KEYS: Expand the capacity of a keyring

Expand the capacity of a keyring to be able to hold a lot more keys by using
the previously added associative array implementation.  Currently the maximum
capacity is:

	(PAGE_SIZE - sizeof(header)) / sizeof(struct key *)

which, on a 64-bit system, is a little more 500.  However, since this is being
used for the NFS uid mapper, we need more than that.  The new implementation
gives us effectively unlimited capacity.

With some alterations, the keyutils testsuite runs successfully to completion
after this patch is applied.  The alterations are because (a) keyrings that
are simply added to no longer appear ordered and (b) some of the errors have
changed a bit.

Signed-off-by: David Howells <dhowells@redhat.com>

6 files changed, 792 insertions, 742 deletions

diff --git a/security/keys/Kconfig b/security/keys/Kconfig
index a90d6d300dbd..15e0dfe8c80f 100644
--- a/security/keys/Kconfig
+++ b/security/keys/Kconfig
@@ -4,6 +4,7 @@
 
 config KEYS
 	bool "Enable access key retention support"
+	select ASSOCIATIVE_ARRAY
 	help
 	  This option provides support for retaining authentication tokens and
 	  access keys in the kernel.
diff --git a/security/keys/gc.c b/security/keys/gc.c
index d67c97bb1025..cce621c33dce 100644
--- a/security/keys/gc.c
+++ b/security/keys/gc.c
@@ -130,6 +130,13 @@ void key_gc_keytype(struct key_type *ktype)
 	kleave("");
 }
 
+static int key_gc_keyring_func(const void *object, void *iterator_data)
+{
+	const struct key *key = object;
+	time_t *limit = iterator_data;
+	return key_is_dead(key, *limit);
+}
+
 /*
  * Garbage collect pointers from a keyring.
  *
@@ -138,10 +145,9 @@ void key_gc_keytype(struct key_type *ktype)
  */
 static void key_gc_keyring(struct key *keyring, time_t limit)
 {
-	struct keyring_list *klist;
-	int loop;
+	int result;
 
-	kenter("%x", key_serial(keyring));
+	kenter("%x{%s}", keyring->serial, keyring->description ?: "");
 
 	if (keyring->flags & ((1 << KEY_FLAG_INVALIDATED) |
 			      (1 << KEY_FLAG_REVOKED)))
@@ -149,27 +155,17 @@ static void key_gc_keyring(struct key *keyring, time_t limit)
 
 	/* scan the keyring looking for dead keys */
 	rcu_read_lock();
-	klist = rcu_dereference(keyring->payload.subscriptions);
-	if (!klist)
-		goto unlock_dont_gc;
-
-	loop = klist->nkeys;
-	smp_rmb();
-	for (loop--; loop >= 0; loop--) {
-		struct key *key = rcu_dereference(klist->keys[loop]);
-		if (key_is_dead(key, limit))
-			goto do_gc;
-	}
-
-unlock_dont_gc:
+	result = assoc_array_iterate(&keyring->keys,
+				     key_gc_keyring_func, &limit);
 	rcu_read_unlock();
+	if (result == true)
+		goto do_gc;
+
 dont_gc:
 	kleave(" [no gc]");
 	return;
 
 do_gc:
-	rcu_read_unlock();
-
 	keyring_gc(keyring, limit);
 	kleave(" [gc]");
 }
@@ -392,7 +388,6 @@ found_unreferenced_key:
 	 */
 found_keyring:
 	spin_unlock(&key_serial_lock);
-	kdebug("scan keyring %d", key->serial);
 	key_gc_keyring(key, limit);
 	goto maybe_resched;
 
diff --git a/security/keys/internal.h b/security/keys/internal.h
index 73950bf8f875..581c6f688352 100644
--- a/security/keys/internal.h
+++ b/security/keys/internal.h
@@ -90,20 +90,23 @@ extern void key_type_put(struct key_type *ktype);
 
 extern int __key_link_begin(struct key *keyring,
 			    const struct keyring_index_key *index_key,
-			    unsigned long *_prealloc);
+			    struct assoc_array_edit **_edit);
 extern int __key_link_check_live_key(struct key *keyring, struct key *key);
-extern void __key_link(struct key *keyring, struct key *key,
-		       unsigned long *_prealloc);
+extern void __key_link(struct key *key, struct assoc_array_edit **_edit);
 extern void __key_link_end(struct key *keyring,
 			   const struct keyring_index_key *index_key,
-			   unsigned long prealloc);
+			   struct assoc_array_edit *edit);
 
-extern key_ref_t __keyring_search_one(key_ref_t keyring_ref,
-				      const struct keyring_index_key *index_key);
+extern key_ref_t find_key_to_update(key_ref_t keyring_ref,
+				    const struct keyring_index_key *index_key);
 
 extern struct key *keyring_search_instkey(struct key *keyring,
 					  key_serial_t target_id);
 
+extern int iterate_over_keyring(const struct key *keyring,
+				int (*func)(const struct key *key, void *data),
+				void *data);
+
 typedef int (*key_match_func_t)(const struct key *, const void *);
 
 struct keyring_search_context {
@@ -119,6 +122,8 @@ struct keyring_search_context {
 #define KEYRING_SEARCH_NO_CHECK_PERM	0x0010	/* Don't check permissions */
 #define KEYRING_SEARCH_DETECT_TOO_DEEP	0x0020	/* Give an error on excessive depth */
 
+	int (*iterator)(const void *object, void *iterator_data);
+
 	/* Internal stuff */
 	int			skipped_ret;
 	bool			possessed;
diff --git a/security/keys/key.c b/security/keys/key.c
index 7d716b82a61e..a819b5c7d4ec 100644
--- a/security/keys/key.c
+++ b/security/keys/key.c
@@ -409,7 +409,7 @@ static int __key_instantiate_and_link(struct key *key,
 				      struct key_preparsed_payload *prep,
 				      struct key *keyring,
 				      struct key *authkey,
-				      unsigned long *_prealloc)
+				      struct assoc_array_edit **_edit)
 {
 	int ret, awaken;
 
@@ -436,7 +436,7 @@ static int __key_instantiate_and_link(struct key *key,
 
 			/* and link it into the destination keyring */
 			if (keyring)
-				__key_link(keyring, key, _prealloc);
+				__key_link(key, _edit);
 
 			/* disable the authorisation key */
 			if (authkey)
@@ -476,7 +476,7 @@ int key_instantiate_and_link(struct key *key,
 			     struct key *authkey)
 {
 	struct key_preparsed_payload prep;
-	unsigned long prealloc;
+	struct assoc_array_edit *edit;
 	int ret;
 
 	memset(&prep, 0, sizeof(prep));
@@ -490,16 +490,15 @@ int key_instantiate_and_link(struct key *key,
 	}
 
 	if (keyring) {
-		ret = __key_link_begin(keyring, &key->index_key, &prealloc);
+		ret = __key_link_begin(keyring, &key->index_key, &edit);
 		if (ret < 0)
 			goto error_free_preparse;
 	}
 
-	ret = __key_instantiate_and_link(key, &prep, keyring, authkey,
-					 &prealloc);
+	ret = __key_instantiate_and_link(key, &prep, keyring, authkey, &edit);
 
 	if (keyring)
-		__key_link_end(keyring, &key->index_key, prealloc);
+		__key_link_end(keyring, &key->index_key, edit);
 
 error_free_preparse:
 	if (key->type->preparse)
@@ -537,7 +536,7 @@ int key_reject_and_link(struct key *key,
 			struct key *keyring,
 			struct key *authkey)
 {
-	unsigned long prealloc;
+	struct assoc_array_edit *edit;
 	struct timespec now;
 	int ret, awaken, link_ret = 0;
 
@@ -548,7 +547,7 @@ int key_reject_and_link(struct key *key,
 	ret = -EBUSY;
 
 	if (keyring)
-		link_ret = __key_link_begin(keyring, &key->index_key, &prealloc);
+		link_ret = __key_link_begin(keyring, &key->index_key, &edit);
 
 	mutex_lock(&key_construction_mutex);
 
@@ -570,7 +569,7 @@ int key_reject_and_link(struct key *key,
 
 		/* and link it into the destination keyring */
 		if (keyring && link_ret == 0)
-			__key_link(keyring, key, &prealloc);
+			__key_link(key, &edit);
 
 		/* disable the authorisation key */
 		if (authkey)
@@ -580,7 +579,7 @@ int key_reject_and_link(struct key *key,
 	mutex_unlock(&key_construction_mutex);
 
 	if (keyring)
-		__key_link_end(keyring, &key->index_key, prealloc);
+		__key_link_end(keyring, &key->index_key, edit);
 
 	/* wake up anyone waiting for a key to be constructed */
 	if (awaken)
@@ -783,8 +782,8 @@ key_ref_t key_create_or_update(key_ref_t keyring_ref,
 		.description	= description,
 	};
 	struct key_preparsed_payload prep;
+	struct assoc_array_edit *edit;
 	const struct cred *cred = current_cred();
-	unsigned long prealloc;
 	struct key *keyring, *key = NULL;
 	key_ref_t key_ref;
 	int ret;
@@ -828,7 +827,7 @@ key_ref_t key_create_or_update(key_ref_t keyring_ref,
 	}
 	index_key.desc_len = strlen(index_key.description);
 
-	ret = __key_link_begin(keyring, &index_key, &prealloc);
+	ret = __key_link_begin(keyring, &index_key, &edit);
 	if (ret < 0) {
 		key_ref = ERR_PTR(ret);
 		goto error_free_prep;
@@ -847,8 +846,8 @@ key_ref_t key_create_or_update(key_ref_t keyring_ref,
 	 * update that instead if possible
 	 */
 	if (index_key.type->update) {
-		key_ref = __keyring_search_one(keyring_ref, &index_key);
-		if (!IS_ERR(key_ref))
+		key_ref = find_key_to_update(keyring_ref, &index_key);
+		if (key_ref)
 			goto found_matching_key;
 	}
 
@@ -874,7 +873,7 @@ key_ref_t key_create_or_update(key_ref_t keyring_ref,
 	}
 
 	/* instantiate it and link it into the target keyring */
-	ret = __key_instantiate_and_link(key, &prep, keyring, NULL, &prealloc);
+	ret = __key_instantiate_and_link(key, &prep, keyring, NULL, &edit);
 	if (ret < 0) {
 		key_put(key);
 		key_ref = ERR_PTR(ret);
@@ -884,7 +883,7 @@ key_ref_t key_create_or_update(key_ref_t keyring_ref,
 	key_ref = make_key_ref(key, is_key_possessed(keyring_ref));
 
 error_link_end:
-	__key_link_end(keyring, &index_key, prealloc);
+	__key_link_end(keyring, &index_key, edit);
 error_free_prep:
 	if (index_key.type->preparse)
 		index_key.type->free_preparse(&prep);
@@ -897,7 +896,7 @@ error:
 	/* we found a matching key, so we're going to try to update it
 	 * - we can drop the locks first as we have the key pinned
 	 */
-	__key_link_end(keyring, &index_key, prealloc);
+	__key_link_end(keyring, &index_key, edit);
 
 	key_ref = __key_update(key_ref, &prep);
 	goto error_free_prep;
diff --git a/security/keys/keyring.c b/security/keys/keyring.c
index eeef1a073db4..f7cdea22214f 100644
--- a/security/keys/keyring.c
+++ b/security/keys/keyring.c
@@ -1,6 +1,6 @@
 /* Keyring handling
  *
- * Copyright (C) 2004-2005, 2008 Red Hat, Inc. All Rights Reserved.
+ * Copyright (C) 2004-2005, 2008, 2013 Red Hat, Inc. All Rights Reserved.
  * Written by David Howells (dhowells@redhat.com)
  *
  * This program is free software; you can redistribute it and/or
@@ -17,25 +17,11 @@
 #include <linux/seq_file.h>
 #include <linux/err.h>
 #include <keys/keyring-type.h>
+#include <keys/user-type.h>
+#include <linux/assoc_array_priv.h>
 #include <linux/uaccess.h>
 #include "internal.h"
 
-#define rcu_dereference_locked_keyring(keyring)				\
-	(rcu_dereference_protected(					\
-		(keyring)->payload.subscriptions,			\
-		rwsem_is_locked((struct rw_semaphore *)&(keyring)->sem)))
-
-#define rcu_deref_link_locked(klist, index, keyring)			\
-	(rcu_dereference_protected(					\
-		(klist)->keys[index],					\
-		rwsem_is_locked((struct rw_semaphore *)&(keyring)->sem)))
-
-#define MAX_KEYRING_LINKS						\
-	min_t(size_t, USHRT_MAX - 1,					\
-	      ((PAGE_SIZE - sizeof(struct keyring_list)) / sizeof(struct key *)))
-
-#define KEY_LINK_FIXQUOTA 1UL
-
 /*
  * When plumbing the depths of the key tree, this sets a hard limit
  * set on how deep we're willing to go.
@@ -47,6 +33,28 @@
  */
 #define KEYRING_NAME_HASH_SIZE	(1 << 5)
 
+/*
+ * We mark pointers we pass to the associative array with bit 1 set if
+ * they're keyrings and clear otherwise.
+ */
+#define KEYRING_PTR_SUBTYPE	0x2UL
+
+static inline bool keyring_ptr_is_keyring(const struct assoc_array_ptr *x)
+{
+	return (unsigned long)x & KEYRING_PTR_SUBTYPE;
+}
+static inline struct key *keyring_ptr_to_key(const struct assoc_array_ptr *x)
+{
+	void *object = assoc_array_ptr_to_leaf(x);
+	return (struct key *)((unsigned long)object & ~KEYRING_PTR_SUBTYPE);
+}
+static inline void *keyring_key_to_ptr(struct key *key)
+{
+	if (key->type == &key_type_keyring)
+		return (void *)((unsigned long)key | KEYRING_PTR_SUBTYPE);
+	return key;
+}
+
 static struct list_head	keyring_name_hash[KEYRING_NAME_HASH_SIZE];
 static DEFINE_RWLOCK(keyring_name_lock);
 
@@ -67,7 +75,6 @@ static inline unsigned keyring_hash(const char *desc)
  */
 static int keyring_instantiate(struct key *keyring,
 			       struct key_preparsed_payload *prep);
-static int keyring_match(const struct key *keyring, const void *criterion);
 static void keyring_revoke(struct key *keyring);
 static void keyring_destroy(struct key *keyring);
 static void keyring_describe(const struct key *keyring, struct seq_file *m);
@@ -76,9 +83,9 @@ static long keyring_read(const struct key *keyring,
 
 struct key_type key_type_keyring = {
 	.name		= "keyring",
-	.def_datalen	= sizeof(struct keyring_list),
+	.def_datalen	= 0,
 	.instantiate	= keyring_instantiate,
-	.match		= keyring_match,
+	.match		= user_match,
 	.revoke		= keyring_revoke,
 	.destroy	= keyring_destroy,
 	.describe	= keyring_describe,
@@ -127,6 +134,7 @@ static int keyring_instantiate(struct key *keyring,
 
 	ret = -EINVAL;
 	if (prep->datalen == 0) {
+		assoc_array_init(&keyring->keys);
 		/* make the keyring available by name if it has one */
 		keyring_publish_name(keyring);
 		ret = 0;
@@ -136,15 +144,226 @@ static int keyring_instantiate(struct key *keyring,
 }
 
 /*
- * Match keyrings on their name
+ * Multiply 64-bits by 32-bits to 96-bits and fold back to 64-bit.  Ideally we'd
+ * fold the carry back too, but that requires inline asm.
+ */
+static u64 mult_64x32_and_fold(u64 x, u32 y)
+{
+	u64 hi = (u64)(u32)(x >> 32) * y;
+	u64 lo = (u64)(u32)(x) * y;
+	return lo + ((u64)(u32)hi << 32) + (u32)(hi >> 32);
+}
+
+/*
+ * Hash a key type and description.
+ */
+static unsigned long hash_key_type_and_desc(const struct keyring_index_key *index_key)
+{
+	const unsigned level_shift = ASSOC_ARRAY_LEVEL_STEP;
+	const unsigned long level_mask = ASSOC_ARRAY_LEVEL_STEP_MASK;
+	const char *description = index_key->description;
+	unsigned long hash, type;
+	u32 piece;
+	u64 acc;
+	int n, desc_len = index_key->desc_len;
+
+	type = (unsigned long)index_key->type;
+
+	acc = mult_64x32_and_fold(type, desc_len + 13);
+	acc = mult_64x32_and_fold(acc, 9207);
+	for (;;) {
+		n = desc_len;
+		if (n <= 0)
+			break;
+		if (n > 4)
+			n = 4;
+		piece = 0;
+		memcpy(&piece, description, n);
+		description += n;
+		desc_len -= n;
+		acc = mult_64x32_and_fold(acc, piece);
+		acc = mult_64x32_and_fold(acc, 9207);
+	}
+
+	/* Fold the hash down to 32 bits if need be. */
+	hash = acc;
+	if (ASSOC_ARRAY_KEY_CHUNK_SIZE == 32)
+		hash ^= acc >> 32;
+
+	/* Squidge all the keyrings into a separate part of the tree to
+	 * ordinary keys by making sure the lowest level segment in the hash is
+	 * zero for keyrings and non-zero otherwise.
+	 */
+	if (index_key->type != &key_type_keyring && (hash & level_mask) == 0)
+		return hash | (hash >> (ASSOC_ARRAY_KEY_CHUNK_SIZE - level_shift)) | 1;
+	if (index_key->type == &key_type_keyring && (hash & level_mask) != 0)
+		return (hash + (hash << level_shift)) & ~level_mask;
+	return hash;
+}
+
+/*
+ * Build the next index key chunk.
+ *
+ * On 32-bit systems the index key is laid out as:
+ *
+ *	0	4	5	9...
+ *	hash	desclen	typeptr	desc[]
+ *
+ * On 64-bit systems:
+ *
+ *	0	8	9	17...
+ *	hash	desclen	typeptr	desc[]
+ *
+ * We return it one word-sized chunk at a time.
  */
-static int keyring_match(const struct key *keyring, const void *description)
+static unsigned long keyring_get_key_chunk(const void *data, int level)
+{
+	const struct keyring_index_key *index_key = data;
+	unsigned long chunk = 0;
+	long offset = 0;
+	int desc_len = index_key->desc_len, n = sizeof(chunk);
+
+	level /= ASSOC_ARRAY_KEY_CHUNK_SIZE;
+	switch (level) {
+	case 0:
+		return hash_key_type_and_desc(index_key);
+	case 1:
+		return ((unsigned long)index_key->type << 8) | desc_len;
+	case 2:
+		if (desc_len == 0)
+			return (u8)((unsigned long)index_key->type >>
+				    (ASSOC_ARRAY_KEY_CHUNK_SIZE - 8));
+		n--;
+		offset = 1;
+	default:
+		offset += sizeof(chunk) - 1;
+		offset += (level - 3) * sizeof(chunk);
+		if (offset >= desc_len)
+			return 0;
+		desc_len -= offset;
+		if (desc_len > n)
+			desc_len = n;
+		offset += desc_len;
+		do {
+			chunk <<= 8;
+			chunk |= ((u8*)index_key->description)[--offset];
+		} while (--desc_len > 0);
+
+		if (level == 2) {
+			chunk <<= 8;
+			chunk |= (u8)((unsigned long)index_key->type >>
+				      (ASSOC_ARRAY_KEY_CHUNK_SIZE - 8));
+		}
+		return chunk;
+	}
+}
+
+static unsigned long keyring_get_object_key_chunk(const void *object, int level)
+{
+	const struct key *key = keyring_ptr_to_key(object);
+	return keyring_get_key_chunk(&key->index_key, level);
+}
+
+static bool keyring_compare_object(const void *object, const void *data)
 {
-	return keyring->description &&
-		strcmp(keyring->description, description) == 0;
+	const struct keyring_index_key *index_key = data;
+	const struct key *key = keyring_ptr_to_key(object);
+
+	return key->index_key.type == index_key->type &&
+		key->index_key.desc_len == index_key->desc_len &&
+		memcmp(key->index_key.description, index_key->description,
+		       index_key->desc_len) == 0;
 }
 
 /*
+ * Compare the index keys of a pair of objects and determine the bit position
+ * at which they differ - if they differ.
+ */
+static int keyring_diff_objects(const void *_a, const void *_b)
+{
+	const struct key *key_a = keyring_ptr_to_key(_a);
+	const struct key *key_b = keyring_ptr_to_key(_b);
+	const struct keyring_index_key *a = &key_a->index_key;
+	const struct keyring_index_key *b = &key_b->index_key;
+	unsigned long seg_a, seg_b;
+	int level, i;
+
+	level = 0;
+	seg_a = hash_key_type_and_desc(a);
+	seg_b = hash_key_type_and_desc(b);
+	if ((seg_a ^ seg_b) != 0)
+		goto differ;
+
+	/* The number of bits contributed by the hash is controlled by a
+	 * constant in the assoc_array headers.  Everything else thereafter we
+	 * can deal with as being machine word-size dependent.
+	 */
+	level += ASSOC_ARRAY_KEY_CHUNK_SIZE / 8;
+	seg_a = a->desc_len;
+	seg_b = b->desc_len;
+	if ((seg_a ^ seg_b) != 0)
+		goto differ;
+
+	/* The next bit may not work on big endian */
+	level++;
+	seg_a = (unsigned long)a->type;
+	seg_b = (unsigned long)b->type;
+	if ((seg_a ^ seg_b) != 0)
+		goto differ;
+
+	level += sizeof(unsigned long);
+	if (a->desc_len == 0)
+		goto same;
+
+	i = 0;
+	if (((unsigned long)a->description | (unsigned long)b->description) &
+	    (sizeof(unsigned long) - 1)) {
+		do {
+			seg_a = *(unsigned long *)(a->description + i);
+			seg_b = *(unsigned long *)(b->description + i);
+			if ((seg_a ^ seg_b) != 0)
+				goto differ_plus_i;
+			i += sizeof(unsigned long);
+		} while (i < (a->desc_len & (sizeof(unsigned long) - 1)));
+	}
+
+	for (; i < a->desc_len; i++) {
+		seg_a = *(unsigned char *)(a->description + i);
+		seg_b = *(unsigned char *)(b->description + i);
+		if ((seg_a ^ seg_b) != 0)
+			goto differ_plus_i;
+	}
+
+same:
+	return -1;
+
+differ_plus_i:
+	level += i;
+differ:
+	i = level * 8 + __ffs(seg_a ^ seg_b);
+	return i;
+}
+
+/*
+ * Free an object after stripping the keyring flag off of the pointer.
+ */
+static void keyring_free_object(void *object)
+{
+	key_put(keyring_ptr_to_key(object));
+}
+
+/*
+ * Operations for keyring management by the index-tree routines.
+ */
+static const struct assoc_array_ops keyring_assoc_array_ops = {
+	.get_key_chunk		= keyring_get_key_chunk,
+	.get_object_key_chunk	= keyring_get_object_key_chunk,
+	.compare_object		= keyring_compare_object,
+	.diff_objects		= keyring_diff_objects,
+	.free_object		= keyring_free_object,
+};
+
+/*
  * Clean up a keyring when it is destroyed.  Unpublish its name if it had one
  * and dispose of its data.
  *
@@ -155,9 +374,6 @@ static int keyring_match(const struct key *keyring, const void *description)
  */
 static void keyring_destroy(struct key *keyring)
 {
-	struct keyring_list *klist;
-	int loop;
-
 	if (keyring->description) {
 		write_lock(&keyring_name_lock);
 
@@ -168,12 +384,7 @@ static void keyring_destroy(struct key *keyring)
 		write_unlock(&keyring_name_lock);
 	}
 
-	klist = rcu_access_pointer(keyring->payload.subscriptions);
-	if (klist) {
-		for (loop = klist->nkeys - 1; loop >= 0; loop--)
-			key_put(rcu_access_pointer(klist->keys[loop]));
-		kfree(klist);
-	}
+	assoc_array_destroy(&keyring->keys, &keyring_assoc_array_ops);
 }
 
 /*
@@ -181,76 +392,88 @@ static void keyring_destroy(struct key *keyring)
  */
 static void keyring_describe(const struct key *keyring, struct seq_file *m)
 {
-	struct keyring_list *klist;
-
 	if (keyring->description)
 		seq_puts(m, keyring->description);
 	else
 		seq_puts(m, "[anon]");
 
 	if (key_is_instantiated(keyring)) {
-		rcu_read_lock();
-		klist = rcu_dereference(keyring->payload.subscriptions);
-		if (klist)
-			seq_printf(m, ": %u/%u", klist->nkeys, klist->maxkeys);
+		if (keyring->keys.nr_leaves_on_tree != 0)
+			seq_printf(m, ": %lu", keyring->keys.nr_leaves_on_tree);
 		else
 			seq_puts(m, ": empty");
-		rcu_read_unlock();
 	}
 }
 
+struct keyring_read_iterator_context {
+	size_t			qty;
+	size_t			count;
+	key_serial_t __user	*buffer;
+};
+
+static int keyring_read_iterator(const void *object, void *data)
+{
+	struct keyring_read_iterator_context *ctx = data;
+	const struct key *key = keyring_ptr_to_key(object);
+	int ret;
+
+	kenter("{%s,%d},,{%zu/%zu}",
+	       key->type->name, key->serial, ctx->count, ctx->qty);
+
+	if (ctx->count >= ctx->qty)
+		return 1;
+
+	ret = put_user(key->serial, ctx->buffer);
+	if (ret < 0)
+		return ret;
+	ctx->buffer++;
+	ctx->count += sizeof(key->serial);
+	return 0;
+}
+
 /*
  * Read a list of key IDs from the keyring's contents in binary form
  *
- * The keyring's semaphore is read-locked by the caller.
+ * The keyring's semaphore is read-locked by the caller.  This prevents someone
+ * from modifying it under us - which could cause us to read key IDs multiple
+ * times.
  */
 static long keyring_read(const struct key *keyring,
 			 char __user *buffer, size_t buflen)
 {
-	struct keyring_list *klist;
-	struct key *key;
-	size_t qty, tmp;
-	int loop, ret;
+	struct keyring_read_iterator_context ctx;
+	unsigned long nr_keys;
+	int ret;
 
-	ret = 0;
-	klist = rcu_dereference_locked_keyring(keyring);
-	if (klist) {
-		/* calculate how much data we could return */
-		qty = klist->nkeys * sizeof(key_serial_t);
-
-		if (buffer && buflen > 0) {
-			if (buflen > qty)
-				buflen = qty;
-
-			/* copy the IDs of the subscribed keys into the
-			 * buffer */
-			ret = -EFAULT;
-
-			for (loop = 0; loop < klist->nkeys; loop++) {
-				key = rcu_deref_link_locked(klist, loop,
-							    keyring);
-
-				tmp = sizeof(key_serial_t);
-				if (tmp > buflen)
-					tmp = buflen;
-
-				if (copy_to_user(buffer,
-						 &key->serial,
-						 tmp) != 0)
-					goto error;
-
-				buflen -= tmp;
-				if (buflen == 0)
-					break;
-				buffer += tmp;
-			}
-		}
+	kenter("{%d},,%zu", key_serial(keyring), buflen);
+
+	if (buflen & (sizeof(key_serial_t) - 1))
+		return -EINVAL;
+
+	nr_keys = keyring->keys.nr_leaves_on_tree;
+	if (nr_keys == 0)
+		return 0;
 
-		ret = qty;
+	/* Calculate how much data we could return */
+	ctx.qty = nr_keys * sizeof(key_serial_t);
+
+	if (!buffer || !buflen)
+		return ctx.qty;
+
+	if (buflen > ctx.qty)
+		ctx.qty = buflen;
+
+	/* Copy the IDs of the subscribed keys into the buffer */
+	ctx.buffer = (key_serial_t __user *)buffer;
+	ctx.count = 0;
+	ret = assoc_array_iterate(&keyring->keys, keyring_read_iterator, &ctx);
+	if (ret < 0) {
+		kleave(" = %d [iterate]", ret);
+		return ret;
 	}
 
-error:
-	return ret;
+	kleave(" = %zu [ok]", ctx.count);
+	return ctx.count;
 }
 
 /*
@@ -277,219 +500,360 @@ struct key *keyring_alloc(const char *description, kuid_t uid, kgid_t gid,
 }
 EXPORT_SYMBOL(keyring_alloc);
 
-/**
- * keyring_search_aux - Search a keyring tree for a key matching some criteria
- * @keyring_ref: A pointer to the keyring with possession indicator.
- * @ctx: The keyring search context.
- *
- * Search the supplied keyring tree for a key that matches the criteria given.
- * The root keyring and any linked keyrings must grant Search permission to the
- * caller to be searchable and keys can only be found if they too grant Search
- * to the caller. The possession flag on the root keyring pointer controls use
- * of the possessor bits in permissions checking of the entire tree.  In
- * addition, the LSM gets to forbid keyring searches and key matches.
- *
- * The search is performed as a breadth-then-depth search up to the prescribed
- * limit (KEYRING_SEARCH_MAX_DEPTH).
- *
- * Keys are matched to the type provided and are then filtered by the match
- * function, which is given the description to use in any way it sees fit.  The
- * match function may use any attributes of a key that it wishes to to
- * determine the match.  Normally the match function from the key type would be
- * used.
- *
- * RCU is used to prevent the keyring key lists from disappearing without the
- * need to take lots of locks.
- *
- * Returns a pointer to the found key and increments the key usage count if
- * successful; -EAGAIN if no matching keys were found, or if expired or revoked
- * keys were found; -ENOKEY if only negative keys were found; -ENOTDIR if the
- * specified keyring wasn't a keyring.
- *
- * In the case of a successful return, the possession attribute from
- * @keyring_ref is propagated to the returned key reference.
+/*
+ * Iteration function to consider each key found.
  */
-key_ref_t keyring_search_aux(key_ref_t keyring_ref,
-			     struct keyring_search_context *ctx)
+static int keyring_search_iterator(const void *object, void *iterator_data)
 {
-	struct {
-		/* Need a separate keylist pointer for RCU purposes */
-		struct key *keyring;
-		struct keyring_list *keylist;
-		int kix;
-	} stack[KEYRING_SEARCH_MAX_DEPTH];
-
-	struct keyring_list *keylist;
-	unsigned long kflags;
-	struct key *keyring, *key;
-	key_ref_t key_ref;
-	long err;
-	int sp, nkeys, kix;
+	struct keyring_search_context *ctx = iterator_data;
+	const struct key *key = keyring_ptr_to_key(object);
+	unsigned long kflags = key->flags;
 
-	keyring = key_ref_to_ptr(keyring_ref);
-	ctx->possessed = is_key_possessed(keyring_ref);
-	key_check(keyring);
+	kenter("{%d}", key->serial);
 
-	/* top keyring must have search permission to begin the search */
-	err = key_task_permission(keyring_ref, ctx->cred, KEY_SEARCH);
-	if (err < 0) {
-		key_ref = ERR_PTR(err);
-		goto error;
+	/* ignore keys not of this type */
+	if (key->type != ctx->index_key.type) {
+		kleave(" = 0 [!type]");
+		return 0;
 	}
 
-	key_ref = ERR_PTR(-ENOTDIR);
-	if (keyring->type != &key_type_keyring)
-		goto error;
+	/* skip invalidated, revoked and expired keys */
+	if (ctx->flags & KEYRING_SEARCH_DO_STATE_CHECK) {
+		if (kflags & ((1 << KEY_FLAG_INVALIDATED) |
+			      (1 << KEY_FLAG_REVOKED))) {
+			ctx->result = ERR_PTR(-EKEYREVOKED);
+			kleave(" = %d [invrev]", ctx->skipped_ret);
+			goto skipped;
+		}
 
-	rcu_read_lock();
+		if (key->expiry && ctx->now.tv_sec >= key->expiry) {
+			ctx->result = ERR_PTR(-EKEYEXPIRED);
+			kleave(" = %d [expire]", ctx->skipped_ret);
+			goto skipped;
+		}
+	}
 
-	ctx->now = current_kernel_time();
-	err = -EAGAIN;
-	sp = 0;
-
-	/* firstly we should check to see if this top-level keyring is what we
-	 * are looking for */
-	key_ref = ERR_PTR(-EAGAIN);
-	kflags = keyring->flags;
-	if (keyring->type == ctx->index_key.type &&
-	    ctx->match(keyring, ctx->match_data)) {
-		key = keyring;
-		if (ctx->flags & KEYRING_SEARCH_NO_STATE_CHECK)
-			goto found;
+	/* keys that don't match */
+	if (!ctx->match(key, ctx->match_data)) {
+		kleave(" = 0 [!match]");
+		return 0;
+	}
 
-		/* check it isn't negative and hasn't expired or been
-		 * revoked */
-		if (kflags & (1 << KEY_FLAG_REVOKED))
-			goto error_2;
-		if (key->expiry && ctx->now.tv_sec >= key->expiry)
-			goto error_2;
-		key_ref = ERR_PTR(key->type_data.reject_error);
-		if (kflags & (1 << KEY_FLAG_NEGATIVE))
-			goto error_2;
-		goto found;
+	/* key must have search permissions */
+	if (!(ctx->flags & KEYRING_SEARCH_NO_CHECK_PERM) &&
+	    key_task_permission(make_key_ref(key, ctx->possessed),
+				ctx->cred, KEY_SEARCH) < 0) {
+		ctx->result = ERR_PTR(-EACCES);
+		kleave(" = %d [!perm]", ctx->skipped_ret);
+		goto skipped;
 	}
 
-	/* otherwise, the top keyring must not be revoked, expired, or
-	 * negatively instantiated if we are to search it */
-	key_ref = ERR_PTR(-EAGAIN);
-	if (kflags & ((1 << KEY_FLAG_INVALIDATED) |
-		      (1 << KEY_FLAG_REVOKED) |
-		      (1 << KEY_FLAG_NEGATIVE)) ||
-	    (keyring->expiry && ctx->now.tv_sec >= keyring->expiry))
-		goto error_2;
-
-	/* start processing a new keyring */
-descend:
-	kflags = keyring->flags;
-	if (kflags & ((1 << KEY_FLAG_INVALIDATED) |
-		      (1 << KEY_FLAG_REVOKED)))
-		goto not_this_keyring;
+	if (ctx->flags & KEYRING_SEARCH_DO_STATE_CHECK) {
+		/* we set a different error code if we pass a negative key */
+		if (kflags & (1 << KEY_FLAG_NEGATIVE)) {
+			ctx->result = ERR_PTR(key->type_data.reject_error);
+			kleave(" = %d [neg]", ctx->skipped_ret);
+			goto skipped;
+		}
+	}
 
-	keylist = rcu_dereference(keyring->payload.subscriptions);
-	if (!keylist)
-		goto not_this_keyring;
+	/* Found */
+	ctx->result = make_key_ref(key, ctx->possessed);
+	kleave(" = 1 [found]");
+	return 1;
 
-	/* iterate through the keys in this keyring first */
-	nkeys = keylist->nkeys;
-	smp_rmb();
-	for (kix = 0; kix < nkeys; kix++) {
-		key = rcu_dereference(keylist->keys[kix]);
-		kflags = key->flags;
+skipped:
+	return ctx->skipped_ret;
+}
 
-		/* ignore keys not of this type */
-		if (key->type != ctx->index_key.type)
-			continue;
+/*
+ * Search inside a keyring for a key.  We can search by walking to it
+ * directly based on its index-key or we can iterate over the entire
+ * tree looking for it, based on the match function.
+ */
+static int search_keyring(struct key *keyring, struct keyring_search_context *ctx)
+{
+	if ((ctx->flags & KEYRING_SEARCH_LOOKUP_TYPE) ==
+	    KEYRING_SEARCH_LOOKUP_DIRECT) {
+		const void *object;
+
+		object = assoc_array_find(&keyring->keys,
+					  &keyring_assoc_array_ops,
+					  &ctx->index_key);
+		return object ? ctx->iterator(object, ctx) : 0;
+	}
+	return assoc_array_iterate(&keyring->keys, ctx->iterator, ctx);
+}
 
-		/* skip invalidated, revoked and expired keys */
-		if (!(ctx->flags & KEYRING_SEARCH_NO_STATE_CHECK)) {
-			if (kflags & ((1 << KEY_FLAG_INVALIDATED) |
-				      (1 << KEY_FLAG_REVOKED)))
-				continue;
+/*
+ * Search a tree of keyrings that point to other keyrings up to the maximum
+ * depth.
+ */
+static bool search_nested_keyrings(struct key *keyring,
+				   struct keyring_search_context *ctx)
+{
+	struct {
+		struct key *keyring;
+		struct assoc_array_node *node;
+		int slot;
+	} stack[KEYRING_SEARCH_MAX_DEPTH];
 
-			if (key->expiry && ctx->now.tv_sec >= key->expiry)
-				continue;
-		}
+	struct assoc_array_shortcut *shortcut;
+	struct assoc_array_node *node;
+	struct assoc_array_ptr *ptr;
+	struct key *key;
+	int sp = 0, slot;
 
-		/* keys that don't match */
-		if (!ctx->match(key, ctx->match_data))
-			continue;
+	kenter("{%d},{%s,%s}",
+	       keyring->serial,
+	       ctx->index_key.type->name,
+	       ctx->index_key.description);
 
-		/* key must have search permissions */
-		if (key_task_permission(make_key_ref(key, ctx->possessed),
-					ctx->cred, KEY_SEARCH) < 0)
-			continue;
+	if (ctx->index_key.description)
+		ctx->index_key.desc_len = strlen(ctx->index_key.description);
 
-		if (ctx->flags & KEYRING_SEARCH_NO_STATE_CHECK)
+	/* Check to see if this top-level keyring is what we are looking for
+	 * and whether it is valid or not.
+	 */
+	if (ctx->flags & KEYRING_SEARCH_LOOKUP_ITERATE ||
+	    keyring_compare_object(keyring, &ctx->index_key)) {
+		ctx->skipped_ret = 2;
+		ctx->flags |= KEYRING_SEARCH_DO_STATE_CHECK;
+		switch (ctx->iterator(keyring_key_to_ptr(keyring), ctx)) {
+		case 1:
 			goto found;
-
-		/* we set a different error code if we pass a negative key */
-		if (kflags & (1 << KEY_FLAG_NEGATIVE)) {
-			err = key->type_data.reject_error;
-			continue;
+		case 2:
+			return false;
+		default:
+			break;
 		}
+	}
 
+	ctx->skipped_ret = 0;
+	if (ctx->flags & KEYRING_SEARCH_NO_STATE_CHECK)
+		ctx->flags &= ~KEYRING_SEARCH_DO_STATE_CHECK;
+
+	/* Start processing a new keyring */
+descend_to_keyring:
+	kdebug("descend to %d", keyring->serial);
+	if (keyring->flags & ((1 << KEY_FLAG_INVALIDATED) |
+			      (1 << KEY_FLAG_REVOKED)))
+		goto not_this_keyring;
+
+	/* Search through the keys in this keyring before its searching its
+	 * subtrees.
+	 */
+	if (search_keyring(keyring, ctx))
 		goto found;
-	}
 
-	/* search through the keyrings nested in this one */
-	kix = 0;
-ascend:
-	nkeys = keylist->nkeys;
-	smp_rmb();
-	for (; kix < nkeys; kix++) {
-		key = rcu_dereference(keylist->keys[kix]);
-		if (key-