Merge tag 'keys-namespace-20190627' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs

Pull keyring namespacing from David Howells:
 "These patches help make keys and keyrings more namespace aware.

  Firstly some miscellaneous patches to make the process easier:

   - Simplify key index_key handling so that the word-sized chunks
     assoc_array requires don't have to be shifted about, making it
     easier to add more bits into the key.

   - Cache the hash value in the key so that we don't have to calculate
     on every key we examine during a search (it involves a bunch of
     multiplications).

   - Allow keying_search() to search non-recursively.

  Then the main patches:

   - Make it so that keyring names are per-user_namespace from the point
     of view of KEYCTL_JOIN_SESSION_KEYRING so that they're not
     accessible cross-user_namespace.

     keyctl_capabilities() shows KEYCTL_CAPS1_NS_KEYRING_NAME for this.

   - Move the user and user-session keyrings to the user_namespace
     rather than the user_struct. This prevents them propagating
     directly across user_namespaces boundaries (ie. the KEY_SPEC_*
     flags will only pick from the current user_namespace).

   - Make it possible to include the target namespace in which the key
     shall operate in the index_key. This will allow the possibility of
     multiple keys with the same description, but different target
     domains to be held in the same keyring.

     keyctl_capabilities() shows KEYCTL_CAPS1_NS_KEY_TAG for this.

   - Make it so that keys are implicitly invalidated by removal of a
     domain tag, causing them to be garbage collected.

   - Institute a network namespace domain tag that allows keys to be
     differentiated by the network namespace in which they operate. New
     keys that are of a type marked 'KEY_TYPE_NET_DOMAIN' are assigned
     the network domain in force when they are created.

   - Make it so that the desired network namespace can be handed down
     into the request_key() mechanism. This allows AFS, NFS, etc. to
     request keys specific to the network namespace of the superblock.

     This also means that the keys in the DNS record cache are
     thenceforth namespaced, provided network filesystems pass the
     appropriate network namespace down into dns_query().

     For DNS, AFS and NFS are good, whilst CIFS and Ceph are not. Other
     cache keyrings, such as idmapper keyrings, also need to set the
     domain tag - for which they need access to the network namespace of
     the superblock"

* tag 'keys-namespace-20190627' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
  keys: Pass the network namespace into request_key mechanism
  keys: Network namespace domain tag
  keys: Garbage collect keys for which the domain has been removed
  keys: Include target namespace in match criteria
  keys: Move the user and user-session keyrings to the user_namespace
  keys: Namespace keyring names
  keys: Add a 'recurse' flag for keyring searches
  keys: Cache the hash value to avoid lots of recalculation
  keys: Simplify key description management

1 files changed, 149 insertions, 114 deletions

diff --git a/security/keys/keyring.c b/security/keys/keyring.c
index e4de4070c754..febf36c6ddc5 100644
--- a/security/keys/keyring.c
+++ b/security/keys/keyring.c
@@ -12,10 +12,13 @@
 #include <linux/security.h>
 #include <linux/seq_file.h>
 #include <linux/err.h>
+#include <linux/user_namespace.h>
+#include <linux/nsproxy.h>
 #include <keys/keyring-type.h>
 #include <keys/user-type.h>
 #include <linux/assoc_array_priv.h>
 #include <linux/uaccess.h>
+#include <net/net_namespace.h>
 #include "internal.h"
 
 /*
@@ -25,11 +28,6 @@
 #define KEYRING_SEARCH_MAX_DEPTH 6
 
 /*
- * We keep all named keyrings in a hash to speed looking them up.
- */
-#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.
  */
@@ -51,17 +49,21 @@ static inline void *keyring_key_to_ptr(struct key *key)
 	return key;
 }
 
-static struct list_head	keyring_name_hash[KEYRING_NAME_HASH_SIZE];
 static DEFINE_RWLOCK(keyring_name_lock);
 
-static inline unsigned keyring_hash(const char *desc)
+/*
+ * Clean up the bits of user_namespace that belong to us.
+ */
+void key_free_user_ns(struct user_namespace *ns)
 {
-	unsigned bucket = 0;
-
-	for (; *desc; desc++)
-		bucket += (unsigned char)*desc;
-
-	return bucket & (KEYRING_NAME_HASH_SIZE - 1);
+	write_lock(&keyring_name_lock);
+	list_del_init(&ns->keyring_name_list);
+	write_unlock(&keyring_name_lock);
+
+	key_put(ns->user_keyring_register);
+#ifdef CONFIG_PERSISTENT_KEYRINGS
+	key_put(ns->persistent_keyring_register);
+#endif
 }
 
 /*
@@ -100,23 +102,17 @@ static DEFINE_MUTEX(keyring_serialise_link_lock);
 
 /*
  * Publish the name of a keyring so that it can be found by name (if it has
- * one).
+ * one and it doesn't begin with a dot).
  */
 static void keyring_publish_name(struct key *keyring)
 {
-	int bucket;
-
-	if (keyring->description) {
-		bucket = keyring_hash(keyring->description);
+	struct user_namespace *ns = current_user_ns();
 
+	if (keyring->description &&
+	    keyring->description[0] &&
+	    keyring->description[0] != '.') {
 		write_lock(&keyring_name_lock);
-
-		if (!keyring_name_hash[bucket].next)
-			INIT_LIST_HEAD(&keyring_name_hash[bucket]);
-
-		list_add_tail(&keyring->name_link,
-			      &keyring_name_hash[bucket]);
-
+		list_add_tail(&keyring->name_link, &ns->keyring_name_list);
 		write_unlock(&keyring_name_lock);
 	}
 }
@@ -164,7 +160,7 @@ static u64 mult_64x32_and_fold(u64 x, u32 y)
 /*
  * Hash a key type and description.
  */
-static unsigned long hash_key_type_and_desc(const struct keyring_index_key *index_key)
+static void hash_key_type_and_desc(struct keyring_index_key *index_key)
 {
 	const unsigned level_shift = ASSOC_ARRAY_LEVEL_STEP;
 	const unsigned long fan_mask = ASSOC_ARRAY_FAN_MASK;
@@ -175,9 +171,12 @@ static unsigned long hash_key_type_and_desc(const struct keyring_index_key *inde
 	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);
+	piece = (unsigned long)index_key->domain_tag;
+	acc = mult_64x32_and_fold(acc, piece);
+	acc = mult_64x32_and_fold(acc, 9207);
+
 	for (;;) {
 		n = desc_len;
 		if (n <= 0)
@@ -202,24 +201,67 @@ static unsigned long hash_key_type_and_desc(const struct keyring_index_key *inde
 	 * zero for keyrings and non-zero otherwise.
 	 */
 	if (index_key->type != &key_type_keyring && (hash & fan_mask) == 0)
-		return hash | (hash >> (ASSOC_ARRAY_KEY_CHUNK_SIZE - level_shift)) | 1;
-	if (index_key->type == &key_type_keyring && (hash & fan_mask) != 0)
-		return (hash + (hash << level_shift)) & ~fan_mask;
-	return hash;
+		hash |= (hash >> (ASSOC_ARRAY_KEY_CHUNK_SIZE - level_shift)) | 1;
+	else if (index_key->type == &key_type_keyring && (hash & fan_mask) != 0)
+		hash = (hash + (hash << level_shift)) & ~fan_mask;
+	index_key->hash = 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[]
+ * Finalise an index key to include a part of the description actually in the
+ * index key, to set the domain tag and to calculate the hash.
+ */
+void key_set_index_key(struct keyring_index_key *index_key)
+{
+	static struct key_tag default_domain_tag = { .usage = REFCOUNT_INIT(1), };
+	size_t n = min_t(size_t, index_key->desc_len, sizeof(index_key->desc));
+
+	memcpy(index_key->desc, index_key->description, n);
+
+	if (!index_key->domain_tag) {
+		if (index_key->type->flags & KEY_TYPE_NET_DOMAIN)
+			index_key->domain_tag = current->nsproxy->net_ns->key_domain;
+		else
+			index_key->domain_tag = &default_domain_tag;
+	}
+
+	hash_key_type_and_desc(index_key);
+}
+
+/**
+ * key_put_tag - Release a ref on a tag.
+ * @tag: The tag to release.
  *
- * On 64-bit systems:
+ * This releases a reference the given tag and returns true if that ref was the
+ * last one.
+ */
+bool key_put_tag(struct key_tag *tag)
+{
+	if (refcount_dec_and_test(&tag->usage)) {
+		kfree_rcu(tag, rcu);
+		return true;
+	}
+
+	return false;
+}
+
+/**
+ * key_remove_domain - Kill off a key domain and gc its keys
+ * @domain_tag: The domain tag to release.
  *
- *	0	8	9	17...
- *	hash	desclen	typeptr	desc[]
+ * This marks a domain tag as being dead and releases a ref on it.  If that
+ * wasn't the last reference, the garbage collector is poked to try and delete
+ * all keys that were in the domain.
+ */
+void key_remove_domain(struct key_tag *domain_tag)
+{
+	domain_tag->removed = true;
+	if (!key_put_tag(domain_tag))
+		key_schedule_gc_links();
+}
+
+/*
+ * Build the next index key chunk.
  *
  * We return it one word-sized chunk at a time.
  */
@@ -227,41 +269,33 @@ 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;
+	const u8 *d;
 	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);
+		return index_key->hash;
 	case 1:
-		return ((unsigned long)index_key->type << 8) | desc_len;
+		return index_key->x;
 	case 2:
-		if (desc_len == 0)
-			return (u8)((unsigned long)index_key->type >>
-				    (ASSOC_ARRAY_KEY_CHUNK_SIZE - 8));
-		n--;
-		offset = 1;
-		/* fall through */
+		return (unsigned long)index_key->type;
+	case 3:
+		return (unsigned long)index_key->domain_tag;
 	default:
-		offset += sizeof(chunk) - 1;
-		offset += (level - 3) * sizeof(chunk);
-		if (offset >= desc_len)
+		level -= 4;
+		if (desc_len <= sizeof(index_key->desc))
 			return 0;
-		desc_len -= offset;
+
+		d = index_key->description + sizeof(index_key->desc);
+		d += level * sizeof(long);
+		desc_len -= sizeof(index_key->desc);
 		if (desc_len > n)
 			desc_len = n;
-		offset += desc_len;
 		do {
 			chunk <<= 8;
-			chunk |= ((u8*)index_key->description)[--offset];
+			chunk |= *d++;
 		} while (--desc_len > 0);
-
-		if (level == 2) {
-			chunk <<= 8;
-			chunk |= (u8)((unsigned long)index_key->type >>
-				      (ASSOC_ARRAY_KEY_CHUNK_SIZE - 8));
-		}
 		return chunk;
 	}
 }
@@ -278,6 +312,7 @@ static bool keyring_compare_object(const void *object, const void *data)
 	const struct key *key = keyring_ptr_to_key(object);
 
 	return key->index_key.type == index_key->type &&
+		key->index_key.domain_tag == index_key->domain_tag &&
 		key->index_key.desc_len == index_key->desc_len &&
 		memcmp(key->index_key.description, index_key->description,
 		       index_key->desc_len) == 0;
@@ -296,43 +331,38 @@ static int keyring_diff_objects(const void *object, const void *data)
 	int level, i;
 
 	level = 0;
-	seg_a = hash_key_type_and_desc(a);
-	seg_b = hash_key_type_and_desc(b);
+	seg_a = a->hash;
+	seg_b = b->hash;
 	if ((seg_a ^ seg_b) != 0)
 		goto differ;
+	level += ASSOC_ARRAY_KEY_CHUNK_SIZE / 8;
 
 	/* 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;
+	seg_a = a->x;
+	seg_b = b->x;
 	if ((seg_a ^ seg_b) != 0)
 		goto differ;
+	level += sizeof(unsigned long);
 
 	/* 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);
 
+	seg_a = (unsigned long)a->domain_tag;
+	seg_b = (unsigned long)b->domain_tag;
+	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)));
-	}
+	i = sizeof(a->desc);
+	if (a->desc_len <= i)
+		goto same;
 
 	for (; i < a->desc_len; i++) {
 		seg_a = *(unsigned char *)(a->description + i);
@@ -658,6 +688,9 @@ static bool search_nested_keyrings(struct key *keyring,
 	BUG_ON((ctx->flags & STATE_CHECKS) == 0 ||
 	       (ctx->flags & STATE_CHECKS) == STATE_CHECKS);
 
+	if (ctx->index_key.description)
+		key_set_index_key(&ctx->index_key);
+
 	/* Check to see if this top-level keyring is what we are looking for
 	 * and whether it is valid or not.
 	 */
@@ -697,6 +730,9 @@ descend_to_keyring:
 	 * Non-keyrings avoid the leftmost branch of the root entirely (root
 	 * slots 1-15).
 	 */
+	if (!(ctx->flags & KEYRING_SEARCH_RECURSE))
+		goto not_this_keyring;
+
 	ptr = READ_ONCE(keyring->keys.root);
 	if (!ptr)
 		goto not_this_keyring;
@@ -897,13 +933,15 @@ key_ref_t keyring_search_rcu(key_ref_t keyring_ref,
  * @keyring: The root of the keyring tree to be searched.
  * @type: The type of keyring we want to find.
  * @description: The name of the keyring we want to find.
+ * @recurse: True to search the children of @keyring also
  *
  * As keyring_search_rcu() above, but using the current task's credentials and
  * type's default matching function and preferred search method.
  */
 key_ref_t keyring_search(key_ref_t keyring,
 			 struct key_type *type,
-			 const char *description)
+			 const char *description,
+			 bool recurse)
 {
 	struct keyring_search_context ctx = {
 		.index_key.type		= type,
@@ -918,6 +956,8 @@ key_ref_t keyring_search(key_ref_t keyring,
 	key_ref_t key;
 	int ret;
 
+	if (recurse)
+		ctx.flags |= KEYRING_SEARCH_RECURSE;
 	if (type->match_preparse) {
 		ret = type->match_preparse(&ctx.match_data);
 		if (ret < 0)
@@ -1102,50 +1142,44 @@ found:
  */
 struct key *find_keyring_by_name(const char *name, bool uid_keyring)
 {
+	struct user_namespace *ns = current_user_ns();
 	struct key *keyring;
-	int bucket;
 
 	if (!name)
 		return ERR_PTR(-EINVAL);
 
-	bucket = keyring_hash(name);
-
 	read_lock(&keyring_name_lock);
 
-	if (keyring_name_hash[bucket].next) {
-		/* search this hash bucket for a keyring with a matching name
-		 * that's readable and that hasn't been revoked */
-		list_for_each_entry(keyring,
-				    &keyring_name_hash[bucket],
-				    name_link
-				    ) {
-			if (!kuid_has_mapping(current_user_ns(), keyring->user->uid))
-				continue;
-
-			if (test_bit(KEY_FLAG_REVOKED, &keyring->flags))
-				continue;
+	/* Search this hash bucket for a keyring with a matching name that
+	 * grants Search permission and that hasn't been revoked
+	 */
+	list_for_each_entry(keyring, &ns->keyring_name_list, name_link) {
+		if (!kuid_has_mapping(ns, keyring->user->uid))
+			continue;
 
-			if (strcmp(keyring->description, name) != 0)
-				continue;
+		if (test_bit(KEY_FLAG_REVOKED, &keyring->flags))
+			continue;
 
-			if (uid_keyring) {
-				if (!test_bit(KEY_FLAG_UID_KEYRING,
-					      &keyring->flags))
-					continue;
-			} else {
-				if (key_permission(make_key_ref(keyring, 0),
-						   KEY_NEED_SEARCH) < 0)
-					continue;
-			}
+		if (strcmp(keyring->description, name) != 0)
+			continue;
 
-			/* we've got a match but we might end up racing with
-			 * key_cleanup() if the keyring is currently 'dead'
-			 * (ie. it has a zero usage count) */
-			if (!refcount_inc_not_zero(&keyring->usage))
+		if (uid_keyring) {
+			if (!test_bit(KEY_FLAG_UID_KEYRING,
+				      &keyring->flags))
+				continue;
+		} else {
+			if (key_permission(make_key_ref(keyring, 0),
+					   KEY_NEED_SEARCH) < 0)
 				continue;
-			keyring->last_used_at = ktime_get_real_seconds();
-			goto out;
 		}
+
+		/* we've got a match but we might end up racing with
+		 * key_cleanup() if the keyring is currently 'dead'
+		 * (ie. it has a zero usage count) */
+		if (!refcount_inc_not_zero(&keyring->usage))
+			continue;
+		keyring->last_used_at = ktime_get_real_seconds();
+		goto out;
 	}
 
 	keyring = ERR_PTR(-ENOKEY);
@@ -1188,7 +1222,8 @@ static int keyring_detect_cycle(struct key *A, struct key *B)
 		.flags			= (KEYRING_SEARCH_NO_STATE_CHECK |
 					   KEYRING_SEARCH_NO_UPDATE_TIME |
 					   KEYRING_SEARCH_NO_CHECK_PERM |
-					   KEYRING_SEARCH_DETECT_TOO_DEEP),
+					   KEYRING_SEARCH_DETECT_TOO_DEEP |
+					   KEYRING_SEARCH_RECURSE),
 	};
 
 	rcu_read_lock();