ [maple_arange_64] = ULONG_MAX,
+ [maple_copy] = ULONG_MAX,
+};
+#define mt_node_max(x) mt_max[mte_node_type(x)]
+#endif
+
+static const unsigned char mt_slots[] = {
+ [maple_dense] = MAPLE_NODE_SLOTS,
+ [maple_leaf_64] = MAPLE_RANGE64_SLOTS,
+ [maple_range_64] = MAPLE_RANGE64_SLOTS,
+ [maple_arange_64] = MAPLE_ARANGE64_SLOTS,
+ [maple_copy] = 3,
+};
+#define mt_slot_count(x) mt_slots[mte_node_type(x)]
+
+static const unsigned char mt_pivots[] = {
+ [maple_dense] = 0,
+ [maple_leaf_64] = MAPLE_RANGE64_SLOTS - 1,
+ [maple_range_64] = MAPLE_RANGE64_SLOTS - 1,
+ [maple_arange_64] = MAPLE_ARANGE64_SLOTS - 1,
+ [maple_copy] = 3,
+};
+#define mt_pivot_count(x) mt_pivots[mte_node_type(x)]
+
+static const unsigned char mt_min_slots[] = {
+ [maple_dense] = MAPLE_NODE_SLOTS / 2,
+ [maple_leaf_64] = (MAPLE_RANGE64_SLOTS / 2) - 2,
+ [maple_range_64] = (MAPLE_RANGE64_SLOTS / 2) - 2,
+ [maple_arange_64] = (MAPLE_ARANGE64_SLOTS / 2) - 1,
+ [maple_copy] = 1, /* Should never be used */
+};
+#define mt_min_slot_count(x) mt_min_slots[mte_node_type(x)]
+
+/* Functions */
+static inline struct maple_node *mt_alloc_one(gfp_t gfp)
+{
+ return kmem_cache_alloc(maple_node_cache, gfp);
+}
+
+static inline void mt_free_bulk(size_t size, void __rcu **nodes)
+{
+ kmem_cache_free_bulk(maple_node_cache, size, (void **)nodes);
+}
+
+static void mt_return_sheaf(struct slab_sheaf *sheaf)
+{
+ kmem_cache_return_sheaf(maple_node_cache, GFP_NOWAIT, sheaf);
+}
+
+static struct slab_sheaf *mt_get_sheaf(gfp_t gfp, int count)
+{
+ return kmem_cache_prefill_sheaf(maple_node_cache, gfp, count);
+}
+
+static int mt_refill_sheaf(gfp_t gfp, struct slab_sheaf **sheaf,
+ unsigned int size)
+{
+ return kmem_cache_refill_sheaf(maple_node_cache, gfp, sheaf, size);
+}
+
+/*
+ * ma_free_rcu() - Use rcu callback to free a maple node
+ * @node: The node to free
+ *
+ * The maple tree uses the parent pointer to indicate this node is no longer in
+ * use and will be freed.
+ */
+static void ma_free_rcu(struct maple_node *node)
+{
+ WARN_ON(node->parent != ma_parent_ptr(node));
+ kfree_rcu(node, rcu);
+}
+
+static void mt_set_height(struct maple_tree *mt, unsigned char height)
+{
+ unsigned int new_flags = mt->ma_flags;
+
+ new_flags &= ~MT_FLAGS_HEIGHT_MASK;
+ MT_BUG_ON(mt, height > MAPLE_HEIGHT_MAX);
+ new_flags |= height << MT_FLAGS_HEIGHT_OFFSET;
+ mt->ma_flags = new_flags;
+}
+
+static unsigned int mas_mt_height(struct ma_state *mas)
+{
+ return mt_height(mas->tree);
+}
+
+static inline unsigned int mt_attr(struct maple_tree *mt)
+{
+ return mt->ma_flags & ~MT_FLAGS_HEIGHT_MASK;
+}
+
+static __always_inline enum maple_type mte_node_type(
+ const struct maple_enode *entry)
+{
+ return ((unsigned long)entry >> MAPLE_NODE_TYPE_SHIFT) &
+ MAPLE_NODE_TYPE_MASK;
+}
+
+static __always_inline bool ma_is_dense(const enum maple_type type)
+{
+ return type < maple_leaf_64;
+}
+
+static __always_inline bool ma_is_leaf(const enum maple_type type)
+{
+ return type < maple_range_64;
+}
+
+static __always_inline bool mte_is_leaf(const struct maple_enode *entry)
+{
+ return ma_is_leaf(mte_node_type(entry));
+}
+
+/*
+ * We also reserve values with the bottom two bits set to '10' which are
+ * below 4096
+ */
+static __always_inline bool mt_is_reserved(const void *entry)
+{
+ return ((unsigned long)entry < MAPLE_RESERVED_RANGE) &&
+ xa_is_internal(entry);
+}
+
+static __always_inline void mas_set_err(struct ma_state *mas, long err)
+{
+ mas->node = MA_ERROR(err);
+ mas->status = ma_error;
+}
+
+static __always_inline bool mas_is_ptr(const struct ma_state *mas)
+{
+ return mas->status == ma_root;
+}
+
+static __always_inline bool mas_is_start(const struct ma_state *mas)
+{
+ return mas->status == ma_start;
+}
+
+static __always_inline bool mas_is_none(const struct ma_state *mas)
+{
+ return mas->status == ma_none;
+}
+
+static __always_inline bool mas_is_paused(const struct ma_state *mas)
+{
+ return mas->status == ma_pause;
+}
+
+static __always_inline bool mas_is_overflow(struct ma_state *mas)
+{
+ return mas->status == ma_overflow;
+}
+
+static inline bool mas_is_underflow(struct ma_state *mas)
+{
+ return mas->status == ma_underflow;
+}
+
+static inline void mas_make_walkable(struct ma_state *mas)
+{
+ if (!mas_is_active(mas) && !mas_is_start(mas))
+ mas->status = ma_start;
+}
+
+static __always_inline struct maple_node *mte_to_node(
+ const struct maple_enode *entry)
+{
+ return (struct maple_node *)((unsigned long)entry & ~MAPLE_NODE_MASK);
+}
+
+/*
+ * mte_to_mat() - Convert a maple encoded node to a maple topiary node.
+ * @entry: The maple encoded node
+ *
+ * Return: a maple topiary pointer
+ */
+static inline struct maple_topiary *mte_to_mat(const struct maple_enode *entry)
+{
+ return (struct maple_topiary *)
+ ((unsigned long)entry & ~MAPLE_NODE_MASK);
+}
+
+/*
+ * mas_mn() - Get the maple state node.
+ * @mas: The maple state
+ *
+ * Return: the maple node (not encoded - bare pointer).
+ */
+static inline struct maple_node *mas_mn(const struct ma_state *mas)
+{
+ return mte_to_node(mas->node);
+}
+
+/*
+ * mte_set_node_dead() - Set a maple encoded node as dead.
+ * @mn: The maple encoded node.
+ */
+static inline void mte_set_node_dead(struct maple_enode *mn)
+{
+ mte_to_node(mn)->parent = ma_parent_ptr(mte_to_node(mn));
+ smp_wmb(); /* Needed for RCU */
+}
+
+/* Bit 1 indicates the root is a node */
+#define MAPLE_ROOT_NODE 0x02
+/* maple_type stored bit 3-6 */
+#define MAPLE_ENODE_TYPE_SHIFT 0x03
+/* Bit 2 means a NULL somewhere below */
+#define MAPLE_ENODE_NULL 0x04
+
+static inline struct maple_enode *mt_mk_node(const struct maple_node *node,
+ enum maple_type type)
+{
+ return (void *)((unsigned long)node |
+ (type << MAPLE_ENODE_TYPE_SHIFT) | MAPLE_ENODE_NULL);
+}
+
+static inline void ma_init_slot(void __rcu **slot, const struct maple_node *mn,
+ const enum maple_type mt)
+{
+ /* WARNING: this is unsafe if the slot is exposed to readers. */
+ RCU_INIT_POINTER(*slot, (void *)mt_mk_node(mn, mt));
+}
+
+static inline void *mte_mk_root(const struct maple_enode *node)
+{
+ return (void *)((unsigned long)node | MAPLE_ROOT_NODE);
+}
+
+static inline void *mte_safe_root(const struct maple_enode *node)
+{
+ return (void *)((unsigned long)node & ~MAPLE_ROOT_NODE);
+}
+
+static inline void __maybe_unused *mte_set_full(const struct maple_enode *node)
+{
+ return (void *)((unsigned long)node & ~MAPLE_ENODE_NULL);
+}
+
+static inline void __maybe_unused *mte_clear_full(const struct maple_enode *node)
+{
+ return (void *)((unsigned long)node | MAPLE_ENODE_NULL);
+}
+
+static inline bool __maybe_unused mte_has_null(const struct maple_enode *node)
+{
+ return (unsigned long)node & MAPLE_ENODE_NULL;
+}
+
+static __always_inline bool ma_is_root(struct maple_node *node)
+{
+ return ((unsigned long)node->parent & MA_ROOT_PARENT);
+}
+
+static __always_inline bool mte_is_root(const struct maple_enode *node)
+{
+ return ma_is_root(mte_to_node(node));
+}
+
+static __always_inline bool mt_is_alloc(struct maple_tree *mt)
+{
+ return (mt->ma_flags & MT_FLAGS_ALLOC_RANGE);
+}
+
+/*
+ * The Parent Pointer
+ * Excluding root, the parent pointer is 256B aligned like all other tree nodes.
+ * When storing a 32 or 64 bit values, the offset can fit into 5 bits. The 16
+ * bit values need an extra bit to store the offset. This extra bit comes from
+ * a reuse of the last bit in the node type. This is possible by using bit 1 to
+ * indicate if bit 2 is part of the type or the slot.
+ *
+ * Node types:
+ * 0b??1 = Root
+ * 0b?00 = 16 bit nodes
+ * 0b010 = 32 bit nodes
+ * 0b110 = 64 bit nodes
+ *
+ * Slot size and alignment
+ * 0b??1 : Root
+ * 0b?00 : 16 bit values, type in 0-1, slot in 2-7
+ * 0b010 : 32 bit values, type in 0-2, slot in 3-7
+ * 0b110 : 64 bit values, type in 0-2, slot in 3-7
+ */
+
+#define MAPLE_PARENT_ROOT 0x01
+
+#define MAPLE_PARENT_SLOT_SHIFT 0x03
+#define MAPLE_PARENT_SLOT_MASK 0xF8
+
+#define MAPLE_PARENT_16B_SLOT_SHIFT 0x02
+#define MAPLE_PARENT_16B_SLOT_MASK 0xFC
+
+#define MAPLE_PARENT_RANGE64 0x06
+#define MAPLE_PARENT_RANGE32 0x02
+#define MAPLE_PARENT_NOT_RANGE16 0x02
+
+/*
+ * mte_parent_shift() - Get the parent shift for the slot storage.
+ * @parent: The parent pointer cast as an unsigned long
+ * Return: The shift into that pointer to the star to of the slot
+ */
+static inline unsigned long mte_parent_shift(unsigned long parent)
+{
+ /* Note bit 1 == 0 means 16B */
+ if (likely(parent & MAPLE_PARENT_NOT_RANGE16))
+ return MAPLE_PARENT_SLOT_SHIFT;
+
+ return MAPLE_PARENT_16B_SLOT_SHIFT;
+}
+
+/*
+ * mte_parent_slot_mask() - Get the slot mask for the parent.
+ * @parent: The parent pointer cast as an unsigned long.
+ * Return: The slot mask for that parent.
+ */
+static inline unsigned long mte_parent_slot_mask(unsigned long parent)
+{
+ /* Note bit 1 == 0 means 16B */
+ if (likely(parent & MAPLE_PARENT_NOT_RANGE16))
+ return MAPLE_PARENT_SLOT_MASK;
+
+ return MAPLE_PARENT_16B_SLOT_MASK;
+}
+
+/*
+ * mas_parent_type() - Return the maple_type of the parent from the stored
+ * parent type.
+ * @mas: The maple state
+ * @enode: The maple_enode to extract the parent's enum
+ * Return: The node->parent maple_type
+ */
+static inline
+enum maple_type mas_parent_type(struct ma_state *mas, struct maple_enode *enode)
+{
+ unsigned long p_type;
+
+ p_type = (unsigned long)mte_to_node(enode)->parent;
+ if (WARN_ON(p_type & MAPLE_PARENT_ROOT))
+ return 0;
+
+ p_type &= MAPLE_NODE_MASK;
+ p_type &= ~mte_parent_slot_mask(p_type);
+ switch (p_type) {
+ case MAPLE_PARENT_RANGE64: /* or MAPLE_PARENT_ARANGE64 */
+ if (mt_is_alloc(mas->tree))
+ return maple_arange_64;
+ return maple_range_64;
+ }
+
+ return 0;
+}
+
+/*
+ * mte_parent_slot() - get the parent slot of @enode.
+ * @enode: The encoded maple node.
+ *
+ * Return: The slot in the parent node where @enode resides.
+ */
+static __always_inline
+unsigned int mte_parent_slot(const struct maple_enode *enode)
+{
+ unsigned long val = (unsigned long)mte_to_node(enode)->parent;
+
+ if (unlikely(val & MA_ROOT_PARENT))
+ return 0;
+
+ /*
+ * Okay to use MAPLE_PARENT_16B_SLOT_MASK as the last bit will be lost
+ * by shift if the parent shift is MAPLE_PARENT_SLOT_SHIFT
+ */
+ return (val & MAPLE_PARENT_16B_SLOT_MASK) >> mte_parent_shift(val);
+}
+
+/*
+ * mte_parent() - Get the parent of @node.
+ * @enode: The encoded maple node.
+ *
+ * Return: The parent maple node.
+ */
+static __always_inline
+struct maple_node *mte_parent(const struct maple_enode *enode)
+{
+ return (void *)((unsigned long)
+ (mte_to_node(enode)->parent) & ~MAPLE_NODE_MASK);
+}
+
+/*
+ * ma_dead_node() - check if the @enode is dead.
+ * @enode: The encoded maple node
+ *
+ * Return: true if dead, false otherwise.
+ */
+static __always_inline bool ma_dead_node(const struct maple_node *node)
+{
+ struct maple_node *parent;
+
+ /* Do not reorder reads from the node prior to the parent check */
+ smp_rmb();
+ parent = (void *)((unsigned long) node->parent & ~MAPLE_NODE_MASK);
+ return (parent == node);
+}
+
+/*
+ * mte_dead_node() - check if the @enode is dead.
+ * @enode: The encoded maple node
+ *
+ * Return: true if dead, false otherwise.
+ */
+static __always_inline bool mte_dead_node(const struct maple_enode *enode)
+{
+ struct maple_node *node;
+
+ node = mte_to_node(enode);
+ return ma_dead_node(node);
+}
+
+/*
+ * ma_pivots() - Get a pointer to the maple node pivots.
+ * @node: the maple node
+ * @type: the node type
+ *
+ * In the event of a dead node, this array may be %NULL
+ *
+ * Return: A pointer to the maple node pivots
+ */
+static inline unsigned long *ma_pivots(struct maple_node *node,
+ enum maple_type type)
+{
+ switch (type) {
+ case maple_arange_64:
+ return node->ma64.pivot;
+ case maple_range_64:
+ case maple_leaf_64:
+ return node->mr64.pivot;
+ case maple_copy:
+ return node->cp.pivot;
+ case maple_dense:
+ return NULL;
+ }
+ return NULL;
+}
+
+/*
+ * ma_gaps() - Get a pointer to the maple node gaps.
+ * @node: the maple node
+ * @type: the node type
+ *
+ * Return: A pointer to the maple node gaps
+ */
+static inline unsigned long *ma_gaps(struct maple_node *node,
+ enum maple_type type)
+{
+ switch (type) {
+ case maple_arange_64:
+ return node->ma64.gap;
+ case maple_copy:
+ return node->cp.gap;
+ case maple_range_64:
+ case maple_leaf_64:
+ case maple_dense:
+ return NULL;
+ }
+ return NULL;
+}
+
+/*
+ * mas_safe_pivot() - get the pivot at @piv or mas->max.
+ * @mas: The maple state
+ * @pivots: The pointer to the maple node pivots
+ * @piv: The pivot to fetch
+ * @type: The maple node type
+ *
+ * Return: The pivot at @piv within the limit of the @pivots array, @mas->max
+ * otherwise.
+ */
+static __always_inline unsigned long
+mas_safe_pivot(const struct ma_state *mas, unsigned long *pivots,
+ unsigned char piv, enum maple_type type)
+{
+ if (piv >= mt_pivots[type])
+ return mas->max;
+
+ return pivots[piv];
+}
+
+/*
+ * mas_safe_min() - Return the minimum for a given offset.
+ * @mas: The maple state
+ * @pivots: The pointer to the maple node pivots
+ * @offset: The offset into the pivot array
+ *
+ * Return: The minimum range value that is contained in @offset.
+ */
+static inline unsigned long
+mas_safe_min(struct ma_state *mas, unsigned long *pivots, unsigned char offset)
+{
+ if (likely(offset))
+ return pivots[offset - 1] + 1;
+
+ return mas->min;
+}
+
+/*
+ * mte_set_pivot() - Set a pivot to a value in an encoded maple node.
+ * @mn: The encoded maple node
+ * @piv: The pivot offset
+ * @val: The value of the pivot
+ */
+static inline void mte_set_pivot(struct maple_enode *mn, unsigned char piv,
+ unsigned long val)
+{
+ struct maple_node *node = mte_to_node(mn);
+ enum maple_type type = mte_node_type(mn);
+
+ BUG_ON(piv >= mt_pivots[type]);
+ switch (type) {
+ case maple_range_64:
+ case maple_leaf_64:
+ node->mr64.pivot[piv] = val;
+ break;
+ case maple_arange_64:
+ node->ma64.pivot[piv] = val;
+ break;
+ case maple_copy:
+ case maple_dense:
+ break;
+ }
+
+}
+
+/*
+ * ma_slots() - Get a pointer to the maple node slots.
+ * @mn: The maple node
+ * @mt: The maple node type
+ *
+ * Return: A pointer to the maple node slots
+ */
+static inline void __rcu **ma_slots(struct maple_node *mn, enum maple_type mt)
+{
+ switch (mt) {
+ case maple_arange_64:
+ return mn->ma64.slot;
+ case maple_range_64:
+ case maple_leaf_64:
+ return mn->mr64.slot;
+ case maple_copy:
+ return mn->cp.slot;
+ case maple_dense:
+ return mn->slot;
+ }
+
+ return NULL;
+}
+
+static inline bool mt_write_locked(const struct maple_tree *mt)
+{
+ return mt_external_lock(mt) ? mt_write_lock_is_held(mt) :
+ lockdep_is_held(&mt->ma_lock);
+}
+
+static __always_inline bool mt_locked(const struct maple_tree *mt)
+{
+ return mt_external_lock(mt) ? mt_lock_is_held(mt) :
+ lockdep_is_held(&mt->ma_lock);
+}
+
+static __always_inline void *mt_slot(const struct maple_tree *mt,
+ void __rcu **slots, unsigned char offset)
+{
+ return rcu_dereference_check(slots[offset], mt_locked(mt));
+}
+
+static __always_inline void *mt_slot_locked(struct maple_tree *mt,
+ void __rcu **slots, unsigned char offset)
+{
+ return rcu_dereference_protected(slots[offset], mt_write_locked(mt));
+}
+/*
+ * mas_slot_locked() - Get the slot value when holding the maple tree lock.
+ * @mas: The maple state
+ * @slots: The pointer to the slots
+ * @offset: The offset into the slots array to fetch
+ *
+ * Return: The entry stored in @slots at the @offset.
+ */
+static __always_inline void *mas_slot_locked(struct ma_state *mas,
+ void __rcu **slots, unsigned char offset)
+{
+ return mt_slot_locked(mas->tree, slots, offset);
+}
+
+/*
+ * mas_slot() - Get the slot value when not holding the maple tree lock.
+ * @mas: The maple state
+ * @slots: The pointer to the slots
+ * @offset: The offset into the slots array to fetch
+ *
+ * Return: The entry stored in @slots at the @offset
+ */
+static __always_inline void *mas_slot(struct ma_state *mas, void __rcu **slots,
+ unsigned char offset)
+{
+ return mt_slot(mas->tree, slots, offset);
+}
+
+/*
+ * mas_root() - Get the maple tree root.
+ * @mas: The maple state.
+ *
+ * Return: The pointer to the root of the tree
+ */
+static __always_inline void *mas_root(struct ma_state *mas)
+{
+ return rcu_dereference_check(mas->tree->ma_root, mt_locked(mas->tree));
+}
+
+static inline void *mt_root_locked(struct maple_tree *mt)
+{
+ return rcu_dereference_protected(mt->ma_root, mt_write_locked(mt));
+}
+
+/*
+ * mas_root_locked() - Get the maple tree root when holding the maple tree lock.
+ * @mas: The maple state.
+ *
+ * Return: The pointer to the root of the tree
+ */
+static inline void *mas_root_locked(struct ma_state *mas)
+{
+ return mt_root_locked(mas->tree);
+}
+
+static inline struct maple_metadata *ma_meta(struct maple_node *mn,
+ enum maple_type mt)
+{
+ switch (mt) {
+ case maple_arange_64:
+ return &mn->ma64.meta;
+ default:
+ return &mn->mr64.meta;
+ }
+}
+
+/*
+ * ma_set_meta() - Set the metadata information of a node.
+ * @mn: The maple node
+ * @mt: The maple node type
+ * @offset: The offset of the highest sub-gap in this node.
+ * @end: The end of the data in this node.
+ */
+static inline void ma_set_meta(struct maple_node *mn, enum maple_type mt,
+ unsigned char offset, unsigned char end)
+{
+ struct maple_metadata *meta = ma_meta(mn, mt);
+
+ meta->gap = offset;
+ meta->end = end;
+}
+
+/*
+ * mt_clear_meta() - clear the metadata information of a node, if it exists
+ * @mt: The maple tree
+ * @mn: The maple node
+ * @type: The maple node type
+ */
+static inline void mt_clear_meta(struct maple_tree *mt, struct maple_node *mn,
+ enum maple_type type)
+{
+ struct maple_metadata *meta;
+ unsigned long *pivots;
+ void __rcu **slots;
+ void *next;
+
+ switch (type) {
+ case maple_range_64:
+ pivots = mn->mr64.pivot;
+ if (unlikely(pivots[MAPLE_RANGE64_SLOTS - 2])) {
+ slots = mn->mr64.slot;
+ next = mt_slot_locked(mt, slots,
+ MAPLE_RANGE64_SLOTS - 1);
+ if (unlikely((mte_to_node(next) &&
+ mte_node_type(next))))
+ return; /* no metadata, could be node */
+ }
+ fallthrough;
+ case maple_arange_64:
+ meta = ma_meta(mn, type);
+ break;
+ default:
+ return;
+ }
+
+ meta->gap = 0;
+ meta->end = 0;
+}
+
+/*
+ * ma_meta_end() - Get the data end of a node from the metadata
+ * @mn: The maple node
+ * @mt: The maple node type
+ */
+static inline unsigned char ma_meta_end(struct maple_node *mn,
+ enum maple_type mt)
+{
+ struct maple_metadata *meta = ma_meta(mn, mt);
+
+ return meta->end;
+}
+
+/*
+ * ma_meta_gap() - Get the largest gap location of a node from the metadata
+ * @mn: The maple node
+ */
+static inline unsigned char ma_meta_gap(struct maple_node *mn)
+{
+ return mn->ma64.meta.gap;
+}
+
+/*
+ * ma_set_meta_gap() - Set the largest gap location in a nodes metadata
+ * @mn: The maple node
+ * @mt: The maple node type
+ * @offset: The location of the largest gap.
+ */
+static inline void ma_set_meta_gap(struct maple_node *mn, enum maple_type mt,
+ unsigned char offset)
+{
+
+ struct maple_metadata *meta = ma_meta(mn, mt);
+
+ meta->gap = offset;
+}
+
+/*
+ * mas_set_parent_slots() - Bulk operation to set many slot parent pointers
+ * @mas: The maple state
+ * @parent: The encoded maple node that is the parent of @enode.
+ * @slot: The slot that of the @enode.
+ * @start_slot: The offset into @slot
+ * @count: The number of slots to set (eg: exclusive)
+ */
+static inline
+void mas_set_parent_slots(struct ma_state *mas, struct maple_enode *parent,
+ void __rcu **slots, unsigned char start_slot, unsigned char count)
+{
+ unsigned long val;
+ unsigned long shift;
+ unsigned long type;
+ enum maple_type p_type = mte_node_type(parent);
+ unsigned char i;
+
+ MAS_BUG_ON(mas, p_type != maple_range_64 &&
+ p_type != maple_arange_64);
+
+ shift = MAPLE_PARENT_SLOT_SHIFT;
+ type = MAPLE_PARENT_RANGE64;
+
+ val = (unsigned long)parent;
+ val &= ~MAPLE_NODE_MASK;
+
+ for (i = 0; i < count; i++) {
+ unsigned long pval = val | ((start_slot + i) << shift) | type;
+ struct maple_enode *child;
+
+ child = mt_slot_locked(mas->tree, slots, i);
+ mte_to_node(child)->parent = ma_parent_ptr(pval);
+ }
+}
+
+/*
+ * mat_add() - Add a @dead_enode to the ma_topiary of a list of dead nodes.
+ * @mat: the ma_topiary, a linked list of dead nodes.
+ * @dead_enode: the node to be marked as dead and added to the tail of the list
+ *
+ * Add the @dead_enode to the linked list in @mat.
+ */
+static inline void mat_add(struct ma_topiary *mat,
+ struct maple_enode *dead_enode)
+{
+ mte_set_node_dead(dead_enode);
+ mte_to_mat(dead_enode)->next = NULL;
+ if (!mat->tail) {
+ mat->tail = mat->head = dead_enode;
+ return;
+ }
+
+ mte_to_mat(mat->tail)->next = dead_enode;
+ mat->tail = dead_enode;
+}
+
+static void mt_free_walk(struct rcu_head *head);
+static void mt_destroy_walk(struct maple_enode *enode, struct maple_tree *mt,
+ bool free);
+/*
+ * mas_mat_destroy() - Free all nodes and subtrees in a dead list.
+ * @mas: the maple state
+ * @mat: the ma_topiary linked list of dead nodes to free.
+ *
+ * Destroy walk a dead list.
+ */
+static void mas_mat_destroy(struct ma_state *mas, struct ma_topiary *mat)
+{
+ struct maple_enode *next;
+ struct maple_node *node;
+ bool in_rcu = mt_in_rcu(mas->tree);
+
+ while (mat->head) {
+ next = mte_to_mat(mat->head)->next;
+ node = mte_to_node(mat->head);
+ mt_destroy_walk(mat->head, mas->tree, !in_rcu);
+ if (in_rcu)
+ call_rcu(&node->rcu, mt_free_walk);
+ mat->head = next;
+ }
+}
+/*
+ * mas_descend() - Descend into the slot stored in the ma_state.
+ * @mas: the maple state.
+ *
+ * Note: Not RCU safe, only use in write side or debug code.
+ */
+static inline void mas_descend(struct ma_state *mas)
+{
+ enum maple_type type;
+ unsigned long *pivots;
+ struct maple_node *node;
+ void __rcu **slots;
+
+ node = mas_mn(mas);
+ type = mte_node_type(mas->node);
+ pivots = ma_pivots(node, type);
+ slots = ma_slots(node, type);
+
+ if (mas->offset)
+ mas->min = pivots[mas->offset - 1] + 1;
+ mas->max = mas_safe_pivot(mas, pivots, mas->offset, type);
+ mas->node = mas_slot(mas, slots, mas->offset);
+}
+
+/*
+ * mas_ascend() - Walk up a level of the tree.
+ * @mas: The maple state
+ *
+ * Sets the @mas->max and @mas->min for the parent node of mas->node. This
+ * may cause several levels of walking up to find the correct min and max.
+ * May find a dead node which will cause a premature return.
+ * Return: 1 on dead node, 0 otherwise
+ */
+static int mas_ascend(struct ma_state *mas)
+{
+ struct maple_enode *p_enode; /* parent enode. */
+ struct maple_enode *a_enode; /* ancestor enode. */
+ struct maple_node *a_node; /* ancestor node. */
+ struct maple_node *p_node; /* parent node. */
+ unsigned char a_slot;
+ enum maple_type a_type;
+ unsigned long min, max;
+ unsigned long *pivots;
+ bool set_max = false, set_min = false;
+
+ a_node = mas_mn(mas);
+ if (ma_is_root(a_node)) {
+ mas->offset = 0;
+ return 0;
+ }
+
+ p_node = mte_parent(mas->node);
+ if (unlikely(a_node == p_node))
+ return 1;
+
+ a_type = mas_parent_type(mas, mas->node);
+ mas->offset = mte_parent_slot(mas->node);
+ a_enode = mt_mk_node(p_node, a_type);
+
+ /* Check to make sure all parent information is still accurate */
+ if (p_node != mte_parent(mas->node))
+ return 1;
+
+ mas->node = a_enode;
+
+ if (mte_is_root(a_enode)) {
+ mas->max = ULONG_MAX;
+ mas->min = 0;
+ return 0;
+ }
+
+ min = 0;
+ max = ULONG_MAX;
+
+ /*
+ * !mas->offset implies that parent node min == mas->min.
+ * mas->offset > 0 implies that we need to walk up to find the
+ * implied pivot min.
+ */
+ if (!mas->offset) {
+ min = mas->min;
+ set_min = true;
+ }
+
+ if (mas->max == ULONG_MAX)
+ set_max = true;
+
+ do {
+ p_enode = a_enode;
+ a_type = mas_parent_type(mas, p_enode);
+ a_node = mte_parent(p_enode);
+ a_slot = mte_parent_slot(p_enode);
+ a_enode = mt_mk_node(a_node, a_type);
+ pivots = ma_pivots(a_node, a_type);
+
+ if (unlikely(ma_dead_node(a_node)))
+ return 1;
+
+ if (!set_min && a_slot) {
+ set_min = true;
+ min = pivots[a_slot - 1] + 1;
+ }
+
+ if (!set_max && a_slot < mt_pivots[a_type]) {
+ set_max = true;
+ max = pivots[a_slot];
+ }
+
+ if (unlikely(ma_dead_node(a_node)))
+ return 1;
+
+ if (unlikely(ma_is_root(a_node)))
+ break;
+
+ } while (!set_min || !set_max);
+
+ mas->max = max;
+ mas->min = min;
+ return 0;
+}
+
+/*
+ * mas_pop_node() - Get a previously allocated maple node from the maple state.
+ * @mas: The maple state
+ *
+ * Return: A pointer to a maple node.
+ */
+static __always_inline struct maple_node *mas_pop_node(struct ma_state *mas)
+{
+ struct maple_node *ret;
+
+ if (mas->alloc) {
+ ret = mas->alloc;
+ mas->alloc = NULL;
+ goto out;
+ }
+
+ if (WARN_ON_ONCE(!mas->sheaf))
+ return NULL;
+
+ ret = kmem_cache_alloc_from_sheaf(maple_node_cache, GFP_NOWAIT, mas->sheaf);
+
+out:
+ memset(ret, 0, sizeof(*ret));
+ return ret;
+}
+
+/*
+ * mas_alloc_nodes() - Allocate nodes into a maple state
+ * @mas: The maple state
+ * @gfp: The GFP Flags
+ */
+static inline void mas_alloc_nodes(struct ma_state *mas, gfp_t gfp)
+{
+ if (!mas->node_request)
+ return;
+
+ if (mas->node_request == 1) {
+ if (mas->sheaf)
+ goto use_sheaf;
+
+ if (mas->alloc)
+ return;
+
+ mas->alloc = mt_alloc_one(gfp);
+ if (!mas->alloc)
+ goto error;
+
+ mas->node_request = 0;
+ return;
+ }
+
+use_sheaf:
+ if (unlikely(mas->alloc)) {
+ kfree(mas->alloc);
+ mas->alloc = NULL;
+ }
+
+ if (mas->sheaf) {
+ unsigned long refill;
+
+ refill = mas->node_request;
+ if (kmem_cache_sheaf_size(mas->sheaf) >= refill) {
+ mas->node_request = 0;
+ return;
+ }
+
+ if (mt_refill_sheaf(gfp, &mas->sheaf, refill))
+ goto error;
+
+ mas->node_request = 0;
+ return;
+ }
+
+ mas->sheaf = mt_get_sheaf(gfp, mas->node_request);
+ if (likely(mas->sheaf)) {
+ mas->node_request = 0;
+ return;
+ }
+
+error:
+ mas_set_err(mas, -ENOMEM);
+}
+
+static inline void mas_empty_nodes(struct ma_state *mas)
+{
+ mas->node_request = 0;
+ if (mas->sheaf) {
+ mt_return_sheaf(mas->sheaf);
+ mas->sheaf = NULL;
+ }
+
+ if (mas->alloc) {
+ kfree(mas->alloc);
+ mas->alloc = NULL;
+ }
+}
+
+/*
+ * mas_free() - Free an encoded maple node
+ * @mas: The maple state
+ * @used: The encoded maple node to free.
+ *
+ * Uses rcu free if necessary, pushes @used back on the maple state allocations
+ * otherwise.
+ */
+static inline void mas_free(struct ma_state *mas, struct maple_enode *used)
+{
+ ma_free_rcu(mte_to_node(used));
+}
+
+
+#ifdef CONFIG_LOCKDEP
+static struct lockdep_map *mas_lockdep_map(struct ma_state *mas)
+{
+ struct maple_tree *mt = mas->tree;
+
+ if (mt_external_lock(mt))
+ return mt->ma_external_lock;
+
+ return &(mt->ma_lock).dep_map;
+}
+
+#endif
+
+static void mas_lock_check(struct ma_state *mas)
+{
+#ifdef CONFIG_LOCKDEP
+ struct lockdep_map *map;
+ u32 seq;
+
+ if (!mas_is_active(mas))
+ return;
+
+#ifdef CONFIG_RCU_STRICT_GRACE_PERIOD
+ if (!mt_locked(mas->tree)) {
+ if (mt_in_rcu(mas->tree))
+ WARN_ON_ONCE(poll_state_synchronize_rcu(mas->rcu_gp));
+ }
+#endif /* CONFIG_RCU_STRICT_GRACE_PERIOD */
+
+ map = mas_lockdep_map(mas);
+ if (!map)
+ return;
+
+ seq = lock_sequence(map);
+ if (seq != UINT_MAX && mas->ld_seq != UINT_MAX)
+ WARN_ON_ONCE(mas->ld_seq != seq);
+#endif /* CONFIG_LOCKDEP */
+
+}
+
+static void mas_init_lock_check(struct ma_state *mas)
+{
+#ifdef CONFIG_LOCKDEP
+ struct lockdep_map *map;
+#ifdef CONFIG_RCU_STRICT_GRACE_PERIOD
+ if (!mt_locked(mas->tree)) {
+ if (mt_in_rcu(mas->tree))
+ mas->rcu_gp = get_state_synchronize_rcu();
+ return;
+ }
+#endif /* CONFIG_RCU_STRICT_GRACE_PERIOD */
+
+ map = mas_lockdep_map(mas);
+ if (map) /* Update regardless of lock state */
+ mas->ld_seq = lock_sequence(map);
+#endif /* CONFIG_LOCKDEP */
+
+}
+
+static void mas_may_init_lock_check(struct ma_state *mas)
+{
+#ifdef CONFIG_LOCKDEP
+#ifdef CONFIG_RCU_STRICT_GRACE_PERIOD
+ if (mas_is_start(mas) || mas_is_paused(mas)) {
+ mas_init_lock_check(mas);
+ return;
+ }
+#endif /* CONFIG_RCU_STRICT_GRACE_PERIOD */
+ mas_lock_check(mas);
+#endif /* CONFIG_LOCKDEP */
+}
+
+/*
+ * mas_start() - Sets up maple state for operations.
+ * @mas: The maple state.
+ *
+ * If mas->status == ma_start, then set the min, max and depth to
+ * defaults.
+ *
+ * Return:
+ * - If mas->node is an error or not mas_start, return NULL.
+ * - If it's an empty tree: NULL & mas->status == ma_none
+ * - If it's a single entry: The entry & mas->status == ma_root
+ * - If it's a tree: NULL & mas->status == ma_active
+ */
+static inline struct maple_enode *mas_start(struct ma_state *mas)
+{
+ if (likely(mas_is_start(mas))) {
+ struct maple_enode *root;
+
+ mas_init_lock_check(mas);
+ mas->min = 0;
+ mas->max = ULONG_MAX;
+
+retry:
+ mas->depth = 0;
+ root = mas_root(mas);
+ /* Tree with nodes */
+ if (likely(xa_is_node(root))) {
+ mas->depth = 0;
+ mas->status = ma_active;
+ mas->node = mte_safe_root(root);
+ mas->offset = 0;
+ if (mte_dead_node(mas->node))
+ goto retry;
+
+ return NULL;
+ }
+
+ mas->node = NULL;
+ /* empty tree */
+ if (unlikely(!root)) {
+ mas->status = ma_none;
+ mas->offset = MAPLE_NODE_SLOTS;
+ return NULL;
+ }
+
+ /* Single entry tree */
+ mas->status = ma_root;
+ mas->offset = MAPLE_NODE_SLOTS;
+
+ /* Single entry tree. */
+ if (mas->index > 0)
+ return NULL;
+
+ return root;
+ }
+
+ return NULL;
+}
+
+/*
+ * ma_data_end() - Find the end of the data in a node.
+ * @node: The maple node
+ * @type: The maple node type
+ * @pivots: The array of pivots in the node
+ * @max: The maximum value in the node
+ *
+ * Uses metadata to find the end of the data when possible.
+ * Return: The zero indexed last slot with data (may be null).
+ */
+static __always_inline unsigned char ma_data_end(struct maple_node *node,
+ enum maple_type type, unsigned long *pivots, unsigned long max)
+{
+ unsigned char offset;
+
+ if (!pivots)
+ return 0;
+
+ if (type == maple_arange_64)
+ return ma_meta_end(node, type);
+
+ offset = mt_pivots[type] - 1;
+ if (likely(!pivots[offset]))
+ return ma_meta_end(node, type);
+
+ if (likely(pivots[offset] == max))
+ return offset;
+
+ return mt_pivots[type];
+}
+
+/*
+ * mas_data_end() - Find the end of the data (slot).
+ * @mas: the maple state
+ *
+ * This method is optimized to check the metadata of a node if the node type
+ * supports data end metadata.
+ *
+ * Return: The zero indexed last slot with data (may be null).
+ */
+static inline unsigned char mas_data_end(struct ma_state *mas)
+{
+ enum maple_type type;
+ struct maple_node *node;
+ unsigned char offset;
+ unsigned long *pivots;
+
+ type = mte_node_type(mas->node);
+ node = mas_mn(mas);
+ if (type == maple_arange_64)
+ return ma_meta_end(node, type);
+
+ pivots = ma_pivots(node, type);
+ if (unlikely(ma_dead_node(node)))
+ return 0;
+
+ offset = mt_pivots[type] - 1;
+ if (likely(!pivots[offset]))
+ return ma_meta_end(node, type);
+
+ if (likely(pivots[offset] == mas->max))
+ return offset;
+
+ return mt_pivots[type];
+}
+
+static inline
+void wr_mas_setup(struct ma_wr_state *wr_mas, struct ma_state *mas)
+{
+ wr_mas->node = mas_mn(mas);
+ wr_mas->type = mte_node_type(mas->node);
+ wr_mas->pivots = ma_pivots(wr_mas->node, wr_mas->type);
+ wr_mas->slots = ma_slots(wr_mas->node, wr_mas->type);
+ wr_mas->r_min = mas_safe_min(mas, wr_mas->pivots, mas->offset);
+ wr_mas->r_max = mas_safe_pivot(mas, wr_mas->pivots, mas->offset,
+ wr_mas->type);
+}
+
+static inline
+void wr_mas_ascend(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+
+ mas_ascend(mas);
+ wr_mas_setup(wr_mas, mas);
+ mas->end = ma_data_end(wr_mas->node, wr_mas->type, wr_mas->pivots,
+ mas->max);
+ /* Careful, this may be wrong.. */
+ wr_mas->end_piv = wr_mas->r_max;
+ wr_mas->offset_end = mas->offset;
+}
+
+static inline unsigned long ma_leaf_max_gap(struct maple_node *mn,
+ enum maple_type mt, unsigned long min, unsigned long max,
+ unsigned long *pivots, void __rcu **slots)
+{
+ unsigned long pstart, gap, max_gap;
+ unsigned char i;
+ unsigned char max_piv;
+
+ max_gap = 0;
+ if (unlikely(ma_is_dense(mt))) {
+ gap = 0;
+ for (i = 0; i < mt_slots[mt]; i++) {
+ if (slots[i]) {
+ if (gap > max_gap)
+ max_gap = gap;
+ gap = 0;
+ } else {
+ gap++;
+ }
+ }
+ if (gap > max_gap)
+ max_gap = gap;
+ return max_gap;
+ }
+
+ /*
+ * Check the first implied pivot optimizes the loop below and slot 1 may
+ * be skipped if there is a gap in slot 0.
+ */
+ if (likely(!slots[0])) {
+ max_gap = pivots[0] - min + 1;
+ i = 2;
+ } else {
+ i = 1;
+ }
+
+ /* reduce max_piv as the special case is checked before the loop */
+ max_piv = ma_data_end(mn, mt, pivots, max) - 1;
+ /*
+ * Check end implied pivot which can only be a gap on the right most
+ * node.
+ */
+ if (unlikely(max == ULONG_MAX) && !slots[max_piv + 1]) {
+ gap = ULONG_MAX - pivots[max_piv];
+ if (gap > max_gap)
+ max_gap = gap;
+
+ if (max_gap > pivots[max_piv] - min)
+ return max_gap;
+ }
+
+ for (; i <= max_piv; i++) {
+ /* data == no gap. */
+ if (likely(slots[i]))
+ continue;
+
+ pstart = pivots[i - 1];
+ gap = pivots[i] - pstart;
+ if (gap > max_gap)
+ max_gap = gap;
+
+ /* There cannot be two gaps in a row. */
+ i++;
+ }
+ return max_gap;
+}
+
+/*
+ * mas_leaf_max_gap() - Returns the largest gap in a leaf node
+ * @mas: the maple state
+ *
+ * Return: The maximum gap in the leaf.
+ */
+static inline unsigned long mas_leaf_max_gap(struct ma_state *mas)
+{
+ enum maple_type mt;
+ struct maple_node *mn;
+ unsigned long *pivots;
+ void __rcu **slots;
+
+ mn = mas_mn(mas);
+ mt = mte_node_type(mas->node);
+ slots = ma_slots(mn, mt);
+ pivots = ma_pivots(mn, mt);
+
+ return ma_leaf_max_gap(mn, mt, mas->min, mas->max, pivots, slots);
+}
+
+/*
+ * ma_max_gap() - Get the maximum gap in a maple node (non-leaf)
+ * @node: The maple node
+ * @gaps: The pointer to the gaps
+ * @mt: The maple node type
+ * @off: Pointer to store the offset location of the gap.
+ *
+ * Uses the metadata data end to scan backwards across set gaps.
+ *
+ * Return: The maximum gap value
+ */
+static inline unsigned long
+ma_max_gap(struct maple_node *node, unsigned long *gaps, enum maple_type mt,
+ unsigned char *off)
+{
+ unsigned char offset, i;
+ unsigned long max_gap = 0;
+
+ i = offset = ma_meta_end(node, mt);
+ do {
+ if (gaps[i] > max_gap) {
+ max_gap = gaps[i];
+ offset = i;
+ }
+ } while (i--);
+
+ *off = offset;
+ return max_gap;
+}
+
+/*
+ * mas_max_gap() - find the largest gap in a non-leaf node and set the slot.
+ * @mas: The maple state.
+ *
+ * Return: The gap value.
+ */
+static inline unsigned long mas_max_gap(struct ma_state *mas)
+{
+ unsigned long *gaps;
+ unsigned char offset;
+ enum maple_type mt;
+ struct maple_node *node;
+
+ mt = mte_node_type(mas->node);
+ if (ma_is_leaf(mt))
+ return mas_leaf_max_gap(mas);
+
+ node = mas_mn(mas);
+ MAS_BUG_ON(mas, mt != maple_arange_64);
+ offset = ma_meta_gap(node);
+ gaps = ma_gaps(node, mt);
+ return gaps[offset];
+}
+
+/*
+ * mas_parent_gap() - Set the parent gap and any gaps above, as needed
+ * @mas: The maple state
+ * @offset: The gap offset in the parent to set
+ * @new: The new gap value.
+ *
+ * Set the parent gap then continue to set the gap upwards, using the metadata
+ * of the parent to see if it is necessary to check the node above.
+ */
+static inline void mas_parent_gap(struct ma_state *mas, unsigned char offset,
+ unsigned long new)
+{
+ unsigned long meta_gap = 0;
+ struct maple_node *pnode;
+ struct maple_enode *penode;
+ unsigned long *pgaps;
+ unsigned char meta_offset;
+ enum maple_type pmt;
+
+ pnode = mte_parent(mas->node);
+ pmt = mas_parent_type(mas, mas->node);
+ penode = mt_mk_node(pnode, pmt);
+ pgaps = ma_gaps(pnode, pmt);
+
+ascend:
+ MAS_BUG_ON(mas, pmt != maple_arange_64);
+ meta_offset = ma_meta_gap(pnode);
+ meta_gap = pgaps[meta_offset];
+
+ pgaps[offset] = new;
+
+ if (meta_gap == new)
+ return;
+
+ if (offset != meta_offset) {
+ if (meta_gap > new)
+ return;
+
+ ma_set_meta_gap(pnode, pmt, offset);
+ } else if (new < meta_gap) {
+ new = ma_max_gap(pnode, pgaps, pmt, &meta_offset);
+ ma_set_meta_gap(pnode, pmt, meta_offset);
+ }
+
+ if (ma_is_root(pnode))
+ return;
+
+ /* Go to the parent node. */
+ pnode = mte_parent(penode);
+ pmt = mas_parent_type(mas, penode);
+ pgaps = ma_gaps(pnode, pmt);
+ offset = mte_parent_slot(penode);
+ penode = mt_mk_node(pnode, pmt);
+ goto ascend;
+}
+
+static __always_inline void mas_update_gap_known(struct ma_state *mas,
+ unsigned long gap)
+{
+ unsigned char pslot;
+ unsigned long p_gap;
+
+ pslot = mte_parent_slot(mas->node);
+ p_gap = ma_gaps(mte_parent(mas->node),
+ mas_parent_type(mas, mas->node))[pslot];
+
+ if (p_gap != gap)
+ mas_parent_gap(mas, pslot, gap);
+}
+
+/*
+ * mas_update_gap() - Update a nodes gaps and propagate up if necessary.
+ * @mas: the maple state.
+ */
+static inline void mas_update_gap(struct ma_state *mas)
+{
+ unsigned long max_gap;
+
+ if (!mt_is_alloc(mas->tree))
+ return;
+
+ if (mte_is_root(mas->node))
+ return;
+
+ max_gap = mas_max_gap(mas);
+ mas_update_gap_known(mas, max_gap);
+}
+
+/*
+ * mas_adopt_children() - Set the parent pointer of all nodes in @parent to
+ * @parent with the slot encoded.
+ * @mas: the maple state (for the tree)
+ * @parent: the maple encoded node containing the children.
+ */
+static inline void mas_adopt_children(struct ma_state *mas,
+ struct maple_enode *parent)
+{
+ enum maple_type type = mte_node_type(parent);
+ struct maple_node *node = mte_to_node(parent);
+ void __rcu **slots = ma_slots(node, type);
+ unsigned long *pivots = ma_pivots(node, type);
+ unsigned char end;
+
+ end = ma_data_end(node, type, pivots, mas->max);
+ mas_set_parent_slots(mas, parent, slots, 0, end + 1);
+}
+
+/*
+ * mas_put_in_tree() - Put a new node in the tree, smp_wmb(), and mark the old
+ * node as dead.
+ * @mas: the maple state with the new node
+ * @old_enode: The old maple encoded node to replace.
+ * @new_height: if we are inserting a root node, update the height of the tree
+ */
+static inline void mas_put_in_tree(struct ma_state *mas,
+ struct maple_enode *old_enode, char new_height)
+ __must_hold(mas->tree->ma_lock)
+{
+ unsigned char offset;
+ void __rcu **slots;
+
+ if (mte_is_root(mas->node)) {
+ mas_mn(mas)->parent = ma_parent_ptr(mas_tree_parent(mas));
+ rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node));
+ mt_set_height(mas->tree, new_height);
+ } else {
+
+ offset = mte_parent_slot(mas->node);
+ slots = ma_slots(mte_parent(mas->node),
+ mas_parent_type(mas, mas->node));
+ rcu_assign_pointer(slots[offset], mas->node);
+ }
+
+ mte_set_node_dead(old_enode);
+}
+
+/*
+ * mas_replace_node() - Replace a node by putting it in the tree, marking it
+ * dead, and freeing it.
+ * the parent encoding to locate the maple node in the tree.
+ * @mas: the ma_state with @mas->node pointing to the new node.
+ * @old_enode: The old maple encoded node.
+ * @new_height: The new height of the tree as a result of the operation
+ */
+static inline void mas_replace_node(struct ma_state *mas,
+ struct maple_enode *old_enode, unsigned char new_height)
+ __must_hold(mas->tree->ma_lock)
+{
+ mas_put_in_tree(mas, old_enode, new_height);
+ mas_free(mas, old_enode);
+}
+
+/*
+ * mas_find_child() - Find a child who has the parent @mas->node.
+ * @mas: the maple state with the parent.
+ * @child: the maple state to store the child.
+ */
+static inline bool mas_find_child(struct ma_state *mas, struct ma_state *child)
+ __must_hold(mas->tree->ma_lock)
+{
+ enum maple_type mt;
+ unsigned char offset;
+ unsigned char end;
+ unsigned long *pivots;
+ struct maple_enode *entry;
+ struct maple_node *node;
+ void __rcu **slots;
+
+ mt = mte_node_type(mas->node);
+ node = mas_mn(mas);
+ slots = ma_slots(node, mt);
+ pivots = ma_pivots(node, mt);
+ end = ma_data_end(node, mt, pivots, mas->max);
+ for (offset = mas->offset; offset <= end; offset++) {
+ entry = mas_slot_locked(mas, slots, offset);
+ if (mte_parent(entry) == node) {
+ *child = *mas;
+ mas->offset = offset + 1;
+ child->offset = offset;
+ mas_descend(child);
+ child->offset = 0;
+ return true;
+ }
+ }
+ return false;
+}
+
+/*
+ * mas_leaf_set_meta() - Set the metadata of a leaf if possible.
+ * @node: The maple node
+ * @mt: The maple type
+ * @end: The node end
+ */
+static inline void mas_leaf_set_meta(struct maple_node *node,
+ enum maple_type mt, unsigned char end)
+{
+ if (end < mt_slots[mt] - 1)
+ ma_set_meta(node, mt, 0, end);
+}
+
+/*
+ * mas_prev_sibling() - Find the previous node with the same parent.
+ * @mas: the maple state
+ *
+ * Return: True if there is a previous sibling, false otherwise.
+ */
+static inline bool mas_prev_sibling(struct ma_state *mas)
+{
+ unsigned int p_slot = mte_parent_slot(mas->node);
+
+ /* For root node, p_slot is set to 0 by mte_parent_slot(). */
+ if (!p_slot)
+ return false;
+
+ mas_ascend(mas);
+ mas->offset = p_slot - 1;
+ mas_descend(mas);
+ return true;
+}
+
+/*
+ * mas_next_sibling() - Find the next node with the same parent.
+ * @mas: the maple state
+ *
+ * Return: true if there is a next sibling, false otherwise.
+ */
+static inline bool mas_next_sibling(struct ma_state *mas)
+{
+ MA_STATE(parent, mas->tree, mas->index, mas->last);
+
+ if (mte_is_root(mas->node))
+ return false;
+
+ parent = *mas;
+ mas_ascend(&parent);
+ parent.offset = mte_parent_slot(mas->node) + 1;
+ if (parent.offset > mas_data_end(&parent))
+ return false;
+
+ *mas = parent;
+ mas_descend(mas);
+ return true;
+}
+
+/*
+ * mas_wr_node_walk() - Find the correct offset for the index in the @mas.
+ * If @mas->index cannot be found within the containing
+ * node, we traverse to the last entry in the node.
+ * @wr_mas: The maple write state
+ *
+ * Uses mas_slot_locked() and does not need to worry about dead nodes.
+ */
+static inline void mas_wr_node_walk(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+ unsigned char count, offset;
+
+ if (unlikely(ma_is_dense(wr_mas->type))) {
+ wr_mas->r_max = wr_mas->r_min = mas->index;
+ mas->offset = mas->index = mas->min;
+ return;
+ }
+
+ wr_mas->node = mas_mn(wr_mas->mas);
+ wr_mas->pivots = ma_pivots(wr_mas->node, wr_mas->type);
+ count = mas->end = ma_data_end(wr_mas->node, wr_mas->type,
+ wr_mas->pivots, mas->max);
+ offset = mas->offset;
+
+ while (offset < count && mas->index > wr_mas->pivots[offset])
+ offset++;
+
+ wr_mas->r_max = offset < count ? wr_mas->pivots[offset] : mas->max;
+ wr_mas->r_min = mas_safe_min(mas, wr_mas->pivots, offset);
+ wr_mas->offset_end = mas->offset = offset;
+}
+
+static inline void rebalance_sib(struct ma_state *parent, struct ma_state *sib)
+{
+ *sib = *parent;
+ /* Prioritize move right to pull data left */
+ if (sib->offset < sib->end)
+ sib->offset++;
+ else
+ sib->offset--;
+
+ mas_descend(sib);
+ sib->end = mas_data_end(sib);
+}
+
+static inline
+void spanning_sib(struct ma_wr_state *l_wr_mas,
+ struct ma_wr_state *r_wr_mas, struct ma_state *nneighbour)
+{
+ struct ma_state l_tmp = *l_wr_mas->mas;
+ struct ma_state r_tmp = *r_wr_mas->mas;
+ unsigned char depth = 0;
+
+ do {
+ mas_ascend(&r_tmp);
+ mas_ascend(&l_tmp);
+ depth++;
+ if (r_tmp.offset < mas_data_end(&r_tmp)) {
+ r_tmp.offset++;
+ mas_descend(&r_tmp);
+ r_tmp.offset = 0;
+ while (--depth)
+ mas_descend(&r_tmp);
+
+ r_tmp.end = mas_data_end(&r_tmp);
+ *nneighbour = r_tmp;
+ return;
+ } else if (l_tmp.offset) {
+ l_tmp.offset--;
+ do {
+ mas_descend(&l_tmp);
+ l_tmp.offset = mas_data_end(&l_tmp);
+ } while (--depth);
+
+ l_tmp.end = l_tmp.offset;
+ *nneighbour = l_tmp;
+ return;
+ }
+ } while (!mte_is_root(r_tmp.node));
+
+ WARN_ON_ONCE(1);
+}
+
+/*
+ * mas_topiary_node() - Dispose of a single node
+ * @mas: The maple state for pushing nodes
+ * @in_rcu: If the tree is in rcu mode
+ *
+ * The node will either be RCU freed or pushed back on the maple state.
+ */
+static inline void mas_topiary_node(struct ma_state *mas,
+ struct ma_state *tmp_mas, bool in_rcu)
+{
+ struct maple_node *tmp;
+ struct maple_enode *enode;
+
+ if (mas_is_none(tmp_mas))
+ return;
+
+ enode = tmp_mas->node;
+ tmp = mte_to_node(enode);
+ mte_set_node_dead(enode);
+ ma_free_rcu(tmp);
+}
+
+/*
+ * mas_topiary_replace() - Replace the data with new data, then repair the
+ * parent links within the new tree. Iterate over the dead sub-tree and collect
+ * the dead subtrees and topiary the nodes that are no longer of use.
+ *
+ * The new tree will have up to three children with the correct parent. Keep
+ * track of the new entries as they need to be followed to find the next level
+ * of new entries.
+ *
+ * The old tree will have up to three children with the old parent. Keep track
+ * of the old entries as they may have more nodes below replaced. Nodes within
+ * [index, last] are dead subtrees, others need to be freed and followed.
+ *
+ * @mas: The maple state pointing at the new data
+ * @old_enode: The maple encoded node being replaced
+ * @new_height: The new height of the tree as a result of the operation
+ *
+ */
+static inline void mas_topiary_replace(struct ma_state *mas,
+ struct maple_enode *old_enode, unsigned char new_height)
+{
+ struct ma_state tmp[3], tmp_next[3];
+ MA_TOPIARY(subtrees, mas->tree);
+ bool in_rcu;
+ int i, n;
+
+ /* Place data in tree & then mark node as old */
+ mas_put_in_tree(mas, old_enode, new_height);
+
+ /* Update the parent pointers in the tree */
+ tmp[0] = *mas;
+ tmp[0].offset = 0;
+ tmp[1].status = ma_none;
+ tmp[2].status = ma_none;
+ while (!mte_is_leaf(tmp[0].node)) {
+ n = 0;
+ for (i = 0; i < 3; i++) {
+ if (mas_is_none(&tmp[i]))
+ continue;
+
+ while (n < 3) {
+ if (!mas_find_child(&tmp[i], &tmp_next[n]))
+ break;
+ n++;
+ }
+
+ mas_adopt_children(&tmp[i], tmp[i].node);
+ }
+
+ if (MAS_WARN_ON(mas, n == 0))
+ break;
+
+ while (n < 3)
+ tmp_next[n++].status = ma_none;
+
+ for (i = 0; i < 3; i++)
+ tmp[i] = tmp_next[i];
+ }
+
+ /* Collect the old nodes that need to be discarded */
+ if (mte_is_leaf(old_enode))
+ return mas_free(mas, old_enode);
+
+ tmp[0] = *mas;
+ tmp[0].offset = 0;
+ tmp[0].node = old_enode;
+ tmp[1].status = ma_none;
+ tmp[2].status = ma_none;
+ in_rcu = mt_in_rcu(mas->tree);
+ do {
+ n = 0;
+ for (i = 0; i < 3; i++) {
+ if (mas_is_none(&tmp[i]))
+ continue;
+
+ while (n < 3) {
+ if (!mas_find_child(&tmp[i], &tmp_next[n]))
+ break;
+
+ if ((tmp_next[n].min >= tmp_next->index) &&
+ (tmp_next[n].max <= tmp_next->last)) {
+ mat_add(&subtrees, tmp_next[n].node);
+ tmp_next[n].status = ma_none;
+ } else {
+ n++;
+ }
+ }
+ }
+
+ if (MAS_WARN_ON(mas, n == 0))
+ break;
+
+ while (n < 3)
+ tmp_next[n++].status = ma_none;
+
+ for (i = 0; i < 3; i++) {
+ mas_topiary_node(mas, &tmp[i], in_rcu);
+ tmp[i] = tmp_next[i];
+ }
+ } while (!mte_is_leaf(tmp[0].node));
+
+ for (i = 0; i < 3; i++)
+ mas_topiary_node(mas, &tmp[i], in_rcu);
+
+ mas_mat_destroy(mas, &subtrees);
+}
+
+/*
+ * node_copy() - Copy from one node to another.
+ *
+ * @mas: The maple state
+ * @src: The source node
+ * @start: The offset into the src to start copying
+ * @size: The size to copy (non-zero)
+ * @s_max: The source node max
+ * @s_mt: The source maple node type
+ * @dst: The destination
+ * @d_start: The start location in the destination node
+ * @d_mt: The destination maple node type
+ */
+static inline
+unsigned long node_copy(struct ma_state *mas, struct maple_node *src,
+ unsigned char start, unsigned char size, unsigned long s_max,
+ enum maple_type s_mt, struct maple_node *dst, unsigned char d_start,
+ enum maple_type d_mt)
+{
+ unsigned long *s_pivots, *d_pivots;
+ void __rcu **s_slots, **d_slots;
+ unsigned long *s_gaps, *d_gaps;
+ unsigned long d_max;
+
+ d_slots = ma_slots(dst, d_mt) + d_start;
+ d_pivots = ma_pivots(dst, d_mt) + d_start;
+ s_slots = ma_slots(src, s_mt) + start;
+ s_pivots = ma_pivots(src, s_mt) + start;
+ memcpy(d_slots, s_slots, size * sizeof(void __rcu *));
+
+ if (!ma_is_leaf(d_mt) && s_mt == maple_copy)
+ mas_set_parent_slots(mas, mt_mk_node(dst, d_mt),
+ d_slots, d_start, size);
+
+ d_gaps = ma_gaps(dst, d_mt);
+ if (d_gaps) {
+ s_gaps = ma_gaps(src, s_mt) + start;
+ d_gaps += d_start;
+ memcpy(d_gaps, s_gaps, size * sizeof(unsigned long));
+ }
+
+ if (start + size - 1 < mt_pivots[s_mt])
+ d_max = s_pivots[size - 1];
+ else
+ d_max = s_max;
+
+ if (d_start + size <= mt_pivots[d_mt])
+ d_pivots[size - 1] = d_max;
+
+ size--;
+ if (size)
+ memcpy(d_pivots, s_pivots, size * sizeof(unsigned long));
+
+ return d_max;
+}
+
+/*
+ * node_finalise() - Zero out unused area and populate metadata
+ * @node: The maple node
+ * @mt: The maple node type
+ * @end: The end of the used area
+ */
+static inline
+void node_finalise(struct maple_node *node, enum maple_type mt,
+ unsigned char end)
+{
+ unsigned char max_end = mt_slots[mt];
+ unsigned char size;
+ unsigned long *gaps;
+ unsigned char gap_slot;
+
+ gaps = ma_gaps(node, mt);
+ if (end < max_end - 1) {
+ size = max_end - end;
+ memset(ma_slots(node, mt) + end, 0, size * sizeof(void *));
+
+ if (gaps)
+ memset(gaps + end, 0, size * sizeof(unsigned long));
+
+ if (--size)
+ memset(ma_pivots(node, mt) + end, 0, size * sizeof(unsigned long));
+ }
+
+ gap_slot = 0;
+ if (gaps && !ma_is_leaf(mt)) {
+ unsigned long max_gap;
+
+ max_gap = 0;
+ for (int i = 0; i <= end; i++)
+ if (gaps[i] > max_gap) {
+ gap_slot = i;
+ max_gap = gaps[i];
+ }
+ }
+
+ if (mt == maple_arange_64)
+ ma_set_meta(node, mt, gap_slot, end - 1);
+ else if (end <= max_end - 1)
+ ma_set_meta(node, mt, gap_slot, end - 1);
+}
+
+static inline void *mtree_range_walk(struct ma_state *mas)
+{
+ unsigned long *pivots;
+ unsigned char offset;
+ struct maple_node *node;
+ struct maple_enode *next, *last;
+ enum maple_type type;
+ void __rcu **slots;
+ unsigned char end;
+ unsigned long max, min;
+ unsigned long prev_max, prev_min;
+
+ next = mas->node;
+ min = mas->min;
+ max = mas->max;
+ do {
+ last = next;
+ node = mte_to_node(next);
+ type = mte_node_type(next);
+ pivots = ma_pivots(node, type);
+ end = ma_data_end(node, type, pivots, max);
+ prev_min = min;
+ prev_max = max;
+ if (pivots[0] >= mas->index) {
+ offset = 0;
+ max = pivots[0];
+ goto next;
+ }
+
+ offset = 1;
+ while (offset < end) {
+ if (pivots[offset] >= mas->index) {
+ max = pivots[offset];
+ break;
+ }
+ offset++;
+ }
+
+ min = pivots[offset - 1] + 1;
+next:
+ slots = ma_slots(node, type);
+ next = mt_slot(mas->tree, slots, offset);
+ if (unlikely(ma_dead_node(node)))
+ goto dead_node;
+ } while (!ma_is_leaf(type));
+
+ mas->end = end;
+ mas->offset = offset;
+ mas->index = min;
+ mas->last = max;
+ mas->min = prev_min;
+ mas->max = prev_max;
+ mas->node = last;
+ return (void *)next;
+
+dead_node:
+ mas_reset(mas);
+ return NULL;
+}
+
+/*
+ * mas_wmb_replace() - Write memory barrier and replace
+ * @mas: The maple state
+ * @cp: The maple copy node
+ *
+ * Updates gap as necessary.
+ */
+static inline void mas_wmb_replace(struct ma_state *mas, struct maple_copy *cp)
+{
+ struct maple_enode *old_enode;
+
+ old_enode = mas->node;
+ mas->node = mt_slot_locked(mas->tree, cp->slot, 0);
+ /* Insert the new data in the tree */
+ mas_topiary_replace(mas, old_enode, cp->height);
+ if (mt_is_alloc(mas->tree) && !mte_is_root(mas->node))
+ mas_update_gap_known(mas, cp->gap[0]);
+
+ mtree_range_walk(mas);
+}
+
+
+/*
+ * cp_leaf_init() - Initialize a maple_copy node for the leaf level of a
+ * spanning store
+ * @cp: The maple copy node
+ * @mas: The maple state
+ * @l_wr_mas: The left write state of the spanning store
+ * @r_wr_mas: The right write state of the spanning store
+ */
+static inline void cp_leaf_init(struct maple_copy *cp,
+ struct ma_state *mas, struct ma_wr_state *l_wr_mas,
+ struct ma_wr_state *r_wr_mas)
+{
+ unsigned char end = 0;
+
+ /*
+ * WARNING: The use of RCU_INIT_POINTER() makes it extremely important
+ * to not expose the maple_copy node to any readers. Exposure may
+ * result in buggy code when a compiler reorders the instructions.
+ */
+
+ cp->height = 1;
+ /* Create entries to insert including split entries to left and right */
+ if (l_wr_mas->r_min < mas->index) {
+ end++;
+ RCU_INIT_POINTER(cp->slot[0], l_wr_mas->content);
+ cp->pivot[0] = mas->index - 1;
+ }
+ RCU_INIT_POINTER(cp->slot[end], l_wr_mas->entry);
+ cp->pivot[end] = mas->last;
+
+ if (r_wr_mas->end_piv > mas->last) {
+ end++;
+ RCU_INIT_POINTER(cp->slot[end],
+ r_wr_mas->slots[r_wr_mas->offset_end]);
+ cp->pivot[end] = r_wr_mas->end_piv;
+ }
+
+ cp->min = l_wr_mas->r_min;
+ cp->max = cp->pivot[end];
+ cp->end = end;
+}
+
+/*
+ * cp_data_calc() - Calculate the size of the data (1 indexed).
+ * @cp: The maple copy struct with the new data populated.
+ * @l_wr_mas: The maple write state containing the data to the left of the write
+ * @r_wr_mas: The maple write state containing the data to the right of the
+ * write
+ *
+ * cp->data is a size (not indexed by 0).
+ */
+static inline void cp_data_calc(struct maple_copy *cp,
+ struct ma_wr_state *l_wr_mas, struct ma_wr_state *r_wr_mas)
+{
+
+ /* Add 1 every time for the 0th element */
+ cp->data = l_wr_mas->mas->offset;
+ /* Add the new data and any partial overwrites */
+ cp->data += cp->end + 1;
+ /* Data from right (offset + 1 to end), +1 for zero */
+ cp->data += r_wr_mas->mas->end - r_wr_mas->offset_end;
+}
+
+static bool data_fits(struct ma_state *sib, struct ma_state *mas,
+ struct maple_copy *cp)
+{
+ unsigned char new_data;
+ enum maple_type type;
+ unsigned char space;
+ unsigned char end;
+
+ type = mte_node_type(mas->node);
+ space = 2 * mt_slots[type];
+ end = sib->end;
+
+ new_data = end + 1 + cp->data;
+ if (new_data > space)
+ return false;
+
+ /*
+ * This is off by one by design. The extra space is left to reduce
+ * jitter in operations that add then remove two entries.
+ *
+ * end is an index while new space and data are both sizes. Adding one
+ * to end to convert the index to a size means that the below
+ * calculation should be <=, but we want to keep an extra space in nodes
+ * to reduce jitter.
+ *
+ * Note that it is still possible to get a full node on the left by the
+ * NULL landing exactly on the split. The NULL ending of a node happens
+ * in the dst_setup() function, where we will either increase the split
+ * by one or decrease it by one, if possible. In the case of split
+ * (this case), it is always possible to shift the spilt by one - again
+ * because there is at least one slot free by the below checking.
+ */
+ if (new_data < space)
+ return true;
+
+ return false;
+}
+
+static inline void push_data_sib(struct maple_copy *cp, struct ma_state *mas,
+ struct ma_state *sib, struct ma_state *parent)
+{
+
+ if (mte_is_root(mas->node))
+ goto no_push;
+
+
+ *sib = *parent;
+ if (sib->offset) {
+ sib->offset--;
+ mas_descend(sib);
+ sib->end = mas_data_end(sib);
+ if (data_fits(sib, mas, cp)) /* Push left */
+ return;
+
+ *sib = *parent;
+ }
+
+ if (sib->offset >= sib->end)
+ goto no_push;
+
+ sib->offset++;
+ mas_descend(sib);
+ sib->end = mas_data_end(sib);
+ if (data_fits(sib, mas, cp)) /* Push right*/
+ return;
+
+no_push:
+ sib->end = 0;
+}
+
+/*
+ * rebalance_data() - Calculate the @cp data, populate @sib if insufficient or
+ * if the data can be pushed into a sibling.
+ * @cp: The maple copy node
+ * @wr_mas: The left write maple state
+ * @sib: The maple state of the sibling.
+ *
+ * Note: @cp->data is a size and not indexed by 0. @sib->end may be set to 0 to
+ * indicate it will not be used.
+ *
+ */
+static inline void rebalance_data(struct maple_copy *cp,
+ struct ma_wr_state *wr_mas, struct ma_state *sib,
+ struct ma_state *parent)
+{
+ cp_data_calc(cp, wr_mas, wr_mas);
+ sib->end = 0;
+ if (cp->data > mt_slots[wr_mas->type]) {
+ push_data_sib(cp, wr_mas->mas, sib, parent);
+ if (sib->end)
+ goto use_sib;
+ } else if (cp->data <= mt_min_slots[wr_mas->type]) {
+ if ((wr_mas->mas->min != 0) ||
+ (wr_mas->mas->max != ULONG_MAX)) {
+ rebalance_sib(parent, sib);
+ goto use_sib;
+ }
+ }
+
+ return;
+
+use_sib:
+
+ cp->data += sib->end + 1;
+}
+
+/*
+ * spanning_data() - Calculate the @cp data and populate @sib if insufficient
+ * @cp: The maple copy node
+ * @l_wr_mas: The left write maple state
+ * @r_wr_mas: The right write maple state
+ * @sib: The maple state of the sibling.
+ *
+ * Note: @cp->data is a size and not indexed by 0. @sib->end may be set to 0 to
+ * indicate it will not be used.
+ */
+static inline void spanning_data(struct maple_copy *cp,
+ struct ma_wr_state *l_wr_mas, struct ma_wr_state *r_wr_mas,
+ struct ma_state *sib)
+{
+ cp_data_calc(cp, l_wr_mas, r_wr_mas);
+ if (((l_wr_mas->mas->min != 0) || (r_wr_mas->mas->max != ULONG_MAX)) &&
+ (cp->data <= mt_min_slots[l_wr_mas->type])) {
+ spanning_sib(l_wr_mas, r_wr_mas, sib);
+ cp->data += sib->end + 1;
+ } else {
+ sib->end = 0;
+ }
+}
+
+/*
+ * dst_setup() - Set up one or more destinations for the new data.
+ * @cp: The maple copy node
+ * @mas: The maple state
+ * @mt: The source node type
+ */
+static inline
+void dst_setup(struct maple_copy *cp, struct ma_state *mas, enum maple_type mt)
+{
+ /* Data is 1 indexed, every src has +1 added. */
+
+ if (cp->data <= mt_slots[mt]) {
+ cp->split = cp->data - 1;
+ cp->d_count = 1;
+ goto node_setup;
+ }
+
+ cp->split = (cp->data - 1) / 2;
+ cp->d_count = 2;
+ if (cp->data < mt_slots[mt] * 2)
+ goto node_setup;
+
+ if (cp->data == mt_slots[mt] * 2) {
+ unsigned char off;
+ unsigned char s;
+
+ if (!ma_is_leaf(mt))
+ goto node_setup;
+
+ /*
+ * Leaf nodes are a bit tricky because we cannot assume the data
+ * can fit due to the NULL limitation on node ends.
+ */
+ off = cp->split;
+ for (s = 0; s < cp->s_count; s++) {
+ unsigned char s_off;
+
+ s_off = cp->src[s].end - cp->src[s].start;
+ if (s_off >= off)
+ break;
+
+ s_off++;
+ off -= s_off;
+ }
+
+ off += cp->src[s].start;
+ if (ma_slots(cp->src[s].node, cp->src[s].mt)[off])
+ goto node_setup;
+
+ cp->split++;
+ if (cp->split < mt_slots[mt])
+ goto node_setup;
+
+ cp->split -= 2;
+ if (cp->data - 2 - cp->split < mt_slots[mt])
+ goto node_setup;
+
+ }
+
+ /* No other choice but to 3-way split the data */
+ cp->split = (cp->data + 2) / 3;
+ cp->d_count = 3;
+
+node_setup:
+ for (int i = 0; i < cp->d_count; i++) {
+ cp->dst[i].mt = mt;
+ cp->dst[i].node = ma_mnode_ptr(mas_pop_node(mas));
+ }
+}
+
+static inline void append_mas_cp(struct maple_copy *cp,
+ struct ma_state *mas, unsigned char start, unsigned char end)
+{
+ struct maple_node *node;
+ enum maple_type mt;
+ unsigned char count;
+
+ count = cp->s_count;
+ node = mas_mn(mas);
+ mt = mte_node_type(mas->node);
+ cp->src[count].node = node;
+ cp->src[count].mt = mt;
+ if (mas->end <= end)
+ cp->src[count].max = mas->max;
+ else
+ cp->src[count].max = ma_pivots(node, mt)[end];
+
+ cp->src[count].start = start;
+ cp->src[count].end = end;
+ cp->s_count++;
+}
+
+static inline void append_wr_mas_cp(struct maple_copy *cp,
+ struct ma_wr_state *wr_mas, unsigned char start, unsigned char end)
+{
+ unsigned char count;
+
+ count = cp->s_count;
+ cp->src[count].node = wr_mas->node;
+ cp->src[count].mt = wr_mas->type;
+ if (wr_mas->mas->end <= end)
+ cp->src[count].max = wr_mas->mas->max;
+ else
+ cp->src[count].max = wr_mas->pivots[end];
+
+ cp->src[count].start = start;
+ cp->src[count].end = end;
+ cp->s_count++;
+}
+
+static inline void init_cp_src(struct maple_copy *cp)
+{
+ cp->src[cp->s_count].node = ma_mnode_ptr(cp);
+ cp->src[cp->s_count].mt = maple_copy;
+ cp->src[cp->s_count].max = cp->max;
+ cp->src[cp->s_count].start = 0;
+ cp->src[cp->s_count].end = cp->end;
+ cp->s_count++;
+}
+
+/*
+ * multi_src_setup() - Set the @cp node up with multiple sources to copy from.
+ * @cp: The maple copy node
+ * @l_wr_mas: The left write maple state
+ * @r_wr_mas: The right write maple state
+ * @sib: The sibling maple state
+ *
+ * Note: @sib->end == 0 indicates no sibling will be used.
+ */
+static inline
+void multi_src_setup(struct maple_copy *cp, struct ma_wr_state *l_wr_mas,
+ struct ma_wr_state *r_wr_mas, struct ma_state *sib)
+{
+ cp->s_count = 0;
+ if (sib->end && sib->max < l_wr_mas->mas->min)
+ append_mas_cp(cp, sib, 0, sib->end);
+
+ /* Copy left 0 - offset */
+ if (l_wr_mas->mas->offset) {
+ unsigned char off = l_wr_mas->mas->offset - 1;
+
+ append_wr_mas_cp(cp, l_wr_mas, 0, off);
+ cp->src[cp->s_count - 1].max = cp->min - 1;
+ }
+
+ init_cp_src(cp);
+
+ /* Copy right either from offset or offset + 1 pending on r_max */
+ if (r_wr_mas->mas->end != r_wr_mas->offset_end)
+ append_wr_mas_cp(cp, r_wr_mas, r_wr_mas->offset_end + 1,
+ r_wr_mas->mas->end);
+
+ if (sib->end && sib->min > r_wr_mas->mas->max)
+ append_mas_cp(cp, sib, 0, sib->end);
+}
+
+static inline
+void cp_data_write(struct maple_copy *cp, struct ma_state *mas)
+{
+ struct maple_node *dst, *src;
+ unsigned char s, d;
+ unsigned char dst_offset;
+ unsigned char data_offset;
+ unsigned char src_end, s_offset;
+ unsigned char split;
+ unsigned long s_max, d_max;
+ unsigned char dst_size;
+ enum maple_type s_mt, d_mt;
+
+ data_offset = 0;
+ s = d = 0;
+ /* Readability help */
+ src = cp->src[s].node;
+ dst = cp->dst[d].node;
+ s_offset = cp->src[s].start;
+ src_end = cp->src[s].end;
+ split = cp->split;
+ s_max = cp->src[s].max;
+ s_mt = cp->src[s].mt;
+ d_mt = cp->dst[d].mt;
+ do {
+ dst_offset = 0;
+ d_max = 0;
+ dst = cp->dst[d].node;
+ d_mt = cp->dst[d].mt;
+ dst_size = split + 1;
+
+ while (dst_size) {
+ unsigned char size;
+
+ if (src_end - s_offset + 1 < dst_size)
+ size = src_end - s_offset + 1;
+ else
+ size = dst_size;
+
+ d_max = node_copy(mas, src, s_offset, size, s_max, s_mt,
+ dst, dst_offset, d_mt);
+
+ dst_offset += size;
+ s_offset += size;
+ if (s_offset > src_end) {
+ /* This source is exhausted */
+ s++;
+ if (s >= cp->s_count) {
+ cp->dst[d].max = d_max;
+ node_finalise(dst, d_mt, dst_offset);
+ return;
+ }
+ /* Reset local src */
+ src = cp->src[s].node;
+ s_offset = cp->src[s].start;
+ src_end = cp->src[s].end;
+ s_max = cp->src[s].max;
+ s_mt = cp->src[s].mt;
+ }
+
+ dst_size -= size;
+ data_offset += size;
+ }
+
+ split = cp->split;
+ cp->dst[d].max = d_max;
+ /* Handle null entries */
+ if (cp->dst[d].max != ULONG_MAX &&
+ !ma_slots(dst, d_mt)[dst_offset - 1]) {
+ if (s_offset == cp->src[s].start) {
+ s--;
+ src = cp->src[s].node;
+ src_end = cp->src[s].end;
+ s_max = cp->src[s].max;
+ s_mt = cp->src[s].mt;
+ s_offset = src_end;
+ } else {
+ s_offset--;
+ }
+ /* Set dst max and clear pivot */
+ split++;
+ data_offset--;
+ dst_offset--;
+ cp->dst[d].max = ma_pivots(dst, d_mt)[dst_offset - 1];
+ }
+
+ node_finalise(dst, d_mt, dst_offset);
+ ++d; /* Next destination */
+ if (d == cp->d_count - 1)
+ split = cp->data - data_offset;
+
+ if (d >= cp->d_count) {
+ WARN_ON(data_offset < cp->data);
+ return;
+ }
+
+ } while (data_offset <= cp->data);
+}
+
+/*
+ * cp_dst_to_slots() - Migrate the maple copy destination to the maple copy
+ * slots
+ * @cp: The maple copy node
+ * @min: The minimal value represented
+ * @max: The maximum value represented
+ * @mas: The maple state
+ */
+static inline void cp_dst_to_slots(struct maple_copy *cp, unsigned long min,
+ unsigned long max, struct ma_state *mas)
+{
+ unsigned char d;
+ unsigned long slot_min = min;
+
+ for (d = 0; d < cp->d_count; d++) {
+ struct maple_node *mn = cp->dst[d].node;
+ enum maple_type mt = cp->dst[d].mt;
+ unsigned long slot_max = cp->dst[d].max;
+
+ /*
+ * Warning, see cp_leaf_init() comment and rcu_assign_pointer()
+ * documentation. Since these are new nodes, there are no
+ * read-side operations that can view them until they are
+ * inserted into the tree after an rcu_assign_pointer() call.
+ */
+ ma_init_slot(&cp->slot[d], mn, mt);
+ cp->pivot[d] = slot_max;
+ if (mt_is_alloc(mas->tree)) {
+ if (ma_is_leaf(mt)) {
+ cp->gap[d] = ma_leaf_max_gap(mn, mt, slot_min,
+ slot_max, ma_pivots(mn, mt),
+ ma_slots(mn, mt));
+ } else {
+ unsigned long *gaps = ma_gaps(mn, mt);
+
+ if (gaps) {
+ unsigned char gap_slot;
+
+ gap_slot = ma_meta_gap(mn);
+ cp->gap[d] = gaps[gap_slot];
+ }
+ }
+ }
+ slot_min = slot_max + 1;
+ }
+
+ cp->end = cp->d_count - 1;
+ cp->min = min;
+ cp->max = max;
+}
+
+static inline bool cp_is_new_root(struct maple_copy *cp, struct ma_state *mas)
+{
+ if (cp->min || cp->max != ULONG_MAX)
+ return false;
+
+ if (cp->d_count != 1) {
+ enum maple_type mt = maple_arange_64;
+
+ if (!mt_is_alloc(mas->tree))
+ mt = maple_range_64;
+
+ cp->data = cp->d_count;
+ cp->s_count = 0;
+ dst_setup(cp, mas, mt);
+ init_cp_src(cp);
+ node_copy(mas, cp->src[0].node, 0, cp->data, cp->max, maple_copy,
+ cp->dst[0].node, 0, mt);
+ node_finalise(cp->dst[0].node, mt, cp->end + 1);
+ /*
+ * Warning, see cp_leaf_init() comment and rcu_assign_pointer()
+ * documentation. Since this is a new root, there are no
+ * read-side operations that can view it until it is insert into
+ * the tree after an rcu_assign_pointer() call.
+ */
+ ma_init_slot(&cp->slot[0], cp->dst[0].node, mt);
+ cp->height++;
+ }
+ WARN_ON_ONCE(cp->dst[0].node != mte_to_node(
+ mt_slot_locked(mas->tree, cp->slot, 0)));
+ cp->dst[0].node->parent = ma_parent_ptr(mas_tree_parent(mas));
+ mas->min = 0;
+ mas->max = ULONG_MAX;
+ mas->depth = 0;
+ mas->node = mas_root_locked(mas);
+ return true;
+}
+
+static inline bool cp_converged(struct maple_copy *cp, struct ma_state *mas,
+ struct ma_state *sib)
+{
+ if (cp->d_count != 1 || sib->end)
+ return false;
+
+ cp->dst[0].node->parent = ma_parent_ptr(mas_mn(mas)->parent);
+ return true;
+}
+
+/*
+ * spanning_ascend() - See if a spanning store operation has to keep walking up
+ * the tree
+ * @cp: The maple_copy node
+ * @l_wr_mas: The left maple write state
+ * @r_wr_mas: The right maple write state
+ * @sib: the maple state of the sibling
+ *
+ * Returns: True if another iteration is necessary.
+ */
+static bool spanning_ascend(struct maple_copy *cp, struct ma_state *mas,
+ struct ma_wr_state *l_wr_mas, struct ma_wr_state *r_wr_mas,
+ struct ma_state *sib)
+{
+ if (sib->end) {
+ if (sib->max < l_wr_mas->mas->min)
+ *l_wr_mas->mas = *sib;
+ else
+ *r_wr_mas->mas = *sib;
+ }
+
+ cp_dst_to_slots(cp, l_wr_mas->mas->min, r_wr_mas->mas->max, mas);
+ if (cp_is_new_root(cp, mas))
+ return false;
+
+ /* Converged and has a single destination */
+ if ((cp->d_count == 1) &&
+ (l_wr_mas->mas->node == r_wr_mas->mas->node)) {
+ cp->dst[0].node->parent = ma_parent_ptr(mas_mn(mas)->parent);
+ return false;
+ }
+
+ cp->height++;
+ wr_mas_ascend(l_wr_mas);
+ wr_mas_ascend(r_wr_mas);
+ return true;
+}
+
+static inline
+void copy_tree_location(const struct ma_state *src, struct ma_state *dst)
+{
+ dst->node = src->node;
+ dst->offset = src->offset;
+ dst->min = src->min;
+ dst->max = src->max;
+ dst->end = src->end;
+ dst->depth = src->depth;
+}
+
+/*
+ * rebalance_ascend() - Ascend the tree and set up for the next loop - if
+ * necessary
+ *
+ * Return: True if there another rebalancing operation on the next level is
+ * needed, false otherwise.
+ */
+static inline bool rebalance_ascend(struct maple_copy *cp,
+ struct ma_wr_state *wr_mas, struct ma_state *sib,
+ struct ma_state *parent)
+{
+ struct ma_state *mas;
+ unsigned long min, max;
+
+ mas = wr_mas->mas;
+ if (!sib->end) {
+ min = mas->min;
+ max = mas->max;
+ } else if (sib->min > mas->max) { /* Move right succeeded */
+ min = mas->min;
+ max = sib->max;
+ wr_mas->offset_end = parent->offset + 1;
+ } else {
+ min = sib->min;
+ max = mas->max;
+ wr_mas->offset_end = parent->offset;
+ parent->offset--;
+ }
+
+ cp_dst_to_slots(cp, min, max, mas);
+ if (cp_is_new_root(cp, mas))
+ return false;
+
+ if (cp_converged(cp, mas, sib))
+ return false;
+
+ cp->height++;
+ copy_tree_location(parent, mas);
+ wr_mas_setup(wr_mas, mas);
+ return true;
+}
+
+/*
+ * mas_root_expand() - Expand a root to a node
+ * @mas: The maple state
+ * @entry: The entry to store into the tree
+ */
+static inline void mas_root_expand(struct ma_state *mas, void *entry)
+{
+ void *contents = mas_root_locked(mas);
+ enum maple_type type = maple_leaf_64;
+ struct maple_node *node;
+ void __rcu **slots;
+ unsigned long *pivots;
+ int slot = 0;
+
+ node = mas_pop_node(mas);
+ pivots = ma_pivots(node, type);
+ slots = ma_slots(node, type);
+ node->parent = ma_parent_ptr(mas_tree_parent(mas));
+ mas->node = mt_mk_node(node, type);
+ mas->status = ma_active;
+
+ if (mas->index) {
+ if (contents) {
+ rcu_assign_pointer(slots[slot], contents);
+ if (likely(mas->index > 1))
+ slot++;
+ }
+ pivots[slot++] = mas->index - 1;
+ }
+
+ rcu_assign_pointer(slots[slot], entry);
+ mas->offset = slot;
+ pivots[slot] = mas->last;
+ if (mas->last != ULONG_MAX)
+ pivots[++slot] = ULONG_MAX;
+
+ mt_set_height(mas->tree, 1);
+ ma_set_meta(node, maple_leaf_64, 0, slot);
+ /* swap the new root into the tree */
+ rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node));
+}
+
+/*
+ * mas_store_root() - Storing value into root.
+ * @mas: The maple state
+ * @entry: The entry to store.
+ *
+ * There is no root node now and we are storing a value into the root - this
+ * function either assigns the pointer or expands into a node.
+ */
+static inline void mas_store_root(struct ma_state *mas, void *entry)
+{
+ if (!entry) {
+ if (!mas->index)
+ rcu_assign_pointer(mas->tree->ma_root, NULL);
+ } else if (likely((mas->last != 0) || (mas->index != 0)))
+ mas_root_expand(mas, entry);
+ else if (((unsigned long) (entry) & 3) == 2)
+ mas_root_expand(mas, entry);
+ else {
+ rcu_assign_pointer(mas->tree->ma_root, entry);
+ mas->status = ma_start;
+ }
+}
+
+/*
+ * mas_is_span_wr() - Check if the write needs to be treated as a write that
+ * spans the node.
+ * @wr_mas: The maple write state
+ *
+ * Spanning writes are writes that start in one node and end in another OR if
+ * the write of a %NULL will cause the node to end with a %NULL.
+ *
+ * Return: True if this is a spanning write, false otherwise.
+ */
+static bool mas_is_span_wr(struct ma_wr_state *wr_mas)
+{
+ unsigned long max = wr_mas->r_max;
+ unsigned long last = wr_mas->mas->last;
+ enum maple_type type = wr_mas->type;
+ void *entry = wr_mas->entry;
+
+ /* Contained in this pivot, fast path */
+ if (last < max)
+ return false;
+
+ if (ma_is_leaf(type)) {
+ max = wr_mas->mas->max;
+ if (last < max)
+ return false;
+ }
+
+ if (last == max) {
+ /*
+ * The last entry of leaf node cannot be NULL unless it is the
+ * rightmost node (writing ULONG_MAX), otherwise it spans slots.
+ */
+ if (entry || last == ULONG_MAX)
+ return false;
+ }
+
+ trace_ma_write(TP_FCT, wr_mas->mas, wr_mas->r_max, entry);
+ return true;
+}
+
+static inline void mas_wr_walk_descend(struct ma_wr_state *wr_mas)
+{
+ wr_mas->type = mte_node_type(wr_mas->mas->node);
+ mas_wr_node_walk(wr_mas);
+ wr_mas->slots = ma_slots(wr_mas->node, wr_mas->type);
+}
+
+static inline void mas_wr_walk_traverse(struct ma_wr_state *wr_mas)
+{
+ wr_mas->mas->max = wr_mas->r_max;
+ wr_mas->mas->min = wr_mas->r_min;
+ wr_mas->mas->node = wr_mas->content;
+ wr_mas->mas->offset = 0;
+ wr_mas->mas->depth++;
+}
+/*
+ * mas_wr_walk() - Walk the tree for a write.
+ * @wr_mas: The maple write state
+ *
+ * Uses mas_slot_locked() and does not need to worry about dead nodes.
+ *
+ * Return: True if it's contained in a node, false on spanning write.
+ */
+static bool mas_wr_walk(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+
+ while (true) {
+ mas_wr_walk_descend(wr_mas);
+ if (unlikely(mas_is_span_wr(wr_mas)))
+ return false;
+
+ wr_mas->content = mas_slot_locked(mas, wr_mas->slots,
+ mas->offset);
+ if (ma_is_leaf(wr_mas->type))
+ return true;
+
+ if (mas->end < mt_slots[wr_mas->type] - 1)
+ wr_mas->vacant_height = mas->depth + 1;
+
+ if (ma_is_root(mas_mn(mas))) {
+ /* root needs more than 2 entries to be sufficient + 1 */
+ if (mas->end > 2)
+ wr_mas->sufficient_height = 1;
+ } else if (mas->end > mt_min_slots[wr_mas->type] + 1)
+ wr_mas->sufficient_height = mas->depth + 1;
+
+ mas_wr_walk_traverse(wr_mas);
+ }
+
+ return true;
+}
+
+static void mas_wr_walk_index(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+
+ while (true) {
+ mas_wr_walk_descend(wr_mas);
+ wr_mas->content = mas_slot_locked(mas, wr_mas->slots,
+ mas->offset);
+ if (ma_is_leaf(wr_mas->type))
+ return;
+ mas_wr_walk_traverse(wr_mas);
+ }
+}
+/*
+ * mas_extend_spanning_null() - Extend a store of a %NULL to include surrounding %NULLs.
+ * @l_wr_mas: The left maple write state
+ * @r_wr_mas: The right maple write state
+ */
+static inline void mas_extend_spanning_null(struct ma_wr_state *l_wr_mas,
+ struct ma_wr_state *r_wr_mas)
+{
+ struct ma_state *r_mas = r_wr_mas->mas;
+ struct ma_state *l_mas = l_wr_mas->mas;
+ unsigned char l_slot;
+
+ l_slot = l_mas->offset;
+ if (!l_wr_mas->content)
+ l_mas->index = l_wr_mas->r_min;
+
+ if ((l_mas->index == l_wr_mas->r_min) &&
+ (l_slot &&
+ !mas_slot_locked(l_mas, l_wr_mas->slots, l_slot - 1))) {
+ if (l_slot > 1)
+ l_mas->index = l_wr_mas->pivots[l_slot - 2] + 1;
+ else
+ l_mas->index = l_mas->min;
+
+ l_mas->offset = l_slot - 1;
+ l_wr_mas->r_min = l_mas->index;
+ }
+
+ if (!r_wr_mas->content) {
+ if (r_mas->last < r_wr_mas->r_max)
+ r_mas->last = r_wr_mas->r_max;
+ r_mas->offset++;
+ }
+}
+
+static inline void *mas_state_walk(struct ma_state *mas)
+{
+ void *entry;
+
+ entry = mas_start(mas);
+ if (mas_is_none(mas))
+ return NULL;
+
+ if (mas_is_ptr(mas))
+ return entry;
+
+ return mtree_range_walk(mas);
+}
+
+/*
+ * mtree_lookup_walk() - Internal quick lookup that does not keep maple state up
+ * to date.
+ *
+ * @mas: The maple state.
+ *
+ * Note: Leaves mas in undesirable state.
+ * Return: The entry for @mas->index or %NULL on dead node.
+ */
+static inline void *mtree_lookup_walk(struct ma_state *mas)
+{
+ unsigned long *pivots;
+ unsigned char offset;
+ struct maple_node *node;
+ struct maple_enode *next;
+ enum maple_type type;
+ void __rcu **slots;
+ unsigned char end;
+
+ next = mas->node;
+ do {
+ node = mte_to_node(next);
+ type = mte_node_type(next);
+ pivots = ma_pivots(node, type);
+ end = mt_pivots[type];
+ offset = 0;
+ do {
+ if (pivots[offset] >= mas->index)
+ break;
+ } while (++offset < end);
+
+ slots = ma_slots(node, type);
+ next = mt_slot(mas->tree, slots, offset);
+ if (unlikely(ma_dead_node(node)))
+ goto dead_node;
+ } while (!ma_is_leaf(type));
+
+ return (void *)next;
+
+dead_node:
+ mas_reset(mas);
+ return NULL;
+}
+
+static void mte_destroy_walk(struct maple_enode *, struct maple_tree *);
+/*
+ * mas_new_root() - Create a new root node that only contains the entry passed
+ * in.
+ * @mas: The maple state
+ * @entry: The entry to store.
+ *
+ * Only valid when the index == 0 and the last == ULONG_MAX
+ */
+static inline void mas_new_root(struct ma_state *mas, void *entry)
+{
+ struct maple_enode *root = mas_root_locked(mas);
+ enum maple_type type = maple_leaf_64;
+ struct maple_node *node;
+ void __rcu **slots;
+ unsigned long *pivots;
+
+ WARN_ON_ONCE(mas->index || mas->last != ULONG_MAX);
+
+ if (!entry) {
+ mt_set_height(mas->tree, 0);
+ rcu_assign_pointer(mas->tree->ma_root, entry);
+ mas->status = ma_start;
+ goto done;
+ }
+
+ node = mas_pop_node(mas);
+ pivots = ma_pivots(node, type);
+ slots = ma_slots(node, type);
+ node->parent = ma_parent_ptr(mas_tree_parent(mas));
+ mas->node = mt_mk_node(node, type);
+ mas->status = ma_active;
+ rcu_assign_pointer(slots[0], entry);
+ pivots[0] = mas->last;
+ mt_set_height(mas->tree, 1);
+ rcu_assign_pointer(mas->tree->ma_root, mte_mk_root(mas->node));
+
+done:
+ if (xa_is_node(root))
+ mte_destroy_walk(root, mas->tree);
+}
+/*
+ * mas_wr_spanning_store() - Create a subtree with the store operation completed
+ * and new nodes where necessary, then place the sub-tree in the actual tree.
+ * Note that mas is expected to point to the node which caused the store to
+ * span.
+ * @wr_mas: The maple write state
+ */
+static void mas_wr_spanning_store(struct ma_wr_state *wr_mas)
+{
+ struct maple_copy cp;
+ struct ma_state *mas;
+ struct ma_state sib;
+
+ /* Left and Right side of spanning store */
+ MA_STATE(r_mas, NULL, 0, 0);
+ MA_WR_STATE(r_wr_mas, &r_mas, wr_mas->entry);
+
+ /*
+ * A store operation that spans multiple nodes is called a spanning
+ * store and is handled early in the store call stack by the function
+ * mas_is_span_wr(). When a spanning store is identified, the maple
+ * state is duplicated. The first maple state walks the left tree path
+ * to ``index``, the duplicate walks the right tree path to ``last``.
+ * The data in the two nodes are combined into a single node, two nodes,
+ * or possibly three nodes (see the 3-way split above). A ``NULL``
+ * written to the last entry of a node is considered a spanning store as
+ * a rebalance is required for the operation to complete and an overflow
+ * of data may happen.
+ */
+ mas = wr_mas->mas;
+ trace_ma_op(TP_FCT, mas);
+
+ if (unlikely(!mas->index && mas->last == ULONG_MAX))
+ return mas_new_root(mas, wr_mas->entry);
+ /*
+ * Node rebalancing may occur due to this store, so there may be three new
+ * entries per level plus a new root.
+ */
+
+ /*
+ * Set up right side. Need to get to the next offset after the spanning
+ * store to ensure it's not NULL and to combine both the next node and
+ * the node with the start together.
+ */
+ r_mas = *mas;
+ /* Avoid overflow, walk to next slot in the tree. */
+ if (r_mas.last + 1)
+ r_mas.last++;
+
+ r_mas.index = r_mas.last;
+ mas_wr_walk_index(&r_wr_mas);
+ r_mas.last = r_mas.index = mas->last;
+ r_wr_mas.end_piv = r_wr_mas.r_max;
+
+ /* Set up left side. */
+ mas_wr_walk_index(wr_mas);
+
+ if (!wr_mas->entry) {
+ mas_extend_spanning_null(wr_mas, &r_wr_mas);
+ mas->last = r_mas.last;
+ }
+
+ /* expanding NULLs may make this cover the entire range */
+ if (!mas->index && r_mas.last == ULONG_MAX) {
+ mas_set_range(mas, 0, ULONG_MAX);
+ return mas_new_root(mas, wr_mas->entry);
+ }
+
+ cp_leaf_init(&cp, mas, wr_mas, &r_wr_mas);
+ do {
+ spanning_data(&cp, wr_mas, &r_wr_mas, &sib);
+ multi_src_setup(&cp, wr_mas, &r_wr_mas, &sib);
+ dst_setup(&cp, mas, wr_mas->type);
+ cp_data_write(&cp, mas);
+ } while (spanning_ascend(&cp, mas, wr_mas, &r_wr_mas, &sib));
+
+ mas_wmb_replace(mas, &cp);
+}
+
+/*
+ * mas_wr_node_store() - Attempt to store the value in a node
+ * @wr_mas: The maple write state
+ *
+ * Attempts to reuse the node, but may allocate.
+ */
+static inline void mas_wr_node_store(struct ma_wr_state *wr_mas)
+{
+ unsigned char dst_offset, offset_end;
+ unsigned char copy_size, node_pivots, node_slots;
+ struct maple_node reuse, *newnode;
+ unsigned long *dst_pivots;
+ void __rcu **dst_slots;
+ unsigned char new_end;
+ struct ma_state *mas;
+ bool in_rcu;
+
+ mas = wr_mas->mas;
+ trace_ma_op(TP_FCT, mas);
+ in_rcu = mt_in_rcu(mas->tree);
+ offset_end = wr_mas->offset_end;
+ node_pivots = mt_pivots[wr_mas->type];
+ node_slots = mt_slots[wr_mas->type];
+ /* Assume last adds an entry */
+ new_end = mas->end + 1 - offset_end + mas->offset;
+ if (mas->last == wr_mas->end_piv) {
+ offset_end++; /* don't copy this offset */
+ new_end--;
+ }
+
+ /* set up node. */
+ if (in_rcu) {
+ newnode = mas_pop_node(mas);
+ } else {
+ newnode = &reuse;
+ }
+
+ newnode->parent = mas_mn(mas)->parent;
+ dst_pivots = ma_pivots(newnode, wr_mas->type);
+ dst_slots = ma_slots(newnode, wr_mas->type);
+ /* Copy from start to insert point */
+ if (mas->offset) {
+ memcpy(dst_pivots, wr_mas->pivots, sizeof(unsigned long) * mas->offset);
+ memcpy(dst_slots, wr_mas->slots, sizeof(void __rcu *) * mas->offset);
+ }
+
+ /* Handle insert of new range starting after old range */
+ if (wr_mas->r_min < mas->index) {
+ rcu_assign_pointer(dst_slots[mas->offset], wr_mas->content);
+ dst_pivots[mas->offset++] = mas->index - 1;
+ new_end++;
+ }
+
+ /* Store the new entry and range end. */
+ if (mas->offset < node_pivots)
+ dst_pivots[mas->offset] = mas->last;
+ rcu_assign_pointer(dst_slots[mas->offset], wr_mas->entry);
+
+ /*
+ * this range wrote to the end of the node or it overwrote the rest of
+ * the data
+ */
+ if (offset_end > mas->end)
+ goto done;
+
+ dst_offset = mas->offset + 1;
+ /* Copy to the end of node if necessary. */
+ copy_size = mas->end - offset_end + 1;
+ memcpy(dst_slots + dst_offset, wr_mas->slots + offset_end,
+ sizeof(void __rcu *) * copy_size);
+ memcpy(dst_pivots + dst_offset, wr_mas->pivots + offset_end,
+ sizeof(unsigned long) * (copy_size - 1));
+
+ if (new_end < node_pivots)
+ dst_pivots[new_end] = mas->max;
+
+done:
+ if (!in_rcu && new_end + 2 < node_slots) {
+ unsigned char clear_from = new_end + 1;
+
+ /*
+ * Note that the last slot is never cleared, since the metadata
+ * will be stored there or it has a value.
+ */
+ memset(dst_slots + clear_from, 0,
+ sizeof(void __rcu *) * (node_slots - clear_from));
+ if (clear_from < node_pivots)
+ memset(dst_pivots + clear_from, 0,
+ sizeof(unsigned long) * (node_pivots - clear_from));
+ }
+
+ mas_leaf_set_meta(newnode, wr_mas->type, new_end);
+ if (in_rcu) {
+ struct maple_enode *old_enode = mas->node;
+
+ mas->node = mt_mk_node(newnode, wr_mas->type);
+ mas_replace_node(mas, old_enode, mas_mt_height(mas));
+ } else {
+ memcpy(wr_mas->node, newnode, sizeof(struct maple_node));
+ }
+ trace_ma_write(TP_FCT, mas, 0, wr_mas->entry);
+ mas_update_gap(mas);
+ mas->end = new_end;
+}
+
+/*
+ * mas_wr_slot_store: Attempt to store a value in a slot.
+ * @wr_mas: the maple write state
+ */
+static inline void mas_wr_slot_store(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+ unsigned char offset = mas->offset;
+ void __rcu **slots = wr_mas->slots;
+ bool gap = false;
+
+ gap |= !wr_mas->content;
+ gap |= !mt_slot_locked(mas->tree, slots, offset + 1);
+
+ if (wr_mas->offset_end - offset == 1) {
+ if (mas->index == wr_mas->r_min) {
+ /* Overwriting the range and a part of the next one */
+ rcu_assign_pointer(slots[offset], wr_mas->entry);
+ wr_mas->pivots[offset] = mas->last;
+ } else {
+ /* Overwriting a part of the range and the next one */
+ rcu_assign_pointer(slots[offset + 1], wr_mas->entry);
+ wr_mas->pivots[offset] = mas->index - 1;
+ mas->offset++; /* Keep mas accurate. */
+ }
+ } else {
+ WARN_ON_ONCE(mt_in_rcu(mas->tree));
+ /*
+ * Expand the range, only partially overwriting the previous and
+ * next ranges
+ */
+ gap |= !mt_slot_locked(mas->tree, slots, offset + 2);
+ rcu_assign_pointer(slots[offset + 1], wr_mas->entry);
+ wr_mas->pivots[offset] = mas->index - 1;
+ wr_mas->pivots[offset + 1] = mas->last;
+ mas->offset++; /* Keep mas accurate. */
+ }
+
+ trace_ma_write(TP_FCT, mas, 0, wr_mas->entry);
+ /*
+ * Only update gap when the new entry is empty or there is an empty
+ * entry in the original two ranges.
+ */
+ if (!wr_mas->entry || gap)
+ mas_update_gap(mas);
+}
+
+static inline void mas_wr_extend_null(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+
+ if (!wr_mas->slots[wr_mas->offset_end]) {
+ /* If this one is null, the next and prev are not */
+ mas->last = wr_mas->end_piv;
+ } else {
+ /* Check next slot(s) if we are overwriting the end */
+ if ((mas->last == wr_mas->end_piv) &&
+ (mas->end != wr_mas->offset_end) &&
+ !wr_mas->slots[wr_mas->offset_end + 1]) {
+ wr_mas->offset_end++;
+ if (wr_mas->offset_end == mas->end)
+ mas->last = mas->max;
+ else
+ mas->last = wr_mas->pivots[wr_mas->offset_end];
+ wr_mas->end_piv = mas->last;
+ }
+ }
+
+ if (!wr_mas->content) {
+ /* If this one is null, the next and prev are not */
+ mas->index = wr_mas->r_min;
+ } else {
+ /* Check prev slot if we are overwriting the start */
+ if (mas->index == wr_mas->r_min && mas->offset &&
+ !wr_mas->slots[mas->offset - 1]) {
+ mas->offset--;
+ wr_mas->r_min = mas->index =
+ mas_safe_min(mas, wr_mas->pivots, mas->offset);
+ wr_mas->r_max = wr_mas->pivots[mas->offset];
+ }
+ }
+}
+
+static inline void mas_wr_end_piv(struct ma_wr_state *wr_mas)
+{
+ while ((wr_mas->offset_end < wr_mas->mas->end) &&
+ (wr_mas->mas->last > wr_mas->pivots[wr_mas->offset_end]))
+ wr_mas->offset_end++;
+
+ if (wr_mas->offset_end < wr_mas->mas->end)
+ wr_mas->end_piv = wr_mas->pivots[wr_mas->offset_end];
+ else
+ wr_mas->end_piv = wr_mas->mas->max;
+}
+
+static inline unsigned char mas_wr_new_end(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+ unsigned char new_end = mas->end + 2;
+
+ new_end -= wr_mas->offset_end - mas->offset;
+ if (wr_mas->r_min == mas->index)
+ new_end--;
+
+ if (wr_mas->end_piv == mas->last)
+ new_end--;
+
+ return new_end;
+}
+
+/*
+ * mas_wr_append: Attempt to append
+ * @wr_mas: the maple write state
+ *
+ * This is currently unsafe in rcu mode since the end of the node may be cached
+ * by readers while the node contents may be updated which could result in
+ * inaccurate information.
+ */
+static inline void mas_wr_append(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+ void __rcu **slots;
+ unsigned char end = mas->end;
+ unsigned char new_end = mas_wr_new_end(wr_mas);
+
+ if (new_end < mt_pivots[wr_mas->type]) {
+ wr_mas->pivots[new_end] = wr_mas->pivots[end];
+ ma_set_meta(wr_mas->node, wr_mas->type, 0, new_end);
+ }
+
+ slots = wr_mas->slots;
+ if (new_end == end + 1) {
+ if (mas->last == wr_mas->r_max) {
+ /* Append to end of range */
+ rcu_assign_pointer(slots[new_end], wr_mas->entry);
+ wr_mas->pivots[end] = mas->index - 1;
+ mas->offset = new_end;
+ } else {
+ /* Append to start of range */
+ rcu_assign_pointer(slots[new_end], wr_mas->content);
+ wr_mas->pivots[end] = mas->last;
+ rcu_assign_pointer(slots[end], wr_mas->entry);
+ }
+ } else {
+ /* Append to the range without touching any boundaries. */
+ rcu_assign_pointer(slots[new_end], wr_mas->content);
+ wr_mas->pivots[end + 1] = mas->last;
+ rcu_assign_pointer(slots[end + 1], wr_mas->entry);
+ wr_mas->pivots[end] = mas->index - 1;
+ mas->offset = end + 1;
+ }
+
+ if (!wr_mas->content || !wr_mas->entry)
+ mas_update_gap(mas);
+
+ mas->end = new_end;
+ trace_ma_write(TP_FCT, mas, new_end, wr_mas->entry);
+}
+
+/*
+ * split_ascend() - See if a split operation has to keep walking up the tree
+ * @cp: The maple_copy node
+ * @wr_mas: The maple write state
+ * @sib: the maple state of the sibling
+ *
+ * Return: true if another split operation on the next level is needed, false
+ * otherwise
+ */
+static inline bool split_ascend(struct maple_copy *cp,
+ struct ma_wr_state *wr_mas, struct ma_state *sib,
+ struct ma_state *parent)
+{
+ struct ma_state *mas;
+ unsigned long min, max;
+
+ mas = wr_mas->mas;
+ min = mas->min; /* push right, or normal split */
+ max = mas->max;
+ wr_mas->offset_end = parent->offset;
+ if (sib->end) {
+ if (sib->max < mas->min) {
+ min = sib->min; /* push left */
+ parent->offset--;
+ } else {
+ max = sib->max; /* push right */
+ wr_mas->offset_end++;
+ }
+ }
+
+ cp_dst_to_slots(cp, min, max, mas);
+ if (cp_is_new_root(cp, mas))
+ return false;
+
+ if (cp_converged(cp, mas, sib))
+ return false;
+
+ cp->height++;
+ copy_tree_location(parent, mas);
+ wr_mas_setup(wr_mas, mas);
+ return true;
+}
+
+/*
+ * split_data() - Calculate the @cp data, populate @sib if the data can be
+ * pushed into a sibling.
+ * @cp: The maple copy node
+ * @wr_mas: The left write maple state
+ * @sib: The maple state of the sibling.
+ *
+ * Note: @cp->data is a size and not indexed by 0. @sib->end may be set to 0 to
+ * indicate it will not be used.
+ *
+ */
+static inline void split_data(struct maple_copy *cp,
+ struct ma_wr_state *wr_mas, struct ma_state *sib,
+ struct ma_state *parent)
+{
+ cp_data_calc(cp, wr_mas, wr_mas);
+ if (cp->data <= mt_slots[wr_mas->type]) {
+ sib->end = 0;
+ return;
+ }
+
+ push_data_sib(cp, wr_mas->mas, sib, parent);
+ if (sib->end)
+ cp->data += sib->end + 1;
+}
+
+/*
+ * mas_wr_split() - Expand one node into two
+ * @wr_mas: The write maple state
+ */
+static void mas_wr_split(struct ma_wr_state *wr_mas)
+{
+ struct ma_state parent;
+ struct ma_state *mas;
+ struct maple_copy cp;
+ struct ma_state sib;
+
+ mas = wr_mas->mas;
+ trace_ma_write(TP_FCT, wr_mas->mas, 0, wr_mas->entry);
+ parent = *mas;
+ cp_leaf_init(&cp, mas, wr_mas, wr_mas);
+ do {
+ if (!mte_is_root(parent.node)) {
+ mas_ascend(&parent);
+ parent.end = mas_data_end(&parent);
+ }
+ split_data(&cp, wr_mas, &sib, &parent);
+ multi_src_setup(&cp, wr_mas, wr_mas, &sib);
+ dst_setup(&cp, mas, wr_mas->type);
+ cp_data_write(&cp, mas);
+ } while (split_ascend(&cp, wr_mas, &sib, &parent));
+
+ mas_wmb_replace(mas, &cp);
+}
+
+/*
+ * mas_wr_rebalance() - Insufficient data in one node needs to either get data
+ * from a sibling or absorb a sibling all together.
+ * @wr_mas: The write maple state
+ *
+ * Rebalance is different than a spanning store in that the write state is
+ * already at the leaf node that's being altered.
+ */
+static void mas_wr_rebalance(struct ma_wr_state *wr_mas)
+{
+ struct ma_state parent;
+ struct ma_state *mas;
+ struct maple_copy cp;
+ struct ma_state sib;
+
+ /*
+ * Rebalancing occurs if a node is insufficient. Data is rebalanced
+ * against the node to the right if it exists, otherwise the node to the
+ * left of this node is rebalanced against this node. If rebalancing
+ * causes just one node to be produced instead of two, then the parent
+ * is also examined and rebalanced if it is insufficient. Every level
+ * tries to combine the data in the same way. If one node contains the
+ * entire range of the tree, then that node is used as a new root node.
+ */
+
+ mas = wr_mas->mas;
+ trace_ma_op(TP_FCT, mas);
+ parent = *mas;
+ cp_leaf_init(&cp, mas, wr_mas, wr_mas);
+ do {
+ if (!mte_is_root(parent.node)) {
+ mas_ascend(&parent);
+ parent.end = mas_data_end(&parent);
+ }
+ rebalance_data(&cp, wr_mas, &sib, &parent);
+ multi_src_setup(&cp, wr_mas, wr_mas, &sib);
+ dst_setup(&cp, mas, wr_mas->type);
+ cp_data_write(&cp, mas);
+ } while (rebalance_ascend(&cp, wr_mas, &sib, &parent));
+
+ mas_wmb_replace(mas, &cp);
+}
+
+/*
+ * mas_wr_store_entry() - Internal call to store a value
+ * @wr_mas: The maple write state
+ */
+static inline void mas_wr_store_entry(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+
+ switch (mas->store_type) {
+ case wr_exact_fit:
+ rcu_assign_pointer(wr_mas->slots[mas->offset], wr_mas->entry);
+ if (!!wr_mas->entry ^ !!wr_mas->content)
+ mas_update_gap(mas);
+ break;
+ case wr_append:
+ mas_wr_append(wr_mas);
+ break;
+ case wr_slot_store:
+ mas_wr_slot_store(wr_mas);
+ break;
+ case wr_node_store:
+ mas_wr_node_store(wr_mas);
+ break;
+ case wr_spanning_store:
+ mas_wr_spanning_store(wr_mas);
+ break;
+ case wr_split_store:
+ mas_wr_split(wr_mas);
+ break;
+ case wr_rebalance:
+ mas_wr_rebalance(wr_mas);
+ break;
+ case wr_new_root:
+ mas_new_root(mas, wr_mas->entry);
+ break;
+ case wr_store_root:
+ mas_store_root(mas, wr_mas->entry);
+ break;
+ case wr_invalid:
+ MT_BUG_ON(mas->tree, 1);
+ }
+}
+
+static inline void mas_wr_prealloc_setup(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+
+ if (!mas_is_active(mas)) {
+ if (mas_is_start(mas))
+ goto set_content;
+
+ if (unlikely(mas_is_paused(mas)))
+ goto reset;
+
+ if (unlikely(mas_is_none(mas)))
+ goto reset;
+
+ if (unlikely(mas_is_overflow(mas)))
+ goto reset;
+
+ if (unlikely(mas_is_underflow(mas)))
+ goto reset;
+ }
+
+ /*
+ * A less strict version of mas_is_span_wr() where we allow spanning
+ * writes within this node. This is to stop partial walks in
+ * mas_prealloc() from being reset.
+ */
+ if (mas->last > mas->max)
+ goto reset;
+
+ if (wr_mas->entry)
+ goto set_content;
+
+ if (mte_is_leaf(mas->node) && mas->last == mas->max)
+ goto reset;
+
+ goto set_content;
+
+reset:
+ mas_reset(mas);
+set_content:
+ wr_mas->content = mas_start(mas);
+}
+
+/**
+ * mas_prealloc_calc() - Calculate number of nodes needed for a
+ * given store operation
+ * @wr_mas: The maple write state
+ * @entry: The entry to store into the tree
+ *
+ * Return: Number of nodes required for preallocation.
+ */
+static inline void mas_prealloc_calc(struct ma_wr_state *wr_mas, void *entry)
+{
+ struct ma_state *mas = wr_mas->mas;
+ unsigned char height = mas_mt_height(mas);
+ int ret = height * 3 + 1;
+ unsigned char delta = height - wr_mas->vacant_height;
+
+ switch (mas->store_type) {
+ case wr_exact_fit:
+ case wr_append:
+ case wr_slot_store:
+ ret = 0;
+ break;
+ case wr_spanning_store:
+ if (wr_mas->sufficient_height < wr_mas->vacant_height)
+ ret = (height - wr_mas->sufficient_height) * 3 + 1;
+ else
+ ret = delta * 3 + 1;
+ break;
+ case wr_split_store:
+ ret = delta * 2 + 1;
+ break;
+ case wr_rebalance:
+ if (wr_mas->sufficient_height < wr_mas->vacant_height)
+ ret = (height - wr_mas->sufficient_height) * 2 + 1;
+ else
+ ret = delta * 2 + 1;
+ break;
+ case wr_node_store:
+ ret = mt_in_rcu(mas->tree) ? 1 : 0;
+ break;
+ case wr_new_root:
+ ret = 1;
+ break;
+ case wr_store_root:
+ if (likely((mas->last != 0) || (mas->index != 0)))
+ ret = 1;
+ else if (((unsigned long) (entry) & 3) == 2)
+ ret = 1;
+ else
+ ret = 0;
+ break;
+ case wr_invalid:
+ WARN_ON_ONCE(1);
+ }
+
+ mas->node_request = ret;
+}
+
+/*
+ * mas_wr_store_type() - Determine the store type for a given
+ * store operation.
+ * @wr_mas: The maple write state
+ *
+ * Return: the type of store needed for the operation
+ */
+static inline enum store_type mas_wr_store_type(struct ma_wr_state *wr_mas)
+{
+ struct ma_state *mas = wr_mas->mas;
+ unsigned char new_end;
+ bool appending;
+ bool one_slot;
+ bool in_rcu;
+
+ if (unlikely(mas_is_none(mas) || mas_is_ptr(mas)))
+ return wr_store_root;
+
+ if (unlikely(!mas_wr_walk(wr_mas)))
+ return wr_spanning_store;
+
+ /* At this point, we are at the leaf node that needs to be altered. */
+ mas_wr_end_piv(wr_mas);
+ if (!wr_mas->entry)
+ mas_wr_extend_null(wr_mas);
+
+ if ((wr_mas->r_min == mas->index) && (wr_mas->r_max == mas->last))
+ return wr_exact_fit;
+
+ if (unlikely(!mas->index && mas->last == ULONG_MAX))
+ return wr_new_root;
+
+ new_end = mas_wr_new_end(wr_mas);
+ in_rcu = mt_in_rcu(mas->tree);
+ appending = mas->offset == mas->end;
+ one_slot = wr_mas->offset_end - mas->offset == 1;
+
+ /* Potential spanning rebalance collapsing a node */
+ if (new_end < mt_min_slots[wr_mas->type]) {
+ if (!mte_is_root(mas->node))
+ return wr_rebalance;
+ if (!in_rcu) {
+ if (appending)
+ return wr_append;
+ else if (mas->end == new_end && one_slot)
+ return wr_slot_store;
+ }
+ return wr_node_store;
+ }
+
+ if (new_end >= mt_slots[wr_mas->type])
+ return wr_split_store;
+
+ if (!in_rcu && appending)
+ return wr_append;
+
+ if (new_end == mas->end && (!in_rcu || one_slot))
+ return wr_slot_store;
+
+ return wr_node_store;
+}
+
+/**
+ * mas_wr_preallocate() - Preallocate enough nodes for a store operation
+ * @wr_mas: The maple write state
+ * @entry: The entry that will be stored
+ *
+ */
+static inline void mas_wr_preallocate(struct ma_wr_state *wr_mas, void *entry)
+{
+ struct ma_state *mas = wr_mas->mas;
+
+ mas_wr_prealloc_setup(wr_mas);
+ mas->store_type = mas_wr_store_type(wr_mas);
+ mas_prealloc_calc(wr_mas, entry);
+ if (!mas->node_request)
+ return;
+
+ mas_alloc_nodes(mas, GFP_NOWAIT);
+}
+
+/**
+ * mas_insert() - Internal call to insert a value
+ * @mas: The maple state
+ * @entry: The entry to store
+ *
+ * Return: %NULL or the contents that already exists at the requested index
+ * otherwise. The maple state needs to be checked for error conditions.
+ */
+static inline void *mas_insert(struct ma_state *mas, void *entry)
+{
+ MA_WR_STATE(wr_mas, mas, entry);
+
+ /*
+ * Inserting a new range inserts either 0, 1, or 2 pivots within the
+ * tree. If the insert fits exactly into an existing gap with a value
+ * of NULL, then the slot only needs to be written with the new value.
+ * If the range being inserted is adjacent to another range, then only a
+ * single pivot needs to be inserted (as well as writing the entry). If
+ * the new range is within a gap but does not touch any other ranges,
+ * then two pivots need to be inserted: the start - 1, and the end. As
+ * usual, the entry must be written. Most operations require a new node
+ * to be allocated and replace an existing node to ensure RCU safety,
+ * when in RCU mode. The exception to requiring a newly allocated node
+ * is when inserting at the end of a node (appending). When done
+ * carefully, appending can reuse the node in place.
+ */
+ wr_mas.content = mas_start(mas);
+ if (wr_mas.content)
+ goto exists;
+
+ mas_wr_preallocate(&wr_mas, entry);
+ if (mas_is_err(mas))
+ return NULL;
+
+ /* spanning writes always overwrite something */
+ if (mas->store_type == wr_spanning_store)
+ goto exists;
+
+ /* At this point, we are at the leaf node that needs to be altered. */
+ if (mas->store_type != wr_new_root && mas->store_type != wr_store_root) {
+ wr_mas.offset_end = mas->offset;
+ wr_mas.end_piv = wr_mas.r_max;
+
+ if (wr_mas.content || (mas->last > wr_mas.r_max))
+ goto exists;
+ }
+
+ mas_wr_store_entry(&wr_mas);
+ return wr_mas.content;
+
+exists:
+ mas_set_err(mas, -EEXIST);
+ return wr_mas.content;
+
+}
+
+/**
+ * mas_alloc_cyclic() - Internal call to find somewhere to store an entry
+ * @mas: The maple state.
+ * @startp: Pointer to ID.
+ * @range_lo: Lower bound of range to search.
+ * @range_hi: Upper bound of range to search.
+ * @entry: The entry to store.
+ * @next: Pointer to next ID to allocate.
+ * @gfp: The GFP_FLAGS to use for allocations.
+ *
+ * Return: 0 if the allocation succeeded without wrapping, 1 if the
+ * allocation succeeded after wrapping, or -EBUSY if there are no
+ * free entries.
+ */
+int mas_alloc_cyclic(struct ma_state *mas, unsigned long *startp,
+ void *entry, unsigned long range_lo, unsigned long range_hi,
+ unsigned long *next, gfp_t gfp)
+{
+ int ret;
+ unsigned long min;
+
+ min = range_lo;
+ do {
+ range_lo = max(min, *next);
+ ret = mas_empty_area(mas, range_lo, range_hi, 1);
+ if (ret < 0 && range_lo > min) {
+ mas_reset(mas);
+ ret = mas_empty_area(mas, min, range_hi, 1);
+ if (ret == 0)
+ ret = 1;
+ }
+ if (ret < 0)
+ goto out;
+
+ mas_insert(mas, entry);
+ } while (mas_nomem(mas, gfp));
+
+ if (mas_is_err(mas)) {
+ ret = xa_err(mas->node);
+ goto out;
+ }
+
+ if ((mas->tree->ma_flags & MT_FLAGS_ALLOC_WRAPPED) && ret == 0) {
+ mas->tree->ma_flags &= ~MT_FLAGS_ALLOC_WRAPPED;
+ ret = 1;
+ }
+ *startp = mas->index;
+ *next = *startp + 1;
+ if (*next == 0)
+ mas->tree->ma_flags |= MT_FLAGS_ALLOC_WRAPPED;
+
+out:
+ mas_destroy(mas);
+ return ret;
+}
+EXPORT_SYMBOL(mas_alloc_cyclic);
+
+static __always_inline void mas_rewalk(struct ma_state *mas, unsigned long index)
+{
+retry:
+ mas_set(mas, index);
+ mas_state_walk(mas);
+ if (mas_is_start(mas))
+ goto retry;
+}
+
+static __always_inline bool mas_rewalk_if_dead(struct ma_state *mas,
+ struct maple_node *node, const unsigned long index)
+{
+ if (unlikely(ma_dead_node(node))) {
+ mas_rewalk(mas, index);
+ return true;
+ }
+ return false;
+}
+
+/*
+ * mas_prev_node() - Find the prev non-null entry at the same level in the
+ * tree. The prev value will be mas->node[mas->offset] or the status will be
+ * ma_none.
+ * @mas: The maple state
+ * @min: The lower limit to search
+ *
+ * The prev node value will be mas->node[mas->offset] or the status will be
+ * ma_none.
+ * Return: 1 if the node is dead, 0 otherwise.
+ */
+static int mas_prev_node(struct ma_state *mas, unsigned long min)
+{
+ enum maple_type mt;
+ int offset, level;
+ void __rcu **slots;
+ struct maple_node *node;
+ unsigned long *pivots;
+ unsigned long max;
+
+ node = mas_mn(mas);
+ if (!mas->min)
+ goto no_entry;
+
+ max = mas->min - 1;
+ if (max < min)
+ goto no_entry;
+
+ level = 0;
+ do {
+ if (ma_is_root(node))
+ goto no_entry;
+
+ /* Walk up. */
+ if (unlikely(mas_ascend(mas)))
+ return 1;
+ offset = mas->offset;
+ level++;
+ node = mas_mn(mas);
+ } while (!offset);
+
+ offset--;
+ mt = mte_node_type(mas->node);
+ while (level > 1) {
+ level--;
+ slots = ma_slots(node, mt);
+ mas->node = mas_slot(mas, slots, offset);
+ if (unlikely(ma_dead_node(node)))
+ return 1;
+
+ mt = mte_node_type(mas->node);
+ node = mas_mn(mas);
+ pivots = ma_pivots(node, mt);
+ offset = ma_data_end(node, mt, pivots, max);
+ if (unlikely(ma_dead_node(node)))
+ return 1;
+ }
+
+ slots = ma_slots(node, mt);
+ mas->node = mas_slot(mas, slots, offset);
+ pivots = ma_pivots(node, mt);
+ if (unlikely(ma_dead_node(node)))
+ return 1;
+
+ if (likely(offset))
+ mas->min = pivots[offset - 1] + 1;
+ mas->max = max;
+ mas->offset = mas_data_end(mas);
+ if (unlikely(mte_dead_node(mas->node)))
+ return 1;
+
+ mas->end = mas->offset;
+ return 0;
+
+no_entry:
+ if (unlikely(ma_dead_node(node)))
+ return 1;
+
+ mas->status = ma_underflow;
+ return 0;
+}
+
+/*
+ * mas_prev_slot() - Get the entry in the previous slot
+ *
+ * @mas: The maple state
+ * @min: The minimum starting range
+ * @empty: Can be empty
+ *
+ * Return: The entry in the previous slot which is possibly NULL
+ */
+static void *mas_prev_slot(struct ma_state *mas, unsigned long min, bool empty)
+{
+ void *entry;
+ void __rcu **slots;
+ unsigned long pivot;
+ enum maple_type type;
+ unsigned long *pivots;
+ struct maple_node *node;
+ unsigned long save_point = mas->index;
+
+retry:
+ node = mas_mn(mas);
+ type = mte_node_type(mas->node);
+ pivots = ma_pivots(node, type);
+ if (unlikely(mas_rewalk_if_dead(mas, node, save_point)))
+ goto retry;
+
+ if (mas->min <= min) {
+ pivot = mas_safe_min(mas, pivots, mas->offset);
+
+ if (unlikely(mas_rewalk_if_dead(mas, node, save_point)))
+ goto retry;
+
+ if (pivot <= min)
+ goto underflow;
+ }
+
+again:
+ if (likely(mas->offset)) {
+ mas->offset--;
+ mas->last = mas->index - 1;
+ mas->index = mas_safe_min(mas, pivots, mas->offset);
+ } else {
+ if (mas->index <= min)
+ goto underflow;
+
+ if (mas_prev_node(mas, min)) {
+ mas_rewalk(mas, save_point);
+ goto retry;
+ }
+
+ if (WARN_ON_ONCE(mas_is_underflow(mas)))
+ return NULL;
+
+ mas->last = mas->max;
+ node = mas_mn(mas);
+ type = mte_node_type(mas->node);
+ pivots = ma_pivots(node, type);
+ mas->index = pivots[mas->offset - 1] + 1;
+ }
+
+ slots = ma_slots(node, type);
+ entry = mas_slot(mas, slots, mas->offset);
+ if (unlikely(mas_rewalk_if_dead(mas, node, save_point)))
+ goto retry;
+
+ if (likely(entry))
+ return entry;
+
+ if (!empty) {
+ if (mas->index <= min)
+ goto underflow;
+
+ goto again;
+ }
+
+ return entry;
+
+underflow:
+ mas->status = ma_underflow;
+ return NULL;
+}
+
+/*
+ * mas_next_node() - Get the next node at the same level in the tree.
+ * @mas: The maple state
+ * @node: The maple node
+ * @max: The maximum pivot value to check.
+ *
+ * The next value will be mas->node[mas->offset] or the status will have
+ * overflowed.
+ * Return: 1 on dead node, 0 otherwise.
+ */
+static int mas_next_node(struct ma_state *mas, struct maple_node *node,
+ unsigned long max)
+{
+ unsigned long min;
+ unsigned long *pivots;
+ struct maple_enode *enode;
+ struct maple_node *tmp;
+ int level = 0;
+ unsigned char node_end;
+ enum maple_type mt;
+ void __rcu **slots;
+
+ if (mas->max >= max)
+ goto overflow;
+
+ min = mas->max + 1;
+ level = 0;
+ do {
+ if (ma_is_root(node))
+ goto overflow;
+
+ /* Walk up. */
+ if (unlikely(mas_ascend(mas)))
+ return 1;
+
+ level++;
+ node = mas_mn(mas);
+ mt = mte_node_type(mas->node);
+ pivots = ma_pivots(node, mt);
+ node_end = ma_data_end(node, mt, pivots, mas->max);
+ if (unlikely(ma_dead_node(node)))
+ return 1;
+
+ } while (unlikely(mas->offset == node_end));
+
+ slots = ma_slots(node, mt);
+ mas->offset++;
+ enode = mas_slot(mas, slots, mas->offset);
+ if (unlikely(ma_dead_node(node)))
+ return 1;
+
+ if (level > 1)
+ mas->offset = 0;
+
+ while (unlikely(level > 1)) {
+ level--;
+ mas->node = enode;
+ node = mas_mn(mas);
+ mt = mte_node_type(mas->node);
+ slots = ma_slots(node, mt);
+ enode = mas_slot(mas, slots, 0);
+ if (unlikely(ma_dead_node(node)))
+ return 1;
+ }
+
+ if (!mas->offset)
+ pivots = ma_pivots(node, mt);
+
+ mas->max = mas_safe_pivot(mas, pivots, mas->offset, mt);
+ tmp = mte_to_node(enode);
+ mt = mte_node_type(enode);
+ pivots = ma_pivots(tmp, mt);
+ mas->end = ma_data_end(tmp, mt, pivots, mas->max);
+ if (unlikely(ma_dead_node(node)))
+ return 1;
+
+ mas->node = enode;
+ mas->min = min;
+ return 0;
+
+overflow:
+ if (unlikely(ma_dead_node(node)))
+ return 1;
+
+ mas->status = ma_overflow;
+ return 0;
+}
+
+/*
+ * mas_next_slot() - Get the entry in the next slot
+ *
+ * @mas: The maple state
+ * @max: The maximum starting range
+ * @empty: Can be empty
+ *
+ * Return: The entry in the next slot which is possibly NULL
+ */
+static void *mas_next_slot(struct ma_state *mas, unsigned long max, bool empty)
+{
+ void __rcu **slots;
+ unsigned long *pivots;
+ unsigned long pivot;
+ enum maple_type type;
+ struct maple_node *node;
+ unsigned long save_point = mas->last;
+ void *entry;
+
+retry:
+ node = mas_mn(mas);
+ type = mte_node_type(mas->node);
+ pivots = ma_pivots(node, type);
+ if (unlikely(mas_rewalk_if_dead(mas, node, save_point)))
+ goto retry;
+
+ if (mas->max >= max) {
+ if (likely(mas->offset < mas->end))
+ pivot = pivots[mas->offset];
+ else
+ pivot = mas->max;
+
+ if (unlikely(mas_rewalk_if_dead(mas, node, save_point)))
+ goto retry;
+
+ if (pivot >= max) { /* Was at the limit, next will extend beyond */
+ mas->status = ma_overflow;
+ return NULL;
+ }
+ }
+
+ if (likely(mas->offset < mas->end)) {
+ mas->index = pivots[mas->offset] + 1;
+again:
+ mas->offset++;
+ if (likely(mas->offset < mas->end))
+ mas->last = pivots[mas->offset];
+ else
+ mas->last = mas->max;
+ } else {
+ if (mas->last >= max) {
+ mas->status = ma_overflow;
+ return NULL;
+ }
+
+ if (mas_next_node(mas, node, max)) {
+ mas_rewalk(mas, save_point);
+ goto retry;
+ }
+
+ if (WARN_ON_ONCE(mas_is_overflow(mas)))
+ return NULL;
+
+ mas->offset = 0;
+ mas->index = mas->min;
+ node = mas_mn(mas);
+ type = mte_node_type(mas->node);
+ pivots = ma_pivots(node, type);
+ mas->last = pivots[0];
+ }
+
+ slots = ma_slots(node, type);
+ entry = mt_slot(mas->tree, slots, mas->offset);
+ if (unlikely(mas_rewalk_if_dead(mas, node, save_point)))
+ goto retry;
+
+ if (entry)
+ return entry;
+
+
+ if (!empty) {
+ if (mas->last >= max) {
+ mas->status = ma_overflow;
+ return NULL;
+ }
+
+ mas->index = mas->last + 1;
+ goto again;
+ }
+
+ return entry;
+}
+
+/*
+ * mas_rev_awalk() - Internal function. Reverse allocation walk. Find the
+ * highest gap address of a given size in a given node and descend.
+ * @mas: The maple state
+ * @size: The needed size.
+ *
+ * Return: True if found in a leaf, false otherwise.
+ *
+ */
+static bool mas_rev_awalk(struct ma_state *mas, unsigned long size,
+ unsigned long *gap_min, unsigned long *gap_max)
+{
+ enum maple_type type = mte_node_type(mas->node);
+ struct maple_node *node = mas_mn(mas);
+ unsigned long *pivots, *gaps;
+ void __rcu **slots;
+ unsigned long gap = 0;
+ unsigned long max, min;
+ unsigned char offset;
+
+ if (unlikely(mas_is_err(mas)))
+ return true;
+
+ if (ma_is_dense(type)) {
+ /* dense nodes. */
+ mas->offset = (unsigned char)(mas->index - mas->min);
+ return true;
+ }
+
+ pivots = ma_pivots(node, type);
+ slots = ma_slots(node, type);
+ gaps = ma_gaps(node, type);
+ offset = mas->offset;
+ min = mas_safe_min(mas, pivots, offset);
+ /* Skip out of bounds. */
+ while (mas->last < min)
+ min = mas_safe_min(mas, pivots, --offset);
+
+ max = mas_safe_pivot(mas, pivots, offset, type);
+ while (mas->index <= max) {
+ gap = 0;
+ if (gaps)
+ gap = gaps[offset];
+ else if (!mas_slot(mas, slots, offset))
+ gap = max - min + 1;
+
+ if (gap) {
+ if ((size <= gap) && (size <= mas->last - min + 1))
+ break;
+
+ if (!gaps) {
+ /* Skip the next slot, it cannot be a gap. */
+ if (offset < 2)
+ goto ascend;
+
+ offset -= 2;
+ max = pivots[offset];
+ min = mas_safe_min(mas, pivots, offset);
+ continue;
+ }
+ }
+
+ if (!offset)
+ goto ascend;
+
+ offset--;
+ max = min - 1;
+ min = mas_safe_min(mas, pivots, offset);
+ }
+
+ if (unlikely((mas->index > max) || (size - 1 > max - mas->index)))
+ goto no_space;
+
+ if (unlikely(ma_is_leaf(type))) {
+ mas->offset = offset;
+ *gap_min = min;
+ *gap_max = min + gap - 1;
+ return true;
+ }
+
+ /* descend, only happens under lock. */
+ mas->node = mas_slot(mas, slots, offset);
+ mas->min = min;
+ mas->max = max;
+ mas->offset = mas_data_end(mas);
+ return false;
+
+ascend:
+ if (!mte_is_root(mas->node))
+ return false;
+
+no_space:
+ mas_set_err(mas, -EBUSY);
+ return false;
+}
+
+static inline bool mas_anode_descend(struct ma_state *mas, unsigned long size)
+{
+ enum maple_type type = mte_node_type(mas->node);
+ unsigned long pivot, min, gap = 0;
+ unsigned char offset, data_end;
+ unsigned long *gaps, *pivots;
+ void __rcu **slots;
+ struct maple_node *node;
+ bool found = false;
+
+ if (ma_is_dense(type)) {
+ mas->offset = (unsigned char)(mas->index - mas->min);
+ return true;
+ }
+
+ node = mas_mn(mas);
+ pivots = ma_pivots(node, type);
+ slots = ma_slots(node, type);
+ gaps = ma_gaps(node, type);
+ offset = mas->offset;
+ min = mas_safe_min(mas, pivots, offset);
+ data_end = ma_data_end(node, type, pivots, mas->max);
+ for (; offset <= data_end; offset++) {
+ pivot = mas_safe_pivot(mas, pivots, offset, type);
+
+ /* Not within lower bounds */
+ if (mas->index > pivot)
+ goto next_slot;
+
+ if (gaps)
+ gap = gaps[offset];
+ else if (!mas_slot(mas, slots, offset))
+ gap = min(pivot, mas->last) - max(mas->index, min) + 1;
+ else
+ goto next_slot;
+
+ if (gap >= size) {
+ if (ma_is_leaf(type)) {
+ found = true;
+ break;
+ }
+
+ mas->node = mas_slot(mas, slots, offset);
+ mas->min = min;
+ mas->max = pivot;
+ offset = 0;
+ break;
+ }
+next_slot:
+ min = pivot + 1;
+ if (mas->last <= pivot) {
+ mas_set_err(mas, -EBUSY);
+ return true;
+ }
+ }
+
+ mas->offset = offset;
+ return found;
+}
+
+/**
+ * mas_walk() - Search for @mas->index in the tree.
+ * @mas: The maple state.
+ *
+ * mas->index and mas->last will be set to the range if there is a value. If
+ * mas->status is ma_none, reset to ma_start
+ *
+ * Return: the entry at the location or %NULL.
+ */
+void *mas_walk(struct ma_state *mas)
+{
+ void *entry;
+
+ mas_may_init_lock_check(mas);
+ mas_make_walkable(mas);
+retry:
+ entry = mas_state_walk(mas);
+ if (mas_is_start(mas)) {
+ goto retry;
+ } else if (mas_is_none(mas)) {
+ mas->index = 0;
+ mas->last = ULONG_MAX;
+ } else if (mas_is_ptr(mas)) {
+ if (!mas->index) {
+ mas->last = 0;
+ return entry;
+ }
+
+ mas->index = 1;
+ mas->last = ULONG_MAX;
+ mas->status = ma_none;
+ return NULL;
+ }
+
+ return entry;
+}
+EXPORT_SYMBOL_GPL(mas_walk);
+
+static inline bool mas_rewind_node(struct ma_state *mas)
+{
+ unsigned char slot;
+
+ do {
+ if (mte_is_root(mas->node)) {
+ slot = mas->offset;
+ if (!slot)
+ return false;
+ } else {
+ mas_ascend(mas);
+ slot = mas->offset;
+ }
+ } while (!slot);
+
+ mas->offset = --slot;
+ return true;
+}
+
+/*
+ * mas_skip_node() - Internal function. Skip over a node.
+ * @mas: The maple state.
+ *
+ * Return: true if there is another node, false otherwise.
+ */
+static inline bool mas_skip_node(struct ma_state *mas)
+{
+ if (mas_is_err(mas))
+ return false;
+
+ do {
+ if (mte_is_root(mas->node)) {
+ if (mas->offset >= mas_data_end(mas)) {
+ mas_set_err(mas, -EBUSY);
+ return false;
+ }
+ } else {
+ mas_ascend(mas);
+ }
+ } while (mas->offset >= mas_data_end(mas));
+
+ mas->offset++;
+ return true;
+}
+
+/*
+ * mas_awalk() - Allocation walk. Search from low address to high, for a gap of
+ * @size
+ * @mas: The maple state
+ * @size: The size of the gap required
+ *
+ * Search between @mas->index and @mas->last for a gap of @size.
+ */
+static inline void mas_awalk(struct ma_state *mas, unsigned long size)
+{
+ struct maple_enode *last = NULL;
+
+ /*
+ * There are 4 options:
+ * go to child (descend)
+ * go back to parent (ascend)
+ * no gap found. (return, error == -EBUSY)
+ * found the gap. (return)
+ */
+ while (!mas_is_err(mas) && !mas_anode_descend(mas, size)) {
+ if (last == mas->node)
+ mas_skip_node(mas);
+ else
+ last = mas->node;
+ }
+}
+
+/*
+ * mas_sparse_area() - Internal function. Return upper or lower limit when
+ * searching for a gap in an empty tree.
+ * @mas: The maple state
+ * @min: the minimum range
+ * @max: The maximum range
+ * @size: The size of the gap
+ * @fwd: Searching forward or back
+ */
+static inline int mas_sparse_area(struct ma_state *mas, unsigned long min,
+ unsigned long max, unsigned long size, bool fwd)
+{
+ if (!unlikely(mas_is_none(mas)) && min == 0) {
+ min++;
+ /*
+ * At this time, min is increased, we need to recheck whether
+ * the size is satisfied.
+ */
+ if (min > max || max - min + 1 < size)
+ return -EBUSY;
+ }
+ /* mas_is_ptr */
+
+ if (fwd) {
+ mas->index = min;
+ mas->last = min + size - 1;
+ } else {
+ mas->last = max;
+ mas->index = max - size + 1;
+ }
+ return 0;
+}
+
+/*
+ * mas_empty_area() - Get the lowest address within the range that is
+ * sufficient for the size requested.
+ * @mas: The maple state
+ * @min: The lowest value of the range
+ * @max: The highest value of the range
+ * @size: The size needed
+ */
+int mas_empty_area(struct ma_state *mas, unsigned long min,
+ unsigned long max, unsigned long size)
+{
+ unsigned char offset;
+ unsigned long *pivots;
+ enum maple_type mt;
+ struct maple_node *node;
+
+ if (min > max)
+ return -EINVAL;
+
+ if (size == 0 || max - min < size - 1)
+ return -EINVAL;
+
+ if (mas_is_start(mas))
+ mas_start(mas);
+ else if (mas->offset >= 2)
+ mas->offset -= 2;
+ else if (!mas_skip_node(mas))
+ return -EBUSY;
+
+ /* Empty set */
+ if (mas_is_none(mas) || mas_is_ptr(mas))
+ return mas_sparse_area(mas, min, max, size, true);
+
+ /* The start of the window can only be within these values */
+ mas->index = min;
+ mas->last = max;
+ mas_awalk(mas, size);
+
+ if (unlikely(mas_is_err(mas)))
+ return xa_err(mas->node);
+
+ offset = mas->offset;
+ node = mas_mn(mas);
+ mt = mte_node_type(mas->node);
+ pivots = ma_pivots(node, mt);
+ min = mas_safe_min(mas, pivots, offset);
+ if (mas->index < min)
+ mas->index = min;
+ mas->last = mas->index + size - 1;
+ mas->end = ma_data_end(node, mt, pivots, mas->max);
+ return 0;
+}
+EXPORT_SYMBOL_GPL(mas_empty_area);
+
+/*
+ * mas_empty_area_rev() - Get the highest address within the range that is
+ * sufficient for the size requested.
+ * @mas: The maple state
+ * @min: The lowest value of the range
+ * @max: The highest value of the range
+ * @size: The size needed
+ */
+int mas_empty_area_rev(struct ma_state *mas, unsigned long min,
+ unsigned long max, unsigned long size)
+{
+ struct maple_enode *last = mas->node;
+
+ if (min > max)
+ return -EINVAL;
+
+ if (size == 0 || max - min < size - 1)
+ return -EINVAL;
+
+ if (mas_is_start(mas))
+ mas_start(mas);
+ else if ((mas->offset < 2) && (!mas_rewind_node(mas)))
+ return -EBUSY;
+
+ if (unlikely(mas_is_none(mas) || mas_is_ptr(mas)))
+ return mas_sparse_area(mas, min, max, size, false);
+ else if (mas->offset >= 2)
+ mas->offset -= 2;
+ else
+ mas->offset = mas_data_end(mas);
+
+
+ /* The start of the window can only be within these values. */
+ mas->index = min;
+ mas->last = max;
+
+ while (!mas_rev_awalk(mas, size, &min, &max)) {
+ if (last == mas->node) {
+ if (!mas_rewind_node(mas))
+ return -EBUSY;
+ } else {
+ last = mas->node;
+ }
+ }
+
+ if (mas_is_err(mas))
+ return xa_err(mas->node);
+
+ if (unlikely(mas->offset == MAPLE_NODE_SLOTS))
+ return -EBUSY;
+
+ /* Trim the upper limit to the max. */
+ if (max < mas->last)
+ mas->last = max;
+
+ mas->index = mas->last - size + 1;
+ mas->end = mas_data_end(mas);
+ return 0;
+}
+EXPORT_SYMBOL_GPL(mas_empty_area_rev);
+
+/*
+ * mte_dead_leaves() - Mark all leaves of a node as dead.
+ * @enode: the encoded node
+ * @mt: the maple tree
+ * @slots: Pointer to the slot array
+ *
+ * Must hold the write lock.
+ *
+ * Return: The number of leaves marked as dead.
+ */
+static inline
+unsigned char mte_dead_leaves(struct maple_enode *enode, struct maple_tree *mt,
+ void __rcu **slots)
+{
+ struct maple_node *node;
+ enum maple_type type;
+ void *entry;
+ int offset;
+
+ for (offset = 0; offset < mt_slot_count(enode); offset++) {
+ entry = mt_slot(mt, slots, offset);
+ type = mte_node_type(entry);
+ node = mte_to_node(entry);
+ /* Use both node and type to catch LE & BE metadata */
+ if (!node || !type)
+ break;
+
+ mte_set_node_dead(entry);
+ node->type = type;
+ rcu_assign_pointer(slots[offset], node);
+ }
+
+ return offset;
+}
+
+/**
+ * mte_dead_walk() - Walk down a dead tree to just before the leaves
+ * @enode: The maple encoded node
+ * @offset: The starting offset
+ *
+ * Note: This can only be used from the RCU callback context.
+ */
+static void __rcu **mte_dead_walk(struct maple_enode **enode, unsigned char offset)
+{
+ struct maple_node *node, *next;
+ void __rcu **slots = NULL;
+
+ next = mte_to_node(*enode);
+ do {
+ *enode = ma_enode_ptr(next);
+ node = mte_to_node(*enode);
+ slots = ma_slots(node, node->type);
+ next = rcu_dereference_protected(slots[offset],
+ lock_is_held(&rcu_callback_map));
+ offset = 0;
+ } while (!ma_is_leaf(next->type));
+
+ return slots;
+}
+
+/**
+ * mt_free_walk() - Walk & free a tree in the RCU callback context
+ * @head: The RCU head that's within the node.
+ *
+ * Note: This can only be used from the RCU callback context.
+ */
+static void mt_free_walk(struct rcu_head *head)
+{
+ void __rcu **slots;
+ struct maple_node *node, *start;
+ struct maple_enode *enode;
+ unsigned char offset;
+ enum maple_type type;
+
+ node = container_of(head, struct maple_node, rcu);
+
+ if (ma_is_leaf(node->type))
+ goto free_leaf;
+
+ start = node;
+ enode = mt_mk_node(node, node->type);
+ slots = mte_dead_walk(&enode, 0);
+ node = mte_to_node(enode);
+ do {
+ mt_free_bulk(node->slot_len, slots);
+ offset = node->parent_slot + 1;
+ enode = node->piv_parent;
+ if (mte_to_node(enode) == node)
+ goto free_leaf;
+
+ type = mte_node_type(enode);
+ slots = ma_slots(mte_to_node(enode), type);
+ if ((offset < mt_slots[type]) &&
+ rcu_dereference_protected(slots[offset],
+ lock_is_held(&rcu_callback_map)))
+ slots = mte_dead_walk(&enode, offset);
+ node = mte_to_node(enode);
+ } while ((node != start) || (node->slot_len < offset));
+
+ slots = ma_slots(node, node->type);
+ mt_free_bulk(node->slot_len, slots);
+
+free_leaf:
+ kfree(node);
+}
+
+static inline void __rcu **mte_destroy_descend(struct maple_enode **enode,
+ struct maple_tree *mt, struct maple_enode *prev, unsigned char offset)
+{
+ struct maple_node *node;
+ struct maple_enode *next = *enode;
+ void __rcu **slots = NULL;
+ enum maple_type type;
+ unsigned char next_offset = 0;
+
+ do {
+ *enode = next;
+ node = mte_to_node(*enode);
+ type = mte_node_type(*enode);
+ slots = ma_slots(node, type);
+ next = mt_slot_locked(mt, slots, next_offset);
+ if ((mte_dead_node(next)))
+ next = mt_slot_locked(mt, slots, ++next_offset);
+
+ mte_set_node_dead(*enode);
+ node->type = type;
+ node->piv_parent = prev;
+ node->parent_slot = offset;
+ offset = next_offset;
+ next_offset = 0;
+ prev = *enode;
+ } while (!mte_is_leaf(next));
+
+ return slots;
+}
+
+static void mt_destroy_walk(struct maple_enode *enode, struct maple_tree *mt,
+ bool free)
+{
+ void __rcu **slots;
+ struct maple_node *node = mte_to_node(enode);
+ struct maple_enode *start;
+
+ if (mte_is_leaf(enode)) {
+ mte_set_node_dead(enode);
+ node->type = mte_node_type(enode);
+ goto free_leaf;
+ }
+
+ start = enode;
+ slots = mte_destroy_descend(&enode, mt, start, 0);
+ node = mte_to_node(enode); // Updated in the above call.
+ do {
+ enum maple_type type;
+ unsigned char offset;
+ struct maple_enode *parent, *tmp;
+
+ node->slot_len = mte_dead_leaves(enode, mt, slots);
+ if (free)
+ mt_free_bulk(node->slot_len, slots);
+ offset = node->parent_slot + 1;
+ enode = node->piv_parent;
+ if (mte_to_node(enode) == node)
+ goto free_leaf;
+
+ type = mte_node_type(enode);
+ slots = ma_slots(mte_to_node(enode), type);
+ if (offset >= mt_slots[type])
+ goto next;
+
+ tmp = mt_slot_locked(mt, slots, offset);
+ if (mte_node_type(tmp) && mte_to_node(tmp)) {
+ parent = enode;
+ enode = tmp;
+ slots = mte_destroy_descend(&enode, mt, parent, offset);
+ }
+next:
+ node = mte_to_node(enode);
+ } while (start != enode);
+
+ node = mte_to_node(enode);
+ node->slot_len = mte_dead_leaves(enode, mt, slots);
+ if (free)
+ mt_free_bulk(node->slot_len, slots);
+
+free_leaf:
+ if (free)
+ kfree(node);
+ else
+ mt_clear_meta(mt, node, node->type);
+}
+
+/*
+ * mte_destroy_walk() - Free a tree or sub-tree.
+ * @enode: the encoded maple node (maple_enode) to start
+ * @mt: the tree to free - needed for node types.
+ *
+ * Must hold the write lock.
+ */
+static inline void mte_destroy_walk(struct maple_enode *enode,
+ struct maple_tree *mt)
+{
+ struct maple_node *node = mte_to_node(enode);
+
+ if (mt_in_rcu(mt)) {
+ mt_destroy_walk(enode, mt, false);
+ call_rcu(&node->rcu, mt_free_walk);
+ } else {
+ mt_destroy_walk(enode, mt, true);
+ }
+}
+/* Interface */
+
+/**
+ * mas_store() - Store an @entry.
+ * @mas: The maple state.
+ * @entry: The entry to store.
+ *
+ * The @mas->index and @mas->last is used to set the range for the @entry.
+ *
+ * Return: the first entry between mas->index and mas->last or %NULL.
+ */
+void *mas_store(struct ma_state *mas, void *entry)
+{
+ MA_WR_STATE(wr_mas, mas, entry);
+
+ mas_may_init_lock_check(mas);
+ trace_ma_write(TP_FCT, mas, 0, entry);
+#ifdef CONFIG_DEBUG_MAPLE_TREE
+ if (MAS_WARN_ON(mas, mas->index > mas->last))
+ pr_err("Error %lX > %lX " PTR_FMT "\n", mas->index, mas->last,
+ entry);
+
+ if (mas->index > mas->last) {
+ mas_set_err(mas, -EINVAL);
+ return NULL;
+ }
+
+#endif
+
+ /*
+ * Storing is the same operation as insert with the added caveat that it
+ * can overwrite entries. Although this seems simple enough, one may
+ * want to examine what happens if a single store operation was to
+ * overwrite multiple entries within a self-balancing B-Tree.
+ */
+ mas_wr_prealloc_setup(&wr_mas);
+ mas->store_type = mas_wr_store_type(&wr_mas);
+ if (mas->mas_flags & MA_STATE_PREALLOC) {
+ mas_wr_store_entry(&wr_mas);
+ MAS_WR_BUG_ON(&wr_mas, mas_is_err(mas));
+ return wr_mas.content;
+ }
+
+ mas_prealloc_calc(&wr_mas, entry);
+ if (!mas->node_request)
+ goto store;
+
+ mas_alloc_nodes(mas, GFP_NOWAIT);
+ if (mas_is_err(mas))
+ return NULL;
+
+store:
+ mas_wr_store_entry(&wr_mas);
+ mas_destroy(mas);
+ return wr_mas.content;
+}
+EXPORT_SYMBOL_GPL(mas_store);
+
+/**
+ * mas_store_gfp() - Store a value into the tree.
+ * @mas: The maple state
+ * @entry: The entry to store
+ * @gfp: The GFP_FLAGS to use for allocations if necessary.
+ *
+ * Return: 0 on success, -EINVAL on invalid request, -ENOMEM if memory could not
+ * be allocated.
+ */
+int mas_store_gfp(struct ma_state *mas, void *entry, gfp_t gfp)
+{
+ unsigned long index = mas->index;
+ unsigned long last = mas->last;
+ MA_WR_STATE(wr_mas, mas, entry);
+ int ret = 0;
+
+ mas_may_init_lock_check(mas);
+retry:
+ mas_wr_preallocate(&wr_mas, entry);
+ if (unlikely(mas_nomem(mas, gfp))) {
+ if (!entry)
+ __mas_set_range(mas, index, last);
+ goto retry;
+ }
+
+ if (mas_is_err(mas)) {
+ ret = xa_err(mas->node);
+ goto out;
+ }
+
+ mas_wr_store_entry(&wr_mas);
+out:
+ mas_destroy(mas);
+ return ret;
+}
+EXPORT_SYMBOL_GPL(mas_store_gfp);
+
+/**
+ * mas_store_prealloc() - Store a value into the tree using memory
+ * preallocated in the maple state.
+ * @mas: The maple state
+ * @entry: The entry to store.
+ */
+void mas_store_prealloc(struct ma_state *mas, void *entry)
+{
+ MA_WR_STATE(wr_mas, mas, entry);
+
+ mas_lock_check(mas);
+ if (mas->store_type == wr_store_root) {
+ mas_wr_prealloc_setup(&wr_mas);
+ goto store;
+ }
+
+ mas_wr_walk_descend(&wr_mas);
+ if (mas->store_type != wr_spanning_store) {
+ /* set wr_mas->content to current slot */
+ wr_mas.content = mas_slot_locked(mas, wr_mas.slots, mas->offset);
+ mas_wr_end_piv(&wr_mas);
+ }
+
+store:
+ trace_ma_write(TP_FCT, mas, 0, entry);
+ mas_wr_store_entry(&wr_mas);
+ MAS_WR_BUG_ON(&wr_mas, mas_is_err(mas));
+ mas_destroy(mas);
+}
+EXPORT_SYMBOL_GPL(mas_store_prealloc);
+
+/**
+ * mas_preallocate() - Preallocate enough nodes for a store operation
+ * @mas: The maple state
+ * @entry: The entry that will be stored
+ * @gfp: The GFP_FLAGS to use for allocations.
+ *
+ * Return: 0 on success, -ENOMEM if memory could not be allocated.
+ */
+int mas_preallocate(struct ma_state *mas, void *entry, gfp_t gfp)
+{
+ MA_WR_STATE(wr_mas, mas, entry);
+
+ mas_may_init_lock_check(mas);
+ mas_wr_prealloc_setup(&wr_mas);
+ mas->store_type = mas_wr_store_type(&wr_mas);
+ mas_prealloc_calc(&wr_mas, entry);
+ if (!mas->node_request)
+ goto set_flag;
+
+ mas->mas_flags &= ~MA_STATE_PREALLOC;
+ mas_alloc_nodes(mas, gfp);
+ if (mas_is_err(mas)) {
+ int ret = xa_err(mas->node);
+
+ mas->node_request = 0;
+ mas_destroy(mas);
+ mas_reset(mas);
+ return ret;
+ }
+
+set_flag:
+ mas->mas_flags |= MA_STATE_PREALLOC;
+ return 0;
+}
+EXPORT_SYMBOL_GPL(mas_preallocate);
+
+/*
+ * mas_destroy() - destroy a maple state.
+ * @mas: The maple state
+ *
+ * Upon completion, check the left-most node and rebalance against the node to
+ * the right if necessary. Frees any allocated nodes associated with this maple
+ * state.
+ */
+void mas_destroy(struct ma_state *mas)
+{
+ mas->mas_flags &= ~MA_STATE_PREALLOC;
+ mas_empty_nodes(mas);
+}
+EXPORT_SYMBOL_GPL(mas_destroy);
+
+static void mas_may_activate(struct ma_state *mas)
+{
+ if (!mas->node) {
+ mas->status = ma_start;
+ } else if (mas->index > mas->max || mas->index < mas->min) {
+ mas->status = ma_start;
+ } else {
+ mas->status = ma_active;
+ mas_lock_check(mas);
+ }
+}
+
+static bool mas_next_setup(struct ma_state *mas, unsigned long max,
+ void **entry)
+{
+ bool was_none = mas_is_none(mas);
+
+ if (unlikely(mas->last >= max)) {
+ mas->status = ma_overflow;
+ return true;
+ }
+
+ switch (mas->status) {
+ case ma_active:
+ return false;
+ case ma_none:
+ fallthrough;
+ case ma_pause:
+ mas->status = ma_start;
+ fallthrough;
+ case ma_start:
+ mas_walk(mas); /* Retries on dead nodes handled by mas_walk */
+ break;
+ case ma_overflow:
+ /* Overflowed before, but the max changed */
+ mas_may_activate(mas);
+ break;
+ case ma_underflow:
+ /* The user expects the mas to be one before where it is */
+ mas_may_activate(mas);
+ *entry = mas_walk(mas);
+ if (*entry)
+ return true;
+ break;
+ case ma_root:
+ break;
+ case ma_error:
+ return true;
+ }
+
+ if (likely(mas_is_active(mas))) /* Fast path */
+ return false;
+
+ if (mas_is_ptr(mas)) {
+ *entry = NULL;
+ if (was_none && mas->index == 0) {
+ mas->index = mas->last = 0;
+ return true;
+ }
+ mas->index = 1;
+ mas->last = ULONG_MAX;
+ mas->status = ma_none;
+ return true;
+ }
+
+ if (mas_is_none(mas))
+ return true;
+
+ return false;
+}
+
+/**
+ * mas_next() - Get the next entry.
+ * @mas: The maple state
+ * @max: The maximum index to check.
+ *
+ * Returns the next entry after @mas->index.
+ * Must hold rcu_read_lock or the write lock.
+ * Can return the zero entry.
+ *
+ * Return: The next entry or %NULL
+ */
+void *mas_next(struct ma_state *mas, unsigned long max)
+{
+ void *entry = NULL;
+
+ mas_may_init_lock_check(mas);
+ if (mas_next_setup(mas, max, &entry))
+ return entry;
+
+ /* Retries on dead nodes handled by mas_next_slot */
+ return mas_next_slot(mas, max, false);
+}
+EXPORT_SYMBOL_GPL(mas_next);
+
+/**
+ * mas_next_range() - Advance the maple state to the next range
+ * @mas: The maple state
+ * @max: The maximum index to check.
+ *
+ * Sets @mas->index and @mas->last to the range.
+ * Must hold rcu_read_lock or the write lock.
+ * Can return the zero entry.
+ *
+ * Return: The next entry or %NULL
+ */
+void *mas_next_range(struct ma_state *mas, unsigned long max)
+{
+ void *entry = NULL;
+
+ mas_may_init_lock_check(mas);
+ if (mas_next_setup(mas, max, &entry))
+ return entry;
+
+ /* Retries on dead nodes handled by mas_next_slot */
+ return mas_next_slot(mas, max, true);
+}
+EXPORT_SYMBOL_GPL(mas_next_range);
+
+/**
+ * mt_next() - get the next value in the maple tree
+ * @mt: The maple tree
+ * @index: The start index
+ * @max: The maximum index to check
+ *
+ * Takes RCU read lock internally to protect the search, which does not
+ * protect the returned pointer after dropping RCU read lock.
+ * See also: Documentation/core-api/maple_tree.rst
+ *
+ * Return: The entry higher than @index or %NULL if nothing is found.
+ */
+void *mt_next(struct maple_tree *mt, unsigned long index, unsigned long max)
+{
+ void *entry = NULL;
+ MA_STATE(mas, mt, index, index);
+
+ rcu_read_lock();
+ entry = mas_next(&mas, max);
+ rcu_read_unlock();
+ return entry;
+}
+EXPORT_SYMBOL_GPL(mt_next);
+
+static bool mas_prev_setup(struct ma_state *mas, unsigned long min, void **entry)
+{
+ if (unlikely(mas->index <= min)) {
+ mas->status = ma_underflow;
+ return true;
+ }
+
+ switch (mas->status) {
+ case ma_active:
+ return false;
+ case ma_start:
+ break;
+ case ma_none:
+ fallthrough;
+ case ma_pause:
+ mas->status = ma_start;
+ break;
+ case ma_underflow:
+ /* underflowed before but the min changed */
+ mas_may_activate(mas);
+ break;
+ case ma_overflow:
+ /* User expects mas to be one after where it is */
+ mas_may_activate(mas);
+ *entry = mas_walk(mas);
+ if (*entry)
+ return true;
+ break;
+ case ma_root:
+ break;
+ case ma_error:
+ return true;
+ }
+
+ if (mas_is_start(mas))
+ mas_walk(mas);
+
+ if (unlikely(mas_is_ptr(mas))) {
+ if (!mas->index) {
+ mas->status = ma_none;
+ return true;
+ }
+ mas->index = mas->last = 0;
+ *entry = mas_root(mas);
+ return true;
+ }
+
+ if (mas_is_none(mas)) {
+ if (mas->index) {
+ /* Walked to out-of-range pointer? */
+ mas->index = mas->last = 0;
+ mas->status = ma_root;
+ *entry = mas_root(mas);
+ return true;
+ }
+ return true;
+ }
+
+ return false;
+}
+
+/**
+ * mas_prev() - Get the previous entry
+ * @mas: The maple state
+ * @min: The minimum value to check.
+ *
+ * Must hold rcu_read_lock or the write lock.
+ * Will reset mas to ma_start if the status is ma_none. Will stop on not
+ * searchable nodes.
+ *
+ * Return: the previous value or %NULL.
+ */
+void *mas_prev(struct ma_state *mas, unsigned long min)
+{
+ void *entry = NULL;
+
+ mas_may_init_lock_check(mas);
+ if (mas_prev_setup(mas, min, &entry))
+ return entry;
+
+ return mas_prev_slot(mas, min, false);
+}
+EXPORT_SYMBOL_GPL(mas_prev);
+
+/**
+ * mas_prev_range() - Advance to the previous range
+ * @mas: The maple state
+ * @min: The minimum value to check.
+ *
+ * Sets @mas->index and @mas->last to the range.
+ * Must hold rcu_read_lock or the write lock.
+ * Will reset mas to ma_start if the node is ma_none. Will stop on not
+ * searchable nodes.
+ *
+ * Return: the previous value or %NULL.
+ */
+void *mas_prev_range(struct ma_state *mas, unsigned long min)
+{
+ void *entry = NULL;
+
+ mas_may_init_lock_check(mas);
+ if (mas_prev_setup(mas, min, &entry))
+ return entry;
+
+ return mas_prev_slot(mas, min, true);
+}
+EXPORT_SYMBOL_GPL(mas_prev_range);
+
+/**
+ * mt_prev() - get the previous value in the maple tree
+ * @mt: The maple tree
+ * @index: The start index
+ * @min: The minimum index to check
+ *
+ * Takes RCU read lock internally to protect the search, which does not
+ * protect the returned pointer after dropping RCU read lock.
+ * See also: Documentation/core-api/maple_tree.rst
+ *
+ * Return: The entry before @index or %NULL if nothing is found.
+ */
+void *mt_prev(struct maple_tree *mt, unsigned long index, unsigned long min)
+{
+ void *entry = NULL;
+ MA_STATE(mas, mt, index, index);
+
+ rcu_read_lock();
+ entry = mas_prev(&mas, min);
+ rcu_read_unlock();
+ return entry;
+}
+EXPORT_SYMBOL_GPL(mt_prev);
+
+/**
+ * mas_pause() - Pause a mas_find/mas_for_each to drop the lock.
+ * @mas: The maple state to pause
+ *
+ * Some users need to pause a walk and drop the lock they're holding in
+ * order to yield to a higher priority thread or carry out an operation
+ * on an entry. Those users should call this function before they drop
+ * the lock. It resets the @mas to be suitable for the next iteration
+ * of the loop after the user has reacquired the lock. If most entries
+ * found during a walk require you to call mas_pause(), the mt_for_each()
+ * iterator may be more appropriate.
+ *
+ */
+void mas_pause(struct ma_state *mas)
+{
+ mas_lock_check(mas);
+ mas->status = ma_pause;
+ mas->node = NULL;
+}
+EXPORT_SYMBOL_GPL(mas_pause);
+
+/**
+ * mas_find_setup() - Internal function to set up mas_find*().
+ * @mas: The maple state
+ * @max: The maximum index
+ * @entry: Pointer to the entry
+ *
+ * Returns: True if entry is the answer, false otherwise.
+ */
+static __always_inline bool mas_find_setup(struct ma_state *mas, unsigned long max, void **entry)
+{
+ switch (mas->status) {
+ case ma_active:
+ if (mas->last < max)
+ return false;
+ return true;
+ case ma_start:
+ break;
+ case ma_pause:
+ if (unlikely(mas->last >= max))
+ return true;
+
+ mas->index = ++mas->last;
+ mas->status = ma_start;
+ break;
+ case ma_none:
+ if (unlikely(mas->last >= max))
+ return true;
+
+ mas->index = mas->last;
+ mas->status = ma_start;
+ break;
+ case ma_underflow:
+ /* mas is pointing at entry before unable to go lower */
+ if (unlikely(mas->index >= max)) {
+ mas->status = ma_overflow;
+ return true;
+ }
+
+ mas_may_activate(mas);
+ *entry = mas_walk(mas);
+ if (*entry)
+ return true;
+ break;
+ case ma_overflow:
+ if (unlikely(mas->last >= max))
+ return true;
+
+ mas_may_activate(mas);
+ *entry = mas_walk(mas);
+ if (*entry)
+ return true;
+ break;
+ case ma_root:
+ break;
+ case ma_error:
+ return true;
+ }
+
+ if (mas_is_start(mas)) {
+ /* First run or continue */
+ if (mas->index > max)
+ return true;
+
+ *entry = mas_walk(mas);
+ if (*entry)
+ return true;
+
+ }
+
+ if (unlikely(mas_is_ptr(mas)))
+ goto ptr_out_of_range;
+
+ if (unlikely(mas_is_none(mas)))
+ return true;
+
+ if (mas->index == max)
+ return true;
+
+ return false;
+
+ptr_out_of_range:
+ mas->status = ma_none;
+ mas->index = 1;
+ mas->last = ULONG_MAX;
+ return true;
+}
+
+/**
+ * mas_find() - On the first call, find the entry at or after mas->index up to
+ * %max. Otherwise, find the entry after mas->index.
+ * @mas: The maple state
+ * @max: The maximum value to check.
+ *
+ * Must hold rcu_read_lock or the write lock.
+ * If an entry exists, last and index are updated accordingly.
+ * May set @mas->status to ma_overflow.
+ *
+ * Return: The entry or %NULL.
+ */
+void *mas_find(struct ma_state *mas, unsigned long max)
+{
+ void *entry = NULL;
+
+ mas_may_init_lock_check(mas);
+ if (mas_find_setup(mas, max, &entry))
+ return entry;
+
+ /* Retries on dead nodes handled by mas_next_slot */
+ entry = mas_next_slot(mas, max, false);
+ /* Ignore overflow */
+ mas->status = ma_active;
+ return entry;
+}
+EXPORT_SYMBOL_GPL(mas_find);
+
+/**
+ * mas_find_range() - On the first call, find the entry at or after
+ * mas->index up to %max. Otherwise, advance to the next slot mas->index.
+ * @mas: The maple state
+ * @max: The maximum value to check.
+ *
+ * Must hold rcu_read_lock or the write lock.
+ * If an entry exists, last and index are updated accordingly.
+ * May set @mas->status to ma_overflow.
+ *
+ * Return: The entry or %NULL.
+ */
+void *mas_find_range(struct ma_state *mas, unsigned long max)
+{
+ void *entry = NULL;
+
+ mas_may_init_lock_check(mas);
+ if (mas_find_setup(mas, max, &entry))
+ return entry;
+
+ /* Retries on dead nodes handled by mas_next_slot */
+ return mas_next_slot(mas, max, true);
+}
+EXPORT_SYMBOL_GPL(mas_find_range);
+
+/**
+ * mas_find_rev_setup() - Internal function to set up mas_find_*_rev()
+ * @mas: The maple state
+ * @min: The minimum index
+ * @entry: Pointer to the entry
+ *
+ * Returns: True if entry is the answer, false otherwise.
+ */
+static bool mas_find_rev_setup(struct ma_state *mas, unsigned long min,
+ void **entry)
+{
+ switch (mas->status) {
+ case ma_active:
+ goto active;
+ case ma_start:
+ break;
+ case ma_pause:
+ if (unlikely(mas->index <= min)) {
+ mas->status = ma_underflow;
+ return true;
+ }
+ mas->last = --mas->index;
+ mas->status = ma_start;
+ break;
+ case ma_none:
+ if (mas->index <= min)
+ goto none;
+
+ mas->last = mas->index;
+ mas->status = ma_start;
+ break;
+ case ma_overflow: /* user expects the mas to be one after where it is */
+ if (unlikely(mas->index <= min)) {
+ mas->status = ma_underflow;
+ return true;
+ }
+
+ mas->status = ma_active;
+ break;
+ case ma_underflow: /* user expects the mas to be one before where it is */
+ if (unlikely(mas->index <= min))
+ return true;
+
+ mas->status = ma_active;
+ break;
+ case ma_root:
+ break;
+ case ma_error:
+ return true;
+ }
+
+ if (mas_is_start(mas)) {
+ /* First run or continue */
+ if (mas->index < min)
+ return true;
+
+ *entry = mas_walk(mas);
+ if (*entry)
+ return true;
+ }
+
+ if (unlikely(mas_is_ptr(mas)))
+ goto none;
+
+ if (unlikely(mas_is_none(mas))) {
+ /*
+ * Walked to the location, and there was nothing so the previous
+ * location is 0.
+ */
+ mas->last = mas->index = 0;
+ mas->status = ma_root;
+ *entry = mas_root(mas);
+ return true;
+ }
+
+active:
+ if (mas->index < min)
+ return true;
+
+ return false;
+
+none:
+ mas->status = ma_none;
+ return true;
+}
+
+/**
+ * mas_find_rev: On the first call, find the first non-null entry at or below
+ * mas->index down to %min. Otherwise find the first non-null entry below
+ * mas->index down to %min.
+ * @mas: The maple state
+ * @min: The minimum value to check.
+ *
+ * Must hold rcu_read_lock or the write lock.
+ * If an entry exists, last and index are updated accordingly.
+ * May set @mas->status to ma_underflow.
+ *
+ * Return: The entry or %NULL.
+ */
+void *mas_find_rev(struct ma_state *mas, unsigned long min)
+{
+ void *entry = NULL;
+
+ mas_may_init_lock_check(mas);
+ if (mas_find_rev_setup(mas, min, &entry))
+ return entry;
+
+ /* Retries on dead nodes handled by mas_prev_slot */
+ return mas_prev_slot(mas, min, false);
+
+}
+EXPORT_SYMBOL_GPL(mas_find_rev);
+
+/**
+ * mas_find_range_rev: On the first call, find the first non-null entry at or
+ * below mas->index down to %min. Otherwise advance to the previous slot after
+ * mas->index down to %min.
+ * @mas: The maple state
+ * @min: The minimum value to check.
+ *
+ * Must hold rcu_read_lock or the write lock.
+ * If an entry exists, last and index are updated accordingly.
+ * May set @mas->status to ma_underflow.
+ *
+ * Return: The entry or %NULL.
+ */
+void *mas_find_range_rev(struct ma_state *mas, unsigned long min)
+{
+ void *entry = NULL;
+
+ mas_may_init_lock_check(mas);
+ if (mas_find_rev_setup(mas, min, &entry))
+ return entry;
+
+ /* Retries on dead nodes handled by mas_prev_slot */
+ return mas_prev_slot(mas, min, true);
+}
+EXPORT_SYMBOL_GPL(mas_find_range_rev);
+
+/**
+ * mas_erase() - Find the range in which index resides and erase the entire
+ * range.
+ * @mas: The maple state
+ *
+ * Must hold the write lock.
+ * Searches for @mas->index, sets @mas->index and @mas->last to the range and
+ * erases that range.
+ *
+ * Note that erase requires allocations and will use GFP_KERNEL | __GFP_NOFAIL
+ * to do so if necessary. If the allocation fails, the internal lock will be
+ * dropped to retry. An externally locked tree must be protected by a lock that
+ * allows blocking for this API.
+ *
+ * Return: the entry that was erased or %NULL, @mas->index and @mas->last are updated.
+ */
+void *mas_erase(struct ma_state *mas)
+{
+ void *entry;
+ unsigned long index = mas->index;
+ MA_WR_STATE(wr_mas, mas, NULL);
+
+ /*
+ * In low memory situations, the allocation is retried with the gfp flag
+ * GFP_KERNEL | __GFP_NOFAIL. The internal spinlock is dropped in
+ * mas_nomem_nofail(), however the external lock is not dropped.
+ */
+ if (mt_external_lock(mas->tree))
+ might_alloc(GFP_KERNEL);
+
+ mas_make_walkable(mas);
+write_retry:
+ entry = mas_state_walk(mas);
+ if (!entry)
+ goto out;
+
+ /* Must reset to ensure spanning writes of last slot are detected */
+ mas_reset(mas);
+ mas_wr_preallocate(&wr_mas, NULL);
+ if (mas_nomem_nofail(mas, index, index))
+ goto write_retry;
+
+ mas_wr_store_entry(&wr_mas);
+out:
+ mas_destroy(mas);
+ return entry;
+}
+EXPORT_SYMBOL_GPL(mas_erase);
+
+/**
+ * mas_nomem() - Check if there was an error allocating and do the allocation
+ * if necessary.
+ *
+ * @mas: The maple state
+ * @gfp: The GFP_FLAGS to use for allocations
+ * Return: False on no memory. True otherwise (partial success as well)
+ */
+bool mas_nomem(struct ma_state *mas, gfp_t gfp)
+ __must_hold(mas->tree->ma_lock)
+{
+ if (likely(mas->node != MA_ERROR(-ENOMEM)))
+ return false;
+
+ /* Allocations can fail, don't do this. */
+ WARN_ON_ONCE(!gfpflags_allow_blocking(gfp) &&
+ mt_external_lock(mas->tree));
+
+ if (gfpflags_allow_blocking(gfp) && !mt_external_lock(mas->tree)) {
+ mtree_unlock(mas->tree);
+ mas_alloc_nodes(mas, gfp);
+ mtree_lock(mas->tree);
+ } else {
+ mas_alloc_nodes(mas, gfp);
+ }
+
+ /*
+ * Return false on zero forward progress. Partial allocations are kept
+ * so the retry path will attempt to get the rest. The failure should
+ * not happen as we try our best to reclaim. The user would need an
+ * external lock with a non-blocking gfp in a low memory situation.
+ */
+ if (!mas->sheaf && !mas->alloc)
+ return false;
+
+ mas_reset(mas);
+ return true;
+}
+
+/**
+ * mas_nomem_nofail() - Retry allocations with __GFP_NOFAIL, if the maple state
+ * has stored the -ENOMEM error.
+ * @mas: The maple state
+ * @index: The start of the range for the @mas reset
+ * @last: The end of the range for the @mas reset
+ *
+ * Return: false if @mas isn't in an -ENOMEM state. True if the allocation
+ * happens, the state is reset. The internal lock will be dropped and external
+ * locks must allow blocking.
+ */
+bool mas_nomem_nofail(struct ma_state *mas, unsigned long index,
+ unsigned long last)
+ __must_hold(mas->tree->ma_lock)
+{
+ gfp_t gfp;
+
+ if (likely(mas->node != MA_ERROR(-ENOMEM)))
+ return false;
+
+ gfp = GFP_KERNEL | __GFP_NOFAIL;
+ if (!mt_external_lock(mas->tree)) {
+ mtree_unlock(mas->tree);
+ mas_alloc_nodes(mas, gfp);
+ mtree_lock(mas->tree);
+ } else {
+ mas_alloc_nodes(mas, gfp);
+ }
+
+ mas_set_range(mas, index, last);
+ return true;
+}
+
+void __init maple_tree_init(void)
+{
+ struct kmem_cache_args args = {
+ .align = sizeof(struct maple_node),
+ .sheaf_capacity = 32,
+ };
+
+ maple_node_cache = kmem_cache_create("maple_node",
+ sizeof(struct maple_node), &args,
+ SLAB_PANIC);
+}
+
+/**
+ * mtree_load() - Load a value stored in a maple tree
+ * @mt: The maple tree
+ * @index: The index to load
+ *
+ * Return: the entry or %NULL
+ */
+void *mtree_load(struct maple_tree *mt, unsigned long index)
+{
+ MA_STATE(mas, mt, index, index);
+ void *entry;
+
+ trace_ma_read(TP_FCT, &mas);
+ rcu_read_lock();
+retry:
+ entry = mas_start(&mas);
+ if (unlikely(mas_is_none(&mas)))
+ goto unlock;
+
+ if (unlikely(mas_is_ptr(&mas))) {
+ if (index)
+ entry = NULL;
+
+ goto unlock;
+ }
+
+ entry = mtree_lookup_walk(&mas);
+ if (!entry && unlikely(mas_is_start(&mas)))
+ goto retry;
+unlock:
+ rcu_read_unlock();
+ if (xa_is_zero(entry))
+ return NULL;
+
+ return entry;
+}
+EXPORT_SYMBOL(mtree_load);
+
+/**
+ * mtree_store_range() - Store an entry at a given range.
+ * @mt: The maple tree
+ * @index: The start of the range
+ * @last: The end of the range
+ * @entry: The entry to store
+ * @gfp: The GFP_FLAGS to use for allocations
+ *
+ * Return: 0 on success, -EINVAL on invalid request, -ENOMEM if memory could not
+ * be allocated.
+ */
+int mtree_store_range(struct maple_tree *mt, unsigned long index,
+ unsigned long last, void *entry, gfp_t gfp)
+{
+ MA_STATE(mas, mt, index, last);
+ int ret = 0;
+
+ trace_ma_write(TP_FCT, &mas, 0, entry);
+ if (WARN_ON_ONCE(xa_is_advanced(entry)))
+ return -EINVAL;
+
+ if (index > last)
+ return -EINVAL;
+
+ mtree_lock(mt);
+ ret = mas_store_gfp(&mas, entry, gfp);
+ mtree_unlock(mt);
+
+ return ret;
+}
+EXPORT_SYMBOL(mtree_store_range);
+
+/**
+ * mtree_store() - Store an entry at a given index.
+ * @mt: The maple tree
+ * @index: The index to store the value
+ * @entry: The entry to store
+ * @gfp: The GFP_FLAGS to use for allocations
+ *
+ * Return: 0 on success, -EINVAL on invalid request, -ENOMEM if memory could not
+ * be allocated.
+ */
+int mtree_store(struct maple_tree *mt, unsigned long index, void *entry,
+ gfp_t gfp)
+{
+ return mtree_store_range(mt, index, index, entry, gfp);
+}
+EXPORT_SYMBOL(mtree_store);
+
+/**
+ * mtree_insert_range() - Insert an entry from [first, last] at a given range
+ * if there is no value.
+ * @mt: The maple tree
+ * @first: The start of the range
+ * @last: The end of the range (inclusive)
+ * @entry: The entry to store
+ * @gfp: The GFP_FLAGS to use for allocations.
+ *
+ * Note that @last is inclusive. That is, @last = @first + length - 1;
+ *
+ * Return: 0 on success, -EEXISTS if the range is occupied, -EINVAL on invalid
+ * request, -ENOMEM if memory could not be allocated.
+ */
+int mtree_insert_range(struct maple_tree *mt, unsigned long first,
+ unsigned long last, void *entry, gfp_t gfp)
+{
+ MA_STATE(ms, mt, first, last);
+ int ret = 0;
+
+ if (WARN_ON_ONCE(xa_is_advanced(entry)))
+ return -EINVAL;
+
+ if (first > last)
+ return -EINVAL;
+
+ mtree_lock(mt);
+retry:
+ mas_insert(&ms, entry);
+ if (mas_nomem(&ms, gfp))
+ goto retry;
+
+ mtree_unlock(mt);
+ if (mas_is_err(&ms))
+ ret = xa_err(ms.node);
+
+ mas_destroy(&ms);
+ return ret;
+}
+EXPORT_SYMBOL(mtree_insert_range);
+
+/**
+ * mtree_insert() - Insert an entry at a given index if there is no value.
+ * @mt: The maple tree
+ * @index : The index to store the value
+ * @entry: The entry to store
+ * @gfp: The GFP_FLAGS to use for allocations.
+ *
+ * Return: 0 on success, -EEXISTS if the range is occupied, -EINVAL on invalid
+ * request, -ENOMEM if memory could not be allocated.
+ */
+int mtree_insert(struct maple_tree *mt, unsigned long index, void *entry,
+ gfp_t gfp)
+{
+ return mtree_insert_range(mt, index, index, entry, gfp);
+}
+EXPORT_SYMBOL(mtree_insert);
+
+int mtree_alloc_range(struct maple_tree *mt, unsigned long *startp,
+ void *entry, unsigned long size, unsigned long min,
+ unsigned long max, gfp_t gfp)
+{
+ int ret = 0;
+
+ MA_STATE(mas, mt, 0, 0);
+ if (!mt_is_alloc(mt))
+ return -EINVAL;
+
+ if (WARN_ON_ONCE(mt_is_reserved(entry)))
+ return -EINVAL;
+
+ mtree_lock(mt);
+retry:
+ ret = mas_empty_area(&mas, min, max, size);
+ if (ret)
+ goto unlock;
+
+ mas_insert(&mas, entry);
+ /*
+ * mas_nomem() may release the lock, causing the allocated area
+ * to be unavailable, so try to allocate a free area again.
+ */
+ if (mas_nomem(&mas, gfp))
+ goto retry;
+
+ if (mas_is_err(&mas))
+ ret = xa_err(mas.node);
+ else
+ *startp = mas.index;
+
+unlock:
+ mtree_unlock(mt);
+ mas_destroy(&mas);
+ return ret;
+}
+EXPORT_SYMBOL(mtree_alloc_range);
+
+/**
+ * mtree_alloc_cyclic() - Find somewhere to store this entry in the tree.
+ * @mt: The maple tree.
+ * @startp: Pointer to ID.
+ * @range_lo: Lower bound of range to search.
+ * @range_hi: Upper bound of range to search.
+ * @entry: The entry to store.
+ * @next: Pointer to next ID to allocate.
+ * @gfp: The GFP_FLAGS to use for allocations.
+ *
+ * Finds an empty entry in @mt after @next, stores the new index into
+ * the @id pointer, stores the entry at that index, then updates @next.
+ *
+ * @mt must be initialized with the MT_FLAGS_ALLOC_RANGE flag.
+ *
+ * Context: Any context. Takes and releases the mt.lock. May sleep if
+ * the @gfp flags permit.
+ *
+ * Return: 0 if the allocation succeeded without wrapping, 1 if the
+ * allocation succeeded after wrapping, -ENOMEM if memory could not be
+ * allocated, -EINVAL if @mt cannot be used, or -EBUSY if there are no
+ * free entries.
+ */
+int mtree_alloc_cyclic(struct maple_tree *mt, unsigned long *startp,
+ void *entry, unsigned long range_lo, unsigned long range_hi,
+ unsigned long *next, gfp_t gfp)
+{
+ int ret;
+
+ MA_STATE(mas, mt, 0, 0);
+
+ if (!mt_is_alloc(mt))
+ return -EINVAL;
+ if (WARN_ON_ONCE(mt_is_reserved(entry)))
+ return -EINVAL;
+ mtree_lock(mt);
+ ret = mas_alloc_cyclic(&mas, startp, entry, range_lo, range_hi,
+ next, gfp);
+ mtree_unlock(mt);
+ return ret;
+}
+EXPORT_SYMBOL(mtree_alloc_cyclic);
+
+int mtree_alloc_rrange(struct maple_tree *mt, unsigned long *startp,
+ void *entry, unsigned long size, unsigned long min,
+ unsigned long max, gfp_t gfp)
+{
+ int ret = 0;
+
+ MA_STATE(mas, mt, 0, 0);
+ if (!mt_is_alloc(mt))
+ return -EINVAL;
+
+ if (WARN_ON_ONCE(mt_is_reserved(entry)))
+ return -EINVAL;
+
+ mtree_lock(mt);
+retry:
+ ret = mas_empty_area_rev(&mas, min, max, size);
+ if (ret)
+ goto unlock;
+
+ mas_insert(&mas, entry);
+ /*
+ * mas_nomem() may release the lock, causing the allocated area
+ * to be unavailable, so try to allocate a free area again.
+ */
+ if (mas_nomem(&mas, gfp))
+ goto retry;
+
+ if (mas_is_err(&mas))
+ ret = xa_err(mas.node);
+ else
+ *startp = mas.index;
+
+unlock:
+ mtree_unlock(mt);
+ mas_destroy(&mas);
+ return ret;
+}
+EXPORT_SYMBOL(mtree_alloc_rrange);
+
+/**
+ * mtree_erase() - Find an index and erase the entire range.
+ * @mt: The maple tree
+ * @index: The index to erase
+ *
+ * Erasing is the same as a walk to an entry then a store of a NULL to that
+ * ENTIRE range. In fact, it is implemented as such using the advanced API.
+ *
+ * Note that erase requires allocations and will use GFP_KERNEL | __GFP_NOFAIL
+ * to do so if necessary. If the allocation fails, the internal lock will be
+ * dropped to retry.
+ *
+ * Return: The entry stored at the @index or %NULL
+ */
+void *mtree_erase(struct maple_tree *mt, unsigned long index)
+{
+ void *entry = NULL;
+
+ MA_STATE(mas, mt, index, index);
+ trace_ma_op(TP_FCT, &mas);
+
+ might_alloc(GFP_KERNEL);
+ mtree_lock(mt);
+ entry = mas_erase(&mas);
+ mtree_unlock(mt);
+
+ return entry;
+}
+EXPORT_SYMBOL(mtree_erase);
+
+/*
+ * mas_dup_free() - Free an incomplete duplication of a tree.
+ * @mas: The maple state of a incomplete tree.
+ *
+ * The parameter @mas->node passed in indicates that the allocation failed on
+ * this node. This function frees all nodes starting from @mas->node in the
+ * reverse order of mas_dup_build(). There is no need to hold the source tree
+ * lock at this time.
+ */
+static void mas_dup_free(struct ma_state *mas)
+{
+ struct maple_node *node;
+ enum maple_type type;
+ void __rcu **slots;
+ unsigned char count, i;
+
+ /* Maybe the first node allocation failed. */
+ if (mas_is_none(mas))
+ return;
+
+ while (!mte_is_root(mas->node)) {
+ mas_ascend(mas);
+ if (mas->offset) {
+ mas->offset--;
+ do {
+ mas_descend(mas);
+ mas->offset = mas_data_end(mas);
+ } while (!mte_is_leaf(mas->node));
+
+ mas_ascend(mas);
+ }
+
+ node = mte_to_node(mas->node);
+ type = mte_node_type(mas->node);
+ slots = ma_slots(node, type);
+ count = mas_data_end(mas) + 1;
+ for (i = 0; i < count; i++)
+ ((unsigned long *)slots)[i] &= ~MAPLE_NODE_MASK;
+ mt_free_bulk(count, slots);
+ }
+
+ node = mte_to_node(mas->node);
+ kfree(node);
+}
+
+/*
+ * mas_copy_node() - Copy a maple node and replace the parent.
+ * @mas: The maple state of source tree.
+ * @new_mas: The maple state of new tree.
+ * @parent: The parent of the new node.
+ *
+ * Copy @mas->node to @new_mas->node, set @parent to be the parent of
+ * @new_mas->node. If memory allocation fails, @mas is set to -ENOMEM.
+ */
+static inline void mas_copy_node(struct ma_state *mas, struct ma_state *new_mas,
+ struct maple_pnode *parent)
+{
+ struct maple_node *node = mte_to_node(mas->node);
+ struct maple_node *new_node = mte_to_node(new_mas->node);
+ unsigned long val;
+
+ /* Copy the node completely. */
+ memcpy(new_node, node, sizeof(struct maple_node));
+ /* Update the parent node pointer. */
+ val = (unsigned long)node->parent & MAPLE_NODE_MASK;
+ new_node->parent = ma_parent_ptr(val | (unsigned long)parent);
+}
+
+/*
+ * mas_dup_alloc() - Allocate child nodes for a maple node.
+ * @mas: The maple state of source tree.
+ * @new_mas: The maple state of new tree.
+ * @gfp: The GFP_FLAGS to use for allocations.
+ *
+ * This function allocates child nodes for @new_mas->node during the duplication
+ * process. If memory allocation fails, @mas is set to -ENOMEM.
+ */
+static inline void mas_dup_alloc(struct ma_state *mas, struct ma_state *new_mas,
+ gfp_t gfp)
+{
+ struct maple_node *node = mte_to_node(mas->node);
+ struct maple_node *new_node = mte_to_node(new_mas->node);
+ enum maple_type type;
+ unsigned char count, i;
+ void __rcu **slots;
+ void __rcu **new_slots;
+ unsigned long val;
+
+ /* Allocate memory for child nodes. */
+ type = mte_node_type(mas->node);
+ new_slots = ma_slots(new_node, type);
+ count = mas->node_request = mas_data_end(mas) + 1;
+ mas_alloc_nodes(mas, gfp);
+ if (unlikely(mas_is_err(mas)))
+ return;
+
+ slots = ma_slots(node, type);
+ for (i = 0; i < count; i++) {
+ val = (unsigned long)mt_slot_locked(mas->tree, slots, i);
+ val &= MAPLE_NODE_MASK;
+ /*
+ * Warning, see rcu_assign_pointer() documentation. Since this
+ * is a duplication of a tree, there are no readers walking the
+ * tree until after the rcu_assign_pointer() call in
+ * mas_dup_build().
+ */
+ RCU_INIT_POINTER(new_slots[i],
+ ma_mnode_ptr((unsigned long)mas_pop_node(mas) |
+ val));
+ }
+}
+
+/*
+ * mas_dup_build() - Build a new maple tree from a source tree
+ * @mas: The maple state of source tree, need to be in MAS_START state.
+ * @new_mas: The maple state of new tree, need to be in MAS_START state.
+ * @gfp: The GFP_FLAGS to use for allocations.
+ *
+ * This function builds a new tree in DFS preorder. If the memory allocation
+ * fails, the error code -ENOMEM will be set in @mas, and @new_mas points to the
+ * last node. mas_dup_free() will free the incomplete duplication of a tree.
+ *
+ * Note that the attributes of the two trees need to be exactly the same, and the
+ * new tree needs to be empty, otherwise -EINVAL will be set in @mas.
+ */
+static inline void mas_dup_build(struct ma_state *mas, struct ma_state *new_mas,
+ gfp_t gfp)
+{
+ struct maple_node *node;
+ struct maple_pnode *parent = NULL;
+ struct maple_enode *root;
+ enum maple_type type;
+
+ if (unlikely(mt_attr(mas->tree) != mt_attr(new_mas->tree)) ||
+ unlikely(!mtree_empty(new_mas->tree))) {
+ mas_set_err(mas, -EINVAL);
+ return;
+ }
+
+ root = mas_start(mas);
+ if (mas_is_ptr(mas) || mas_is_none(mas))
+ goto set_new_tree;
+
+ node = mt_alloc_one(gfp);
+ if (!node) {
+ new_mas->status = ma_none;
+ mas_set_err(mas, -ENOMEM);
+ return;
+ }
+
+ type = mte_node_type(mas->node);
+ root = mt_mk_node(node, type);
+ new_mas->node = root;
+ new_mas->min = 0;
+ new_mas->max = ULONG_MAX;
+ root = mte_mk_root(root);
+ while (1) {
+ mas_copy_node(mas, new_mas, parent);
+ if (!mte_is_leaf(mas->node)) {
+ /* Only allocate child nodes for non-leaf nodes. */
+ mas_dup_alloc(mas, new_mas, gfp);
+ if (unlikely(mas_is_err(mas)))
+ goto empty_mas;
+ } else {
+ /*
+ * This is the last leaf node and duplication is
+ * completed.
+ */
+ if (mas->max == ULONG_MAX)
+ goto done;
+
+ /* This is not the last leaf node and needs to go up. */
+ do {
+ mas_ascend(mas);
+ mas_ascend(new_mas);
+ } while (mas->offset == mas_data_end(mas));
+
+ /* Move to the next subtree. */
+ mas->offset++;
+ new_mas->offset++;
+ }
+
+ mas_descend(mas);
+ parent = ma_parent_ptr(mte_to_node(new_mas->node));
+ mas_descend(new_mas);
+ mas->offset = 0;
+ new_mas->offset = 0;
+ }
+done:
+ /* Specially handle the parent of the root node. */
+ mte_to_node(root)->parent = ma_parent_ptr(mas_tree_parent(new_mas));
+set_new_tree:
+ /* Make them the same height */
+ new_mas->tree->ma_flags = mas->tree->ma_flags;
+ rcu_assign_pointer(new_mas->tree->ma_root, root);
+empty_mas:
+ mas_empty_nodes(mas);
+}
+
+/**
+ * __mt_dup(): Duplicate an entire maple tree
+ * @mt: The source maple tree
+ * @new: The new maple tree
+ * @gfp: The GFP_FLAGS to use for allocations
+ *
+ * This function duplicates a maple tree in Depth-First Search (DFS) pre-order
+ * traversal. It uses memcpy() to copy nodes in the source tree and allocate
+ * new child nodes in non-leaf nodes. The new node is exactly the same as the
+ * source node except for all the addresses stored in it. It will be faster than
+ * traversing all elements in the source tree and inserting them one by one into
+ * the new tree.
+ * The user needs to ensure that the attributes of the source tree and the new
+ * tree are the same, and the new tree needs to be an empty tree, otherwise
+ * -EINVAL will be returned.
+ * Note that the user needs to manually lock the source tree and the new tree.
+ *
+ * Return: 0 on success, -ENOMEM if memory could not be allocated, -EINVAL If
+ * the attributes of the two trees are different or the new tree is not an empty
+ * tree.
+ */
+int __mt_dup(struct maple_tree *mt, struct maple_tree *new, gfp_t gfp)
+{
+ int ret = 0;
+ MA_STATE(mas, mt, 0, 0);
+ MA_STATE(new_mas, new, 0, 0);
+
+ mas_dup_build(&mas, &new_mas, gfp);
+ if (unlikely(mas_is_err(&mas))) {
+ ret = xa_err(mas.node);
+ if (ret == -ENOMEM)
+ mas_dup_free(&new_mas);
+ }
+
+ return ret;
+}
+EXPORT_SYMBOL(__mt_dup);
+
+/**
+ * mtree_dup(): Duplicate an entire maple tree
+ * @mt: The source maple tree
+ * @new: The new maple tree
+ * @gfp: The GFP_FLAGS to use for allocations
+ *
+ * This function duplicates a maple tree in Depth-First Search (DFS) pre-order
+ * traversal. It uses memcpy() to copy nodes in the source tree and allocate
+ * new child nodes in non-leaf nodes. The new node is exactly the same as the
+ * source node except for all the addresses stored in it. It will be faster than
+ * traversing all elements in the source tree and inserting them one by one into
+ * the new tree.
+ * The user needs to ensure that the attributes of the source tree and the new
+ * tree are the same, and the new tree needs to be an empty tree, otherwise
+ * -EINVAL will be returned.
+ *
+ * Return: 0 on success, -ENOMEM if memory could not be allocated, -EINVAL If
+ * the attributes of the two trees are different or the new tree is not an empty
+ * tree.
+ */
+int mtree_dup(struct maple_tree *mt, struct maple_tree *new, gfp_t gfp)
+{
+ int ret = 0;
+ MA_STATE(mas, mt, 0, 0);
+ MA_STATE(new_mas, new, 0, 0);
+
+ mas_lock(&new_mas);
+ mas_lock_nested(&mas, SINGLE_DEPTH_NESTING);
+ mas_dup_build(&mas, &new_mas, gfp);
+ mas_unlock(&mas);
+ if (unlikely(mas_is_err(&mas))) {
+ ret = xa_err(mas.node);
+ if (ret == -ENOMEM)
+ mas_dup_free(&new_mas);
+ }
+
+ mas_unlock(&new_mas);
+ return ret;
+}
+EXPORT_SYMBOL(mtree_dup);
+
+/**
+ * __mt_destroy() - Walk and free all nodes of a locked maple tree.
+ * @mt: The maple tree
+ *
+ * Note: Does not handle locking.
+ */
+void __mt_destroy(struct maple_tree *mt)
+{
+ void *root = mt_root_locked(mt);
+
+ rcu_assign_pointer(mt->ma_root, NULL);
+ if (xa_is_node(root))
+ mte_destroy_walk(root, mt);
+
+ mt->ma_flags = mt_attr(mt);
+}
+EXPORT_SYMBOL_GPL(__mt_destroy);
+
+/**
+ * mtree_destroy() - Destroy a maple tree
+ * @mt: The maple tree
+ *
+ * Frees all resources used by the tree. Handles locking.
+ */
+void mtree_destroy(struct maple_tree *mt)
+{
+ mtree_lock(mt);
+ __mt_destroy(mt);
+ mtree_unlock(mt);
+}
+EXPORT_SYMBOL(mtree_destroy);
+
+/**
+ * mt_find() - Search from the start up until an entry is found.
+ * @mt: The maple tree
+ * @index: Pointer which contains the start location of the search
+ * @max: The maximum value of the search range
+ *
+ * Takes RCU read lock internally to protect the search, which does not
+ * protect the returned pointer after dropping RCU read lock.
+ * See also: Documentation/core-api/maple_tree.rst
+ *
+ * In case that an entry is found @index is updated to point to the next
+ * possible entry independent whether the found entry is occupying a
+ * single index or a range if indices.
+ *
+ * Return: The entry at or after the @index or %NULL
+ */
+void *mt_find(struct maple_tree *mt, unsigned long *index, unsigned long max)
+{
+ MA_STATE(mas, mt, *index, *index);
+ void *entry;
+#ifdef CONFIG_DEBUG_MAPLE_TREE
+ unsigned long copy = *index;
+#endif
+
+ trace_ma_read(TP_FCT, &mas);
+
+ if ((*index) > max)
+ return NULL;
+
+ rcu_read_lock();
+retry:
+ entry = mas_state_walk(&mas);
+ if (mas_is_start(&mas))
+ goto retry;
+
+ if (unlikely(xa_is_zero(entry)))
+ entry = NULL;
+
+ if (entry)
+ goto unlock;
+
+ while (mas_is_active(&mas) && (mas.last < max)) {
+ entry = mas_next_slot(&mas, max, false);
+ if (likely(entry && !xa_is_zero(entry)))
+ break;
+ }
+
+ if (unlikely(xa_is_zero(entry)))
+ entry = NULL;
+unlock:
+ rcu_read_unlock();
+ if (likely(entry)) {
+ *index = mas.last + 1;
+#ifdef CONFIG_DEBUG_MAPLE_TREE
+ if (MT_WARN_ON(mt, (*index) && ((*index) <= copy)))
+ pr_err("index not increased! %lx <= %lx\n",
+ *index, copy);
+#endif
+ }
+
+ return entry;
+}
+EXPORT_SYMBOL(mt_find);
+
+/**
+ * mt_find_after() - Search from the start up until an entry is found.
+ * @mt: The maple tree
+ * @index: Pointer which contains the start location of the search
+ * @max: The maximum value to check
+ *
+ * Same as mt_find() except that it checks @index for 0 before
+ * searching. If @index == 0, the search is aborted. This covers a wrap
+ * around of @index to 0 in an iterator loop.
+ *
+ * Return: The entry at or after the @index or %NULL
+ */
+void *mt_find_after(struct maple_tree *mt, unsigned long *index,
+ unsigned long max)
+{
+ if (!(*index))
+ return NULL;
+
+ return mt_find(mt, index, max);
+}
+EXPORT_SYMBOL(mt_find_after);
+
+#ifdef CONFIG_DEBUG_MAPLE_TREE
+atomic_t maple_tree_tests_run;
+EXPORT_SYMBOL_GPL(maple_tree_tests_run);
+atomic_t maple_tree_tests_passed;
+EXPORT_SYMBOL_GPL(maple_tree_tests_passed);
+
+#ifndef __KERNEL__
+extern void kmem_cache_set_non_kernel(struct kmem_cache *, unsigned int);
+void mt_set_non_kernel(unsigned int val)
+{
+ kmem_cache_set_non_kernel(maple_node_cache, val);
+}
+
+extern void kmem_cache_set_callback(struct kmem_cache *cachep,
+ void (*callback)(void *));
+void mt_set_callback(void (*callback)(void *))
+{
+ kmem_cache_set_callback(maple_node_cache, callback);
+}
+
+extern void kmem_cache_set_private(struct kmem_cache *cachep, void *private);
+void mt_set_private(void *private)
+{
+ kmem_cache_set_private(maple_node_cache, private);
+}
+
+extern unsigned long kmem_cache_get_alloc(struct kmem_cache *);
+unsigned long mt_get_alloc_size(void)
+{
+ return kmem_cache_get_alloc(maple_node_cache);
+}
+
+extern void kmem_cache_zero_nr_tallocated(struct kmem_cache *);
+void mt_zero_nr_tallocated(void)
+{
+ kmem_cache_zero_nr_tallocated(maple_node_cache);
+}
+
+extern unsigned int kmem_cache_nr_tallocated(struct kmem_cache *);
+unsigned int mt_nr_tallocated(void)
+{
+ return kmem_cache_nr_tallocated(maple_node_cache);
+}
+
+extern unsigned int kmem_cache_nr_allocated(struct kmem_cache *);
+unsigned int mt_nr_allocated(void)
+{
+ return kmem_cache_nr_allocated(maple_node_cache);
+}
+
+void mt_cache_shrink(void)
+{
+}
+#else
+/*
+ * mt_cache_shrink() - For testing, don't use this.
+ *
+ * Certain testcases can trigger an OOM when combined with other memory
+ * debugging configuration options. This function is used to reduce the
+ * possibility of an out of memory even due to kmem_cache objects remaining
+ * around for longer than usual.
+ */
+void mt_cache_shrink(void)
+{
+ kmem_cache_shrink(maple_node_cache);
+
+}
+EXPORT_SYMBOL_GPL(mt_cache_shrink);
+
+#endif /* not defined __KERNEL__ */
+/*
+ * mas_get_slot() - Get the entry in the maple state node stored at @offset.
+ * @mas: The maple state
+ * @offset: The offset into the slot array to fetch.
+ *
+ * Return: The entry stored at @offset.
+ */
+static inline struct maple_enode *mas_get_slot(struct ma_state *mas,
+ unsigned char offset)
+{
+ return mas_slot(mas, ma_slots(mas_mn(mas), mte_node_type(mas->node)),
+ offset);
+}
+
+/* Depth first search, post-order */
+static void mas_dfs_postorder(struct ma_state *mas, unsigned long max)
+{
+
+ struct maple_enode *p, *mn = mas->node;
+ unsigned long p_min, p_max;
+
+ mas_next_node(mas, mas_mn(mas), max);
+ if (!mas_is_overflow(mas))
+ return;
+
+ if (mte_is_root(mn))
+ return;
+
+ mas->node = mn;
+ mas_ascend(mas);
+ do {
+ p = mas->node;
+ p_min = mas->min;
+ p_max = mas->max;
+ mas_prev_node(mas, 0);
+ } while (!mas_is_underflow(mas));
+
+ mas->node = p;
+ mas->max = p_max;
+ mas->min = p_min;
+}
+
+/* Tree validations */
+static void mt_dump_node(const struct maple_tree *mt, void *entry,
+ unsigned long min, unsigned long max, unsigned int depth,
+ enum mt_dump_format format);
+static void mt_dump_range(unsigned long min, unsigned long max,
+ unsigned int depth, enum mt_dump_format format)
+{
+ static const char spaces[] = " ";
+
+ switch (format) {
+ case mt_dump_hex:
+ if (min == max)
+ pr_info("%.*s%lx: ", depth * 2, spaces, min);
+ else
+ pr_info("%.*s%lx-%lx: ", depth * 2, spaces, min, max);
+ break;
+ case mt_dump_dec:
+ if (min == max)
+ pr_info("%.*s%lu: ", depth * 2, spaces, min);
+ else
+ pr_info("%.*s%lu-%lu: ", depth * 2, spaces, min, max);
+ }
+}
+
+static void mt_dump_entry(void *entry, unsigned long min, unsigned long max,
+ unsigned int depth, enum mt_dump_format format)
+{
+ mt_dump_range(min, max, depth, format);
+
+ if (xa_is_value(entry))
+ pr_cont("value %ld (0x%lx) [" PTR_FMT "]\n", xa_to_value(entry),
+ xa_to_value(entry), entry);
+ else if (xa_is_zero(entry))
+ pr_cont("zero (%ld)\n", xa_to_internal(entry));
+ else if (mt_is_reserved(entry))
+ pr_cont("UNKNOWN ENTRY (" PTR_FMT ")\n", entry);
+ else
+ pr_cont(PTR_FMT "\n", entry);
+}
+
+static void mt_dump_range64(const struct maple_tree *mt, void *entry,
+ unsigned long min, unsigned long max, unsigned int depth,
+ enum mt_dump_format format)
+{
+ struct maple_range_64 *node = &mte_to_node(entry)->mr64;
+ bool leaf = mte_is_leaf(entry);
+ unsigned long first = min;
+ int i;
+
+ pr_cont(" contents: ");
+ for (i = 0; i < MAPLE_RANGE64_SLOTS - 1; i++) {
+ switch (format) {
+ case mt_dump_hex:
+ pr_cont(PTR_FMT " %lX ", node->slot[i], node->pivot[i]);
+ break;
+ case mt_dump_dec:
+ pr_cont(PTR_FMT " %lu ", node->slot[i], node->pivot[i]);
+ }
+ }
+ pr_cont(PTR_FMT "\n", node->slot[i]);
+ for (i = 0; i < MAPLE_RANGE64_SLOTS; i++) {
+ unsigned long last = max;
+
+ if (i < (MAPLE_RANGE64_SLOTS - 1))
+ last = node->pivot[i];
+ else if (!node->slot[i] && max != mt_node_max(entry))
+ break;
+ if (last == 0 && i > 0)
+ break;
+ if (leaf)
+ mt_dump_entry(mt_slot(mt, node->slot, i),
+ first, last, depth + 1, format);
+ else if (node->slot[i])
+ mt_dump_node(mt, mt_slot(mt, node->slot, i),
+ first, last, depth + 1, format);
+
+ if (last == max)
+ break;
+ if (last > max) {
+ switch (format) {
+ case mt_dump_hex:
+ pr_err("node " PTR_FMT " last (%lx) > max (%lx) at pivot %d!\n",
+ node, last, max, i);
+ break;
+ case mt_dump_dec:
+ pr_err("node " PTR_FMT " last (%lu) > max (%lu) at pivot %d!\n",
+ node, last, max, i);
+ }
+ }
+ first = last + 1;
+ }
+}
+
+static void mt_dump_arange64(const struct maple_tree *mt, void *entry,
+ unsigned long min, unsigned long max, unsigned int depth,
+ enum mt_dump_format format)
+{
+ struct maple_arange_64 *node = &mte_to_node(entry)->ma64;
+ unsigned long first = min;
+ int i;
+
+ pr_cont(" contents: ");
+ for (i = 0; i < MAPLE_ARANGE64_SLOTS; i++) {
+ switch (format) {
+ case mt_dump_hex:
+ pr_cont("%lx ", node->gap[i]);
+ break;
+ case mt_dump_dec:
+ pr_cont("%lu ", node->gap[i]);
+ }
+ }
+ pr_cont("| %02X %02X| ", node->meta.end, node->meta.gap);
+ for (i = 0; i < MAPLE_ARANGE64_SLOTS - 1; i++) {
+ switch (format) {
+ case mt_dump_hex:
+ pr_cont(PTR_FMT " %lX ", node->slot[i], node->pivot[i]);
+ break;
+ case mt_dump_dec:
+ pr_cont(PTR_FMT " %lu ", node->slot[i], node->pivot[i]);
+ }
+ }
+ pr_cont(PTR_FMT "\n", node->slot[i]);
+ for (i = 0; i < MAPLE_ARANGE64_SLOTS; i++) {
+ unsigned long last = max;
+
+ if (i < (MAPLE_ARANGE64_SLOTS - 1))
+ last = node->pivot[i];
+ else if (!node->slot[i])
+ break;
+ if (last == 0 && i > 0)
+ break;
+ if (node->slot[i])
+ mt_dump_node(mt, mt_slot(mt, node->slot, i),
+ first, last, depth + 1, format);
+
+ if (last == max)
+ break;
+ if (last > max) {
+ switch (format) {
+ case mt_dump_hex:
+ pr_err("node " PTR_FMT " last (%lx) > max (%lx) at pivot %d!\n",
+ node, last, max, i);
+ break;
+ case mt_dump_dec:
+ pr_err("node " PTR_FMT " last (%lu) > max (%lu) at pivot %d!\n",
+ node, last, max, i);
+ }
+ }
+ first = last + 1;
+ }
+}
+
+static void mt_dump_node(const struct maple_tree *mt, void *entry,
+ unsigned long min, unsigned long max, unsigned int depth,
+ enum mt_dump_format format)
+{
+ struct maple_node *node = mte_to_node(entry);
+ unsigned int type = mte_node_type(entry);
+ unsigned int i;
+
+ mt_dump_range(min, max, depth, format);
+
+ pr_cont("node " PTR_FMT " depth %d type %d parent " PTR_FMT, node,
+ depth, type, node ? node->parent : NULL);
+ switch (type) {
+ case maple_dense:
+ pr_cont("\n");
+ for (i = 0; i < MAPLE_NODE_SLOTS; i++) {
+ if (min + i > max)
+ pr_cont("OUT OF RANGE: ");
+ mt_dump_entry(mt_slot(mt, node->slot, i),
+ min + i, min + i, depth, format);
+ }
+ break;
+ case maple_leaf_64:
+ case maple_range_64:
+ mt_dump_range64(mt, entry, min, max, depth, format);
+ break;
+ case maple_arange_64:
+ mt_dump_arange64(mt, entry, min, max, depth, format);
+ break;
+
+ default:
+ pr_cont(" UNKNOWN TYPE\n");
+ }
+}
+
+void mt_dump(const struct maple_tree *mt, enum mt_dump_format format)
+{
+ void *entry = rcu_dereference_check(mt->ma_root, mt_locked(mt));
+
+ pr_info("maple_tree(" PTR_FMT ") flags %X, height %u root " PTR_FMT "\n",
+ mt, mt->ma_flags, mt_height(mt), entry);
+ if (xa_is_node(entry))
+ mt_dump_node(mt, entry, 0, mt_node_max(entry), 0, format);
+ else if (entry)
+ mt_dump_entry(entry, 0, 0, 0, format);
+ else
+ pr_info("(empty)\n");
+}
+EXPORT_SYMBOL_GPL(mt_dump);
+
+/*
+ * Calculate the maximum gap in a node and check if that's what is reported in
+ * the parent (unless root).
+ */
+static void mas_validate_gaps(struct ma_state *mas)
+{
+ struct maple_enode *mte = mas->node;
+ struct maple_node *p_mn, *node = mte_to_node(mte);
+ enum maple_type mt = mte_node_type(mas->node);
+ unsigned long gap = 0, max_gap = 0;
+ unsigned long p_end, p_start = mas->min;
+ unsigned char p_slot, offset;
+ unsigned long *gaps = NULL;
+ unsigned long *pivots = ma_pivots(node, mt);
+ unsigned int i;
+
+ if (ma_is_dense(mt)) {
+ for (i = 0; i < mt_slot_count(mte); i++) {
+ if (mas_get_slot(mas, i)) {
+ if (gap > max_gap)
+ max_gap = gap;
+ gap = 0;
+ continue;
+ }
+ gap++;
+ }
+ goto counted;
+ }
+
+ gaps = ma_gaps(node, mt);
+ for (i = 0; i < mt_slot_count(mte); i++) {
+ p_end = mas_safe_pivot(mas, pivots, i, mt);
+
+ if (!gaps) {
+ if (!mas_get_slot(mas, i))
+ gap = p_end - p_start + 1;
+ } else {
+ void *entry = mas_get_slot(mas, i);
+
+ gap = gaps[i];
+ MT_BUG_ON(mas->tree, !entry);
+
+ if (gap > p_end - p_start + 1) {
+ pr_err(PTR_FMT "[%u] %lu >= %lu - %lu + 1 (%lu)\n",
+ mas_mn(mas), i, gap, p_end, p_start,
+ p_end - p_start + 1);
+ MT_BUG_ON(mas->tree, gap > p_end - p_start + 1);
+ }
+ }
+
+ if (gap > max_gap)
+ max_gap = gap;
+
+ p_start = p_end + 1;
+ if (p_end >= mas->max)
+ break;
+ }
+
+counted:
+ if (mt == maple_arange_64) {
+ MT_BUG_ON(mas->tree, !gaps);
+ offset = ma_meta_gap(node);
+ if (offset > i) {
+ pr_err("gap offset " PTR_FMT "[%u] is invalid\n", node, offset);
+ MT_BUG_ON(mas->tree, 1);
+ }
+
+ if (gaps[offset] != max_gap) {
+ pr_err("gap " PTR_FMT "[%u] is not the largest gap %lu\n",
+ node, offset, max_gap);
+ MT_BUG_ON(mas->tree, 1);
+ }
+
+ for (i++ ; i < mt_slot_count(mte); i++) {
+ if (gaps[i] != 0) {
+ pr_err("gap " PTR_FMT "[%u] beyond node limit != 0\n",
+ node, i);
+ MT_BUG_ON(mas->tree, 1);
+ }
+ }
+ }
+
+ if (mte_is_root(mte))
+ return;
+
+ p_slot = mte_parent_slot(mas->node);
+ p_mn = mte_parent(mte);
+ MT_BUG_ON(mas->tree, max_gap > mas->max);
+ if (ma_gaps(p_mn, mas_parent_type(mas, mte))[p_slot] != max_gap) {
+ pr_err("gap " PTR_FMT "[%u] != %lu\n", p_mn, p_slot, max_gap);
+ mt_dump(mas->tree, mt_dump_hex);
+ MT_BUG_ON(mas->tree, 1);
+ }
+}
+
+static void mas_validate_parent_slot(struct ma_state *mas)
+{
+ struct maple_node *parent;
+ struct maple_enode *node;
+ enum maple_type p_type;
+ unsigned char p_slot;
+ void __rcu **slots;
+ int i;
+
+ if (mte_is_root(mas->node))
+ return;
+
+ p_slot = mte_parent_slot(mas->node);
+ p_type = mas_parent_type(mas, mas->node);
+ parent = mte_parent(mas->node);
+ slots = ma_slots(parent, p_type);
+ MT_BUG_ON(mas->tree, mas_mn(mas) == parent);
+
+ /* Check prev/next parent slot for duplicate node entry */
+
+ for (i = 0; i < mt_slots[p_type]; i++) {
+ node = mas_slot(mas, slots, i);
+ if (i == p_slot) {
+ if (node != mas->node)
+ pr_err("parent " PTR_FMT "[%u] does not have " PTR_FMT "\n",
+ parent, i, mas_mn(mas));
+ MT_BUG_ON(mas->tree, node != mas->node);
+ } else if (node == mas->node) {
+ pr_err("Invalid child " PTR_FMT " at parent " PTR_FMT "[%u] p_slot %u\n",
+ mas_mn(mas), parent, i, p_slot);
+ MT_BUG_ON(mas->tree, node == mas->node);
+ }
+ }
+}
+
+static void mas_validate_child_slot(struct ma_state *mas)
+{
+ enum maple_type type = mte_node_type(mas->node);
+ void __rcu **slots = ma_slots(mte_to_node(mas->node), type);
+ unsigned long *pivots = ma_pivots(mte_to_node(mas->node), type);
+ struct maple_enode *child;
+ unsigned char i;
+
+ if (mte_is_leaf(mas->node))
+ return;
+
+ for (i = 0; i < mt_slots[type]; i++) {
+ child = mas_slot(mas, slots, i);
+
+ if (!child) {
+ pr_err("Non-leaf node lacks child at " PTR_FMT "[%u]\n",
+ mas_mn(mas), i);
+ MT_BUG_ON(mas->tree, 1);
+ }
+
+ if (mte_parent_slot(child) != i) {
+ pr_err("Slot error at " PTR_FMT "[%u]: child " PTR_FMT " has pslot %u\n",
+ mas_mn(mas), i, mte_to_node(child),
+ mte_parent_slot(child));
+ MT_BUG_ON(mas->tree, 1);
+ }
+
+ if (mte_parent(child) != mte_to_node(mas->node)) {
+ pr_err("child " PTR_FMT " has parent " PTR_FMT " not " PTR_FMT "\n",
+ mte_to_node(child), mte_parent(child),
+ mte_to_node(mas->node));
+ MT_BUG_ON(mas->tree, 1);
+ }
+
+ if (i < mt_pivots[type] && pivots[i] == mas->max)
+ break;
+ }
+}
+
+/*
+ * Validate all pivots are within mas->min and mas->max, check metadata ends
+ * where the maximum ends and ensure there is no slots or pivots set outside of
+ * the end of the data.
+ */
+static void mas_validate_limits(struct ma_state *mas)
+{
+ int i;
+ unsigned long prev_piv = 0;
+ enum maple_type type = mte_node_type(mas->node);
+ void __rcu **slots = ma_slots(mte_to_node(mas->node), type);
+ unsigned long *pivots = ma_pivots(mas_mn(mas), type);
+
+ for (i = 0; i < mt_slots[type]; i++) {
+ unsigned long piv;
+
+ piv = mas_safe_pivot(mas, pivots, i, type);
+
+ if (!piv && (i != 0)) {
+ pr_err("Missing node limit pivot at " PTR_FMT "[%u]",
+ mas_mn(mas), i);
+ MAS_WARN_ON(mas, 1);
+ }
+
+ if (prev_piv > piv) {
+ pr_err(PTR_FMT "[%u] piv %lu < prev_piv %lu\n",
+ mas_mn(mas), i, piv, prev_piv);
+ MAS_WARN_ON(mas, piv < prev_piv);
+ }
+
+ if (piv < mas->min) {
+ pr_err(PTR_FMT "[%u] %lu < %lu\n", mas_mn(mas), i,
+ piv, mas->min);
+ MAS_WARN_ON(mas, piv < mas->min);
+ }
+ if (piv > mas->max) {
+ pr_err(PTR_FMT "[%u] %lu > %lu\n", mas_mn(mas), i,
+ piv, mas->max);
+ MAS_WARN_ON(mas, piv > mas->max);
+ }
+ prev_piv = piv;
+ if (piv == mas->max)
+ break;
+ }
+
+ if (mas_data_end(mas) != i) {
+ pr_err("node" PTR_FMT ": data_end %u != the last slot offset %u\n",
+ mas_mn(mas), mas_data_end(mas), i);
+ MT_BUG_ON(mas->tree, 1);
+ }
+
+ for (i += 1; i < mt_slots[type]; i++) {
+ void *entry = mas_slot(mas, slots, i);
+
+ if (entry && (i != mt_slots[type] - 1)) {
+ pr_err(PTR_FMT "[%u] should not have entry " PTR_FMT "\n",
+ mas_mn(mas), i, entry);
+ MT_BUG_ON(mas->tree, entry != NULL);
+ }
+
+ if (i < mt_pivots[type]) {
+ unsigned long piv = pivots[i];
+
+ if (!piv)
+ continue;
+
+ pr_err(PTR_FMT "[%u] should not have piv %lu\n",
+ mas_mn(mas), i, piv);
+ MAS_WARN_ON(mas, i < mt_pivots[type] - 1);
+ }
+ }
+}
+
+static void mt_validate_nulls(struct maple_tree *mt)
+{
+ void *entry, *last = (void *)1;
+ unsigned char offset = 0;
+ void __rcu **slots;
+ MA_STATE(mas, mt, 0, 0);
+
+ mas_start(&mas);
+ if (mas_is_none(&mas) || (mas_is_ptr(&mas)))
+ return;
+
+ while (!mte_is_leaf(mas.node))
+ mas_descend(&mas);
+
+ slots = ma_slots(mte_to_node(mas.node), mte_node_type(mas.node));
+ do {
+ entry = mas_slot(&mas, slots, offset);
+ if (!last && !entry) {
+ pr_err("Sequential nulls end at " PTR_FMT "[%u]\n",
+ mas_mn(&mas), offset);
+ }
+ MT_BUG_ON(mt, !last && !entry);
+ last = entry;
+ if (offset == mas_data_end(&mas)) {
+ mas_next_node(&mas, mas_mn(&mas), ULONG_MAX);
+ if (mas_is_overflow(&mas))
+ return;
+ offset = 0;
+ slots = ma_slots(mte_to_node(mas.node),
+ mte_node_type(mas.node));
+ } else {
+ offset++;
+ }
+
+ } while (!mas_is_overflow(&mas));
+}
+
+/*
+ * validate a maple tree by checking:
+ * 1. The limits (pivots are within mas->min to mas->max)
+ * 2. The gap is correctly set in the parents
+ */
+void mt_validate(struct maple_tree *mt)
+ __must_hold(mas->tree->ma_lock)
+{
+ unsigned char end;
+
+ MA_STATE(mas, mt, 0, 0);
+ mas_start(&mas);
+ if (!mas_is_active(&mas))
+ return;
+
+ while (!mte_is_leaf(mas.node))
+ mas_descend(&mas);
+
+ while (!mas_is_overflow(&mas)) {
+ MAS_WARN_ON(&mas, mte_dead_node(mas.node));
+ end = mas_data_end(&mas);
+ if (MAS_WARN_ON(&mas, (end < mt_min_slot_count(mas.node)) &&
+ (!mte_is_root(mas.node)))) {
+ pr_err("Invalid size %u of " PTR_FMT "\n",
+ end, mas_mn(&mas));
+ }
+
+ mas_validate_parent_slot(&mas);
+ mas_validate_limits(&mas);
+ mas_validate_child_slot(&mas);
+ if (mt_is_alloc(mt))
+ mas_validate_gaps(&mas);
+ mas_dfs_postorder(&mas, ULONG_MAX);
+ }
+ mt_validate_nulls(mt);
+}
+EXPORT_SYMBOL_GPL(mt_validate);
+
+void mas_dump(const struct ma_state *mas)
+{
+ pr_err("MAS: tree=" PTR_FMT " enode=" PTR_FMT " ",
+ mas->tree, mas->node);
+ switch (mas->status) {
+ case ma_active:
+ pr_err("(ma_active)");
+ break;
+ case ma_none:
+ pr_err("(ma_none)");
+ break;
+ case ma_root:
+ pr_err("(ma_root)");
+ break;
+ case ma_start:
+ pr_err("(ma_start) ");
+ break;
+ case ma_pause:
+ pr_err("(ma_pause) ");
+ break;
+ case ma_overflow:
+ pr_err("(ma_overflow) ");
+ break;
+ case ma_underflow:
+ pr_err("(ma_underflow) ");
+ break;
+ case ma_error:
+ pr_err("(ma_error) ");
+ break;
+ }
+
+ pr_err("Store Type: ");
+ switch (mas->store_type) {
+ case wr_invalid:
+ pr_err("invalid store type\n");
+ break;
+ case wr_new_root:
+ pr_err("new_root\n");
+ break;
+ case wr_store_root:
+ pr_err("store_root\n");
+ break;
+ case wr_exact_fit:
+ pr_err("exact_fit\n");
+ break;
+ case wr_split_store:
+ pr_err("split_store\n");
+ break;
+ case wr_slot_store:
+ pr_err("slot_store\n");
+ break;
+ case wr_append:
+ pr_err("append\n");
+ break;
+ case wr_node_store:
+ pr_err("node_store\n");
+ break;
+ case wr_spanning_store:
+ pr_err("spanning_store\n");
+ break;
+ case wr_rebalance:
+ pr_err("rebalance\n");
+ break;
+ }
+
+ pr_err("[%u/%u] index=%lx last=%lx\n", mas->offset, mas->end,
+ mas->index, mas->last);
+ pr_err(" min=%lx max=%lx sheaf=" PTR_FMT ", request %lu depth=%u, flags=%x\n",
+ mas->min, mas->max, mas->sheaf, mas->node_request, mas->depth,
+ mas->mas_flags);
+ if (mas->index > mas->last)
+ pr_err("Check index & last\n");
+}
+EXPORT_SYMBOL_GPL(mas_dump);
+
+void mas_wr_dump(const struct ma_wr_state *wr_mas)
+{
+ pr_err("WR_MAS: node=" PTR_FMT " r_min=%lx r_max=%lx\n",
+ wr_mas->node, wr_mas->r_min, wr_mas->r_max);
+ pr_err(" type=%u off_end=%u, node_end=%u, end_piv=%lx\n",
+ wr_mas->type, wr_mas->offset_end, wr_mas->mas->end,
+ wr_mas->end_piv);
+}
+EXPORT_SYMBOL_GPL(mas_wr_dump);
+
+#endif /* CONFIG_DEBUG_MAPLE_TREE */