Blob Blame History Raw
From: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Date: Wed, 24 Nov 2021 11:14:29 +0900
Subject: af_unix: Save hash in sk_hash.
Patch-mainline: v5.17-rc1
Git-commit: e6b4b873896f0e9298f70d25726f4bb1e1b265ba
References: jsc#PED-1377

To replace unix_table_lock with per-hash locks in the next patch, we need
to save a hash in each socket because /proc/net/unix or BPF prog iterate
sockets while holding a hash table lock and release it later in a different
function.

Currently, we store a real/pseudo hash in struct unix_address.  However, we
do not allocate it to unbound sockets, nor should we do just for that.  For
this purpose, we can use sk_hash.  Then, we no longer use the hash field in
struct unix_address and can remove it.

Also, this patch does
  - rename unix_insert_socket() to unix_insert_unbound_socket()
  - remove the redundant list argument from __unix_insert_socket() and
     unix_insert_unbound_socket()
  - use 'unsigned int' instead of 'unsigned' in __unix_set_addr_hash()
  - remove 'inline' from unix_remove_socket() and
     unix_insert_unbound_socket().

Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Shung-Hsi Yu <shung-hsi.yu@suse.com>
---
 include/net/af_unix.h |    1 -
 net/unix/af_unix.c    |   42 +++++++++++++++++++++++-------------------
 2 files changed, 23 insertions(+), 20 deletions(-)

--- a/include/net/af_unix.h
+++ b/include/net/af_unix.h
@@ -26,7 +26,6 @@ extern struct hlist_head unix_socket_tab
 struct unix_address {
 	refcount_t	refcnt;
 	int		len;
-	unsigned int	hash;
 	struct sockaddr_un name[];
 };
 
--- a/net/unix/af_unix.c
+++ b/net/unix/af_unix.c
@@ -280,31 +280,33 @@ static void __unix_remove_socket(struct
 	sk_del_node_init(sk);
 }
 
-static void __unix_insert_socket(struct hlist_head *list, struct sock *sk)
+static void __unix_insert_socket(struct sock *sk)
 {
 	WARN_ON(!sk_unhashed(sk));
-	sk_add_node(sk, list);
+	sk_add_node(sk, &unix_socket_table[sk->sk_hash]);
 }
 
-static void __unix_set_addr(struct sock *sk, struct unix_address *addr,
-			    unsigned hash)
+static void __unix_set_addr_hash(struct sock *sk, struct unix_address *addr,
+				 unsigned int hash)
 {
 	__unix_remove_socket(sk);
 	smp_store_release(&unix_sk(sk)->addr, addr);
-	__unix_insert_socket(&unix_socket_table[hash], sk);
+
+	sk->sk_hash = hash;
+	__unix_insert_socket(sk);
 }
 
-static inline void unix_remove_socket(struct sock *sk)
+static void unix_remove_socket(struct sock *sk)
 {
 	spin_lock(&unix_table_lock);
 	__unix_remove_socket(sk);
 	spin_unlock(&unix_table_lock);
 }
 
-static inline void unix_insert_socket(struct hlist_head *list, struct sock *sk)
+static void unix_insert_unbound_socket(struct sock *sk)
 {
 	spin_lock(&unix_table_lock);
-	__unix_insert_socket(list, sk);
+	__unix_insert_socket(sk);
 	spin_unlock(&unix_table_lock);
 }
 
@@ -862,6 +864,7 @@ static struct sock *unix_create1(struct
 
 	sock_init_data(sock, sk);
 
+	sk->sk_hash		= unix_unbound_hash(sk);
 	sk->sk_allocation	= GFP_KERNEL_ACCOUNT;
 	sk->sk_write_space	= unix_write_space;
 	sk->sk_max_ack_backlog	= net->unx.sysctl_max_dgram_qlen;
@@ -877,7 +880,7 @@ static struct sock *unix_create1(struct
 	init_waitqueue_head(&u->peer_wait);
 	init_waitqueue_func_entry(&u->peer_wake, unix_dgram_peer_wake_relay);
 	memset(&u->scm_stat, 0, sizeof(struct scm_stat));
-	unix_insert_socket(&unix_socket_table[unix_unbound_hash(sk)], sk);
+	unix_insert_unbound_socket(sk);
 out:
 	if (sk == NULL)
 		atomic_long_dec(&unix_nr_socks);
@@ -1018,6 +1021,7 @@ static int unix_autobind(struct sock *sk
 	struct unix_address *addr;
 	unsigned int retries = 0;
 	static u32 ordernum = 1;
+	unsigned int new_hash;
 	int err;
 
 	err = mutex_lock_interruptible(&u->bindlock);
@@ -1039,13 +1043,13 @@ static int unix_autobind(struct sock *sk
 retry:
 	addr->len = sprintf(addr->name->sun_path + 1, "%05x", ordernum) +
 		offsetof(struct sockaddr_un, sun_path) + 1;
-	addr->hash = unix_abstract_hash(addr->name, addr->len, sk->sk_type);
 
+	new_hash = unix_abstract_hash(addr->name, addr->len, sk->sk_type);
 	spin_lock(&unix_table_lock);
 	ordernum = (ordernum+1)&0xFFFFF;
 
 	if (__unix_find_socket_byname(sock_net(sk), addr->name, addr->len,
-				      addr->hash)) {
+				      new_hash)) {
 		spin_unlock(&unix_table_lock);
 		/*
 		 * __unix_find_socket_byname() may take long time if many names
@@ -1061,7 +1065,7 @@ retry:
 		goto retry;
 	}
 
-	__unix_set_addr(sk, addr, addr->hash);
+	__unix_set_addr_hash(sk, addr, new_hash);
 	spin_unlock(&unix_table_lock);
 	err = 0;
 
@@ -1077,9 +1081,9 @@ static int unix_bind_bsd(struct sock *sk
 	struct unix_sock *u = unix_sk(sk);
 	struct user_namespace *ns; // barf...
 	struct unix_address *addr;
+	unsigned int new_hash;
 	struct dentry *dentry;
 	struct path parent;
-	unsigned int hash;
 	int err;
 
 	unix_mkname_bsd(sunaddr, addr_len);
@@ -1115,12 +1119,11 @@ static int unix_bind_bsd(struct sock *sk
 	if (u->addr)
 		goto out_unlock;
 
-	addr->hash = UNIX_HASH_SIZE;
-	hash = unix_bsd_hash(d_backing_inode(dentry));
+	new_hash = unix_bsd_hash(d_backing_inode(dentry));
 	spin_lock(&unix_table_lock);
 	u->path.mnt = mntget(parent.mnt);
 	u->path.dentry = dget(dentry);
-	__unix_set_addr(sk, addr, hash);
+	__unix_set_addr_hash(sk, addr, new_hash);
 	spin_unlock(&unix_table_lock);
 	mutex_unlock(&u->bindlock);
 	done_path_create(&parent, dentry);
@@ -1144,6 +1147,7 @@ static int unix_bind_abstract(struct soc
 {
 	struct unix_sock *u = unix_sk(sk);
 	struct unix_address *addr;
+	unsigned int new_hash;
 	int err;
 
 	addr = unix_create_addr(sunaddr, addr_len);
@@ -1159,14 +1163,14 @@ static int unix_bind_abstract(struct soc
 		goto out_mutex;
 	}
 
-	addr->hash = unix_abstract_hash(addr->name, addr->len, sk->sk_type);
+	new_hash = unix_abstract_hash(addr->name, addr->len, sk->sk_type);
 	spin_lock(&unix_table_lock);
 
 	if (__unix_find_socket_byname(sock_net(sk), addr->name, addr->len,
-				      addr->hash))
+				      new_hash))
 		goto out_spin;
 
-	__unix_set_addr(sk, addr, addr->hash);
+	__unix_set_addr_hash(sk, addr, new_hash);
 	spin_unlock(&unix_table_lock);
 	mutex_unlock(&u->bindlock);
 	return 0;