xarray: Add XArray iterators

The xa_for_each iterator allows the user to efficiently walk a range
of the array, executing the loop body once for each entry in that
range that matches the filter.  This commit also includes xa_find()
and xa_find_after() which are helper functions for xa_for_each() but
may also be useful in their own right.

In the xas family of functions, we have xas_for_each(), xas_find(),
xas_next_entry(), xas_for_each_tagged(), xas_find_tagged(),
xas_next_tagged() and xas_pause().

Signed-off-by: Matthew Wilcox <willy@infradead.org>

2 files changed, 475 insertions, 0 deletions

diff --git a/lib/test_xarray.c b/lib/test_xarray.c
index fb472258b639..e3c2d4d00b15 100644
--- a/lib/test_xarray.c
+++ b/lib/test_xarray.c
@@ -75,6 +75,48 @@ static noinline void check_xa_err(struct xarray *xa)
 //	XA_BUG_ON(xa, xa_err(xa_store(xa, 0, xa_mk_internal(0), 0)) != -EINVAL);
 }
 
+static noinline void check_xas_retry(struct xarray *xa)
+{
+	XA_STATE(xas, xa, 0);
+	void *entry;
+
+	xa_store_index(xa, 0, GFP_KERNEL);
+	xa_store_index(xa, 1, GFP_KERNEL);
+
+	rcu_read_lock();
+	XA_BUG_ON(xa, xas_find(&xas, ULONG_MAX) != xa_mk_value(0));
+	xa_erase_index(xa, 1);
+	XA_BUG_ON(xa, !xa_is_retry(xas_reload(&xas)));
+	XA_BUG_ON(xa, xas_retry(&xas, NULL));
+	XA_BUG_ON(xa, xas_retry(&xas, xa_mk_value(0)));
+	xas_reset(&xas);
+	XA_BUG_ON(xa, xas.xa_node != XAS_RESTART);
+	XA_BUG_ON(xa, xas_next_entry(&xas, ULONG_MAX) != xa_mk_value(0));
+	XA_BUG_ON(xa, xas.xa_node != NULL);
+
+	XA_BUG_ON(xa, xa_store_index(xa, 1, GFP_KERNEL) != NULL);
+	XA_BUG_ON(xa, !xa_is_internal(xas_reload(&xas)));
+	xas.xa_node = XAS_RESTART;
+	XA_BUG_ON(xa, xas_next_entry(&xas, ULONG_MAX) != xa_mk_value(0));
+	rcu_read_unlock();
+
+	/* Make sure we can iterate through retry entries */
+	xas_lock(&xas);
+	xas_set(&xas, 0);
+	xas_store(&xas, XA_RETRY_ENTRY);
+	xas_set(&xas, 1);
+	xas_store(&xas, XA_RETRY_ENTRY);
+
+	xas_set(&xas, 0);
+	xas_for_each(&xas, entry, ULONG_MAX) {
+		xas_store(&xas, xa_mk_value(xas.xa_index));
+	}
+	xas_unlock(&xas);
+
+	xa_erase_index(xa, 0);
+	xa_erase_index(xa, 1);
+}
+
 static noinline void check_xa_load(struct xarray *xa)
 {
 	unsigned long i, j;
@@ -217,6 +259,44 @@ static noinline void check_cmpxchg(struct xarray *xa)
 	XA_BUG_ON(xa, !xa_empty(xa));
 }
 
+static noinline void check_xas_erase(struct xarray *xa)
+{
+	XA_STATE(xas, xa, 0);
+	void *entry;
+	unsigned long i, j;
+
+	for (i = 0; i < 200; i++) {
+		for (j = i; j < 2 * i + 17; j++) {
+			xas_set(&xas, j);
+			do {
+				xas_lock(&xas);
+				xas_store(&xas, xa_mk_value(j));
+				xas_unlock(&xas);
+			} while (xas_nomem(&xas, GFP_KERNEL));
+		}
+
+		xas_set(&xas, ULONG_MAX);
+		do {
+			xas_lock(&xas);
+			xas_store(&xas, xa_mk_value(0));
+			xas_unlock(&xas);
+		} while (xas_nomem(&xas, GFP_KERNEL));
+
+		xas_lock(&xas);
+		xas_store(&xas, NULL);
+
+		xas_set(&xas, 0);
+		j = i;
+		xas_for_each(&xas, entry, ULONG_MAX) {
+			XA_BUG_ON(xa, entry != xa_mk_value(j));
+			xas_store(&xas, NULL);
+			j++;
+		}
+		xas_unlock(&xas);
+		XA_BUG_ON(xa, !xa_empty(xa));
+	}
+}
+
 static noinline void check_multi_store(struct xarray *xa)
 {
 #ifdef CONFIG_XARRAY_MULTI
@@ -285,16 +365,119 @@ static noinline void check_multi_store(struct xarray *xa)
 #endif
 }
 
+static noinline void check_multi_find(struct xarray *xa)
+{
+#ifdef CONFIG_XARRAY_MULTI
+	unsigned long index;
+
+	xa_store_order(xa, 12, 2, xa_mk_value(12), GFP_KERNEL);
+	XA_BUG_ON(xa, xa_store_index(xa, 16, GFP_KERNEL) != NULL);
+
+	index = 0;
+	XA_BUG_ON(xa, xa_find(xa, &index, ULONG_MAX, XA_PRESENT) !=
+			xa_mk_value(12));
+	XA_BUG_ON(xa, index != 12);
+	index = 13;
+	XA_BUG_ON(xa, xa_find(xa, &index, ULONG_MAX, XA_PRESENT) !=
+			xa_mk_value(12));
+	XA_BUG_ON(xa, (index < 12) || (index >= 16));
+	XA_BUG_ON(xa, xa_find_after(xa, &index, ULONG_MAX, XA_PRESENT) !=
+			xa_mk_value(16));
+	XA_BUG_ON(xa, index != 16);
+
+	xa_erase_index(xa, 12);
+	xa_erase_index(xa, 16);
+	XA_BUG_ON(xa, !xa_empty(xa));
+#endif
+}
+
+static noinline void check_multi_find_2(struct xarray *xa)
+{
+	unsigned int max_order = IS_ENABLED(CONFIG_XARRAY_MULTI) ? 10 : 1;
+	unsigned int i, j;
+	void *entry;
+
+	for (i = 0; i < max_order; i++) {
+		unsigned long index = 1UL << i;
+		for (j = 0; j < index; j++) {
+			XA_STATE(xas, xa, j + index);
+			xa_store_index(xa, index - 1, GFP_KERNEL);
+			xa_store_order(xa, index, i, xa_mk_value(index),
+					GFP_KERNEL);
+			rcu_read_lock();
+			xas_for_each(&xas, entry, ULONG_MAX) {
+				xa_erase_index(xa, index);
+			}
+			rcu_read_unlock();
+			xa_erase_index(xa, index - 1);
+			XA_BUG_ON(xa, !xa_empty(xa));
+		}
+	}
+}
+
+static noinline void check_find(struct xarray *xa)
+{
+	unsigned long i, j, k;
+
+	XA_BUG_ON(xa, !xa_empty(xa));
+
+	/*
+	 * Check xa_find with all pairs between 0 and 99 inclusive,
+	 * starting at every index between 0 and 99
+	 */
+	for (i = 0; i < 100; i++) {
+		XA_BUG_ON(xa, xa_store_index(xa, i, GFP_KERNEL) != NULL);
+		xa_set_mark(xa, i, XA_MARK_0);
+		for (j = 0; j < i; j++) {
+			XA_BUG_ON(xa, xa_store_index(xa, j, GFP_KERNEL) !=
+					NULL);
+			xa_set_mark(xa, j, XA_MARK_0);
+			for (k = 0; k < 100; k++) {
+				unsigned long index = k;
+				void *entry = xa_find(xa, &index, ULONG_MAX,
+								XA_PRESENT);
+				if (k <= j)
+					XA_BUG_ON(xa, index != j);
+				else if (k <= i)
+					XA_BUG_ON(xa, index != i);
+				else
+					XA_BUG_ON(xa, entry != NULL);
+
+				index = k;
+				entry = xa_find(xa, &index, ULONG_MAX,
+								XA_MARK_0);
+				if (k <= j)
+					XA_BUG_ON(xa, index != j);
+				else if (k <= i)
+					XA_BUG_ON(xa, index != i);
+				else
+					XA_BUG_ON(xa, entry != NULL);
+			}
+			xa_erase_index(xa, j);
+			XA_BUG_ON(xa, xa_get_mark(xa, j, XA_MARK_0));
+			XA_BUG_ON(xa, !xa_get_mark(xa, i, XA_MARK_0));
+		}
+		xa_erase_index(xa, i);
+		XA_BUG_ON(xa, xa_get_mark(xa, i, XA_MARK_0));
+	}
+	XA_BUG_ON(xa, !xa_empty(xa));
+	check_multi_find(xa);
+	check_multi_find_2(xa);
+}
+
 static DEFINE_XARRAY(array);
 
 static int xarray_checks(void)
 {
 	check_xa_err(&array);
+	check_xas_retry(&array);
 	check_xa_load(&array);
 	check_xa_mark(&array);
 	check_xa_shrink(&array);
+	check_xas_erase(&array);
 	check_cmpxchg(&array);
 	check_multi_store(&array);
+	check_find(&array);
 
 	printk("XArray: %u of %u tests passed\n", tests_passed, tests_run);
 	return (tests_run == tests_passed) ? 0 : -EINVAL;
diff --git a/lib/xarray.c b/lib/xarray.c
index 2ba5a98ec618..24494f42daa6 100644
--- a/lib/xarray.c
+++ b/lib/xarray.c
@@ -128,6 +128,11 @@ static unsigned int get_offset(unsigned long index, struct xa_node *node)
 	return (index >> node->shift) & XA_CHUNK_MASK;
 }
 
+static void xas_set_offset(struct xa_state *xas)
+{
+	xas->xa_offset = get_offset(xas->xa_index, xas->xa_node);
+}
+
 /* move the index either forwards (find) or backwards (sibling slot) */
 static void xas_move_index(struct xa_state *xas, unsigned long offset)
 {
@@ -136,6 +141,12 @@ static void xas_move_index(struct xa_state *xas, unsigned long offset)
 	xas->xa_index += offset << shift;
 }
 
+static void xas_advance(struct xa_state *xas)
+{
+	xas->xa_offset++;
+	xas_move_index(xas, xas->xa_offset);
+}
+
 static void *set_bounds(struct xa_state *xas)
 {
 	xas->xa_node = XAS_BOUNDS;
@@ -830,6 +841,202 @@ void xas_init_marks(const struct xa_state *xas)
 EXPORT_SYMBOL_GPL(xas_init_marks);
 
 /**
+ * xas_pause() - Pause a walk to drop a lock.
+ * @xas: XArray operation state.
+ *
+ * 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 @xas 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 xas_pause(), the xa_for_each()
+ * iterator may be more appropriate.
+ *
+ * Note that xas_pause() only works for forward iteration.  If a user needs
+ * to pause a reverse iteration, we will need a xas_pause_rev().
+ */
+void xas_pause(struct xa_state *xas)
+{
+	struct xa_node *node = xas->xa_node;
+
+	if (xas_invalid(xas))
+		return;
+
+	if (node) {
+		unsigned int offset = xas->xa_offset;
+		while (++offset < XA_CHUNK_SIZE) {
+			if (!xa_is_sibling(xa_entry(xas->xa, node, offset)))
+				break;
+		}
+		xas->xa_index += (offset - xas->xa_offset) << node->shift;
+	} else {
+		xas->xa_index++;
+	}
+	xas->xa_node = XAS_RESTART;
+}
+EXPORT_SYMBOL_GPL(xas_pause);
+
+/**
+ * xas_find() - Find the next present entry in the XArray.
+ * @xas: XArray operation state.
+ * @max: Highest index to return.
+ *
+ * If the @xas has not yet been walked to an entry, return the entry
+ * which has an index >= xas.xa_index.  If it has been walked, the entry
+ * currently being pointed at has been processed, and so we move to the
+ * next entry.
+ *
+ * If no entry is found and the array is smaller than @max, the iterator
+ * is set to the smallest index not yet in the array.  This allows @xas
+ * to be immediately passed to xas_store().
+ *
+ * Return: The entry, if found, otherwise %NULL.
+ */
+void *xas_find(struct xa_state *xas, unsigned long max)
+{
+	void *entry;
+
+	if (xas_error(xas))
+		return NULL;
+
+	if (!xas->xa_node) {
+		xas->xa_index = 1;
+		return set_bounds(xas);
+	} else if (xas_top(xas->xa_node)) {
+		entry = xas_load(xas);
+		if (entry || xas_not_node(xas->xa_node))
+			return entry;
+	} else if (!xas->xa_node->shift &&
+		    xas->xa_offset != (xas->xa_index & XA_CHUNK_MASK)) {
+		xas->xa_offset = ((xas->xa_index - 1) & XA_CHUNK_MASK) + 1;
+	}
+
+	xas_advance(xas);
+
+	while (xas->xa_node && (xas->xa_index <= max)) {
+		if (unlikely(xas->xa_offset == XA_CHUNK_SIZE)) {
+			xas->xa_offset = xas->xa_node->offset + 1;
+			xas->xa_node = xa_parent(xas->xa, xas->xa_node);
+			continue;
+		}
+
+		entry = xa_entry(xas->xa, xas->xa_node, xas->xa_offset);
+		if (xa_is_node(entry)) {
+			xas->xa_node = xa_to_node(entry);
+			xas->xa_offset = 0;
+			continue;
+		}
+		if (entry && !xa_is_sibling(entry))
+			return entry;
+
+		xas_advance(xas);
+	}
+
+	if (!xas->xa_node)
+		xas->xa_node = XAS_BOUNDS;
+	return NULL;
+}
+EXPORT_SYMBOL_GPL(xas_find);
+
+/**
+ * xas_find_marked() - Find the next marked entry in the XArray.
+ * @xas: XArray operation state.
+ * @max: Highest index to return.
+ * @mark: Mark number to search for.
+ *
+ * If the @xas has not yet been walked to an entry, return the marked entry
+ * which has an index >= xas.xa_index.  If it has been walked, the entry
+ * currently being pointed at has been processed, and so we return the
+ * first marked entry with an index > xas.xa_index.
+ *
+ * If no marked entry is found and the array is smaller than @max, @xas is
+ * set to the bounds state and xas->xa_index is set to the smallest index
+ * not yet in the array.  This allows @xas to be immediately passed to
+ * xas_store().
+ *
+ * If no entry is found before @max is reached, @xas is set to the restart
+ * state.
+ *
+ * Return: The entry, if found, otherwise %NULL.
+ */
+void *xas_find_marked(struct xa_state *xas, unsigned long max, xa_mark_t mark)
+{
+	bool advance = true;
+	unsigned int offset;
+	void *entry;
+
+	if (xas_error(xas))
+		return NULL;
+
+	if (!xas->xa_node) {
+		xas->xa_index = 1;
+		goto out;
+	} else if (xas_top(xas->xa_node)) {
+		advance = false;
+		entry = xa_head(xas->xa);
+		xas->xa_node = NULL;
+		if (xas->xa_index > max_index(entry))
+			goto bounds;
+		if (!xa_is_node(entry)) {
+			if (xa_marked(xas->xa, mark))
+				return entry;
+			xas->xa_index = 1;
+			goto out;
+		}
+		xas->xa_node = xa_to_node(entry);
+		xas->xa_offset = xas->xa_index >> xas->xa_node->shift;
+	}
+
+	while (xas->xa_index <= max) {
+		if (unlikely(xas->xa_offset == XA_CHUNK_SIZE)) {
+			xas->xa_offset = xas->xa_node->offset + 1;
+			xas->xa_node = xa_parent(xas->xa, xas->xa_node);
+			if (!xas->xa_node)
+				break;
+			advance = false;
+			continue;
+		}
+
+		if (!advance) {
+			entry = xa_entry(xas->xa, xas->xa_node, xas->xa_offset);
+			if (xa_is_sibling(entry)) {
+				xas->xa_offset = xa_to_sibling(entry);
+				xas_move_index(xas, xas->xa_offset);
+			}
+		}
+
+		offset = xas_find_chunk(xas, advance, mark);
+		if (offset > xas->xa_offset) {
+			advance = false;
+			xas_move_index(xas, offset);
+			/* Mind the wrap */
+			if ((xas->xa_index - 1) >= max)
+				goto max;
+			xas->xa_offset = offset;
+			if (offset == XA_CHUNK_SIZE)
+				continue;
+		}
+
+		entry = xa_entry(xas->xa, xas->xa_node, xas->xa_offset);
+		if (!xa_is_node(entry))
+			return entry;
+		xas->xa_node = xa_to_node(entry);
+		xas_set_offset(xas);
+	}
+
+out:
+	if (!max)
+		goto max;
+bounds:
+	xas->xa_node = XAS_BOUNDS;
+	return NULL;
+max:
+	xas->xa_node = XAS_RESTART;
+	return NULL;
+}
+EXPORT_SYMBOL_GPL(xas_find_marked);
+
+/**
  * xa_init_flags() - Initialise an empty XArray with flags.
  * @xa: XArray.
  * @flags: XA_FLAG values.
@@ -1152,6 +1359,91 @@ void xa_clear_mark(struct xarray *xa, unsigned long index, xa_mark_t mark)
 }
 EXPORT_SYMBOL(xa_clear_mark);
 
+/**
+ * xa_find() - Search the XArray for an entry.
+ * @xa: XArray.
+ * @indexp: Pointer to an index.
+ * @max: Maximum index to search to.
+ * @filter: Selection criterion.
+ *
+ * Finds the entry in @xa which matches the @filter, and has the lowest
+ * index that is at least @indexp and no more than @max.
+ * If an entry is found, @indexp is updated to be the index of the entry.
+ * This function is protected by the RCU read lock, so it may not find
+ * entries which are being simultaneously added.  It will not return an
+ * %XA_RETRY_ENTRY; if you need to see retry entries, use xas_find().
+ *
+ * Context: Any context.  Takes and releases the RCU lock.
+ * Return: The entry, if found, otherwise %NULL.
+ */
+void *xa_find(struct xarray *xa, unsigned long *indexp,
+			unsigned long max, xa_mark_t filter)
+{
+	XA_STATE(xas, xa, *indexp);
+	void *entry;
+
+	rcu_read_lock();
+	do {
+		if ((__force unsigned int)filter < XA_MAX_MARKS)
+			entry = xas_find_marked(&xas, max, filter);
+		else
+			entry = xas_find(&xas, max);
+	} while (xas_retry(&xas, entry));
+	rcu_read_unlock();
+
+	if (entry)
+		*indexp = xas.xa_index;
+	return entry;
+}
+EXPORT_SYMBOL(xa_find);
+
+/**
+ * xa_find_after() - Search the XArray for a present entry.
+ * @xa: XArray.
+ * @indexp: Pointer to an index.
+ * @max: Maximum index to search to.
+ * @filter: Selection criterion.
+ *
+ * Finds the entry in @xa which matches the @filter and has the lowest
+ * index that is above @indexp and no more than @max.
+ * If an entry is found, @indexp is updated to be the index of the entry.
+ * This function is protected by the RCU read lock, so it may miss entries
+ * which are being simultaneously added.  It will not return an
+ * %XA_RETRY_ENTRY; if you need to see retry entries, use xas_find().
+ *
+ * Context: Any context.  Takes and releases the RCU lock.
+ * Return: The pointer, if found, otherwise %NULL.
+ */
+void *xa_find_after(struct xarray *xa, unsigned long *indexp,
+			unsigned long max, xa_mark_t filter)
+{
+	XA_STATE(xas, xa, *indexp + 1);
+	void *entry;
+
+	rcu_read_lock();
+	for (;;) {
+		if ((__force unsigned int)filter < XA_MAX_MARKS)
+			entry = xas_find_marked(&xas, max, filter);
+		else
+			entry = xas_find(&xas, max);
+		if (xas.xa_shift) {
+			if (xas.xa_index & ((1UL << xas.xa_shift) - 1))
+				continue;
+		} else {
+			if (xas.xa_offset < (xas.xa_index & XA_CHUNK_MASK))
+				continue;
+		}
+		if (!xas_retry(&xas, entry))
+			break;
+	}
+	rcu_read_unlock();
+
+	if (entry)
+		*indexp = xas.xa_index;
+	return entry;
+}
+EXPORT_SYMBOL(xa_find_after);
+
 #ifdef XA_DEBUG
 void xa_dump_node(const struct xa_node *node)
 {