Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf

Alexei Starovoitov says:

====================
pull-request: bpf 2019-07-25

The following pull-request contains BPF updates for your *net* tree.

The main changes are:

1) fix segfault in libbpf, from Andrii.

2) fix gso_segs access, from Eric.

3) tls/sockmap fixes, from Jakub and John.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

6 files changed, 201 insertions, 66 deletions

diff --git a/net/core/filter.c b/net/core/filter.c
index 4e2a79b2fd77..7878f918b8c0 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -7455,12 +7455,12 @@ static u32 bpf_convert_ctx_access(enum bpf_access_type type,
 	case offsetof(struct __sk_buff, gso_segs):
 		/* si->dst_reg = skb_shinfo(SKB); */
 #ifdef NET_SKBUFF_DATA_USES_OFFSET
-		*insn++ = BPF_LDX_MEM(BPF_FIELD_SIZEOF(struct sk_buff, head),
-				      si->dst_reg, si->src_reg,
-				      offsetof(struct sk_buff, head));
 		*insn++ = BPF_LDX_MEM(BPF_FIELD_SIZEOF(struct sk_buff, end),
 				      BPF_REG_AX, si->src_reg,
 				      offsetof(struct sk_buff, end));
+		*insn++ = BPF_LDX_MEM(BPF_FIELD_SIZEOF(struct sk_buff, head),
+				      si->dst_reg, si->src_reg,
+				      offsetof(struct sk_buff, head));
 		*insn++ = BPF_ALU64_REG(BPF_ADD, si->dst_reg, BPF_REG_AX);
 #else
 		*insn++ = BPF_LDX_MEM(BPF_FIELD_SIZEOF(struct sk_buff, end),
diff --git a/net/core/skmsg.c b/net/core/skmsg.c
index 93bffaad2135..6832eeb4b785 100644
--- a/net/core/skmsg.c
+++ b/net/core/skmsg.c
@@ -585,12 +585,12 @@ EXPORT_SYMBOL_GPL(sk_psock_destroy);
 
 void sk_psock_drop(struct sock *sk, struct sk_psock *psock)
 {
-	rcu_assign_sk_user_data(sk, NULL);
 	sk_psock_cork_free(psock);
 	sk_psock_zap_ingress(psock);
-	sk_psock_restore_proto(sk, psock);
 
 	write_lock_bh(&sk->sk_callback_lock);
+	sk_psock_restore_proto(sk, psock);
+	rcu_assign_sk_user_data(sk, NULL);
 	if (psock->progs.skb_parser)
 		sk_psock_stop_strp(sk, psock);
 	write_unlock_bh(&sk->sk_callback_lock);
diff --git a/net/core/sock_map.c b/net/core/sock_map.c
index 52d4faeee18b..1330a7442e5b 100644
--- a/net/core/sock_map.c
+++ b/net/core/sock_map.c
@@ -247,6 +247,8 @@ static void sock_map_free(struct bpf_map *map)
 	raw_spin_unlock_bh(&stab->lock);
 	rcu_read_unlock();
 
+	synchronize_rcu();
+
 	bpf_map_area_free(stab->sks);
 	kfree(stab);
 }
@@ -276,16 +278,20 @@ static int __sock_map_delete(struct bpf_stab *stab, struct sock *sk_test,
 			     struct sock **psk)
 {
 	struct sock *sk;
+	int err = 0;
 
 	raw_spin_lock_bh(&stab->lock);
 	sk = *psk;
 	if (!sk_test || sk_test == sk)
-		*psk = NULL;
+		sk = xchg(psk, NULL);
+
+	if (likely(sk))
+		sock_map_unref(sk, psk);
+	else
+		err = -EINVAL;
+
 	raw_spin_unlock_bh(&stab->lock);
-	if (unlikely(!sk))
-		return -EINVAL;
-	sock_map_unref(sk, psk);
-	return 0;
+	return err;
 }
 
 static void sock_map_delete_from_link(struct bpf_map *map, struct sock *sk,
@@ -328,6 +334,7 @@ static int sock_map_update_common(struct bpf_map *map, u32 idx,
 				  struct sock *sk, u64 flags)
 {
 	struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
+	struct inet_connection_sock *icsk = inet_csk(sk);
 	struct sk_psock_link *link;
 	struct sk_psock *psock;
 	struct sock *osk;
@@ -338,6 +345,8 @@ static int sock_map_update_common(struct bpf_map *map, u32 idx,
 		return -EINVAL;
 	if (unlikely(idx >= map->max_entries))
 		return -E2BIG;
+	if (unlikely(icsk->icsk_ulp_data))
+		return -EINVAL;
 
 	link = sk_psock_init_link();
 	if (!link)
diff --git a/net/ipv4/tcp_ulp.c b/net/ipv4/tcp_ulp.c
index 3d8a1d835471..4849edb62d52 100644
--- a/net/ipv4/tcp_ulp.c
+++ b/net/ipv4/tcp_ulp.c
@@ -96,6 +96,19 @@ void tcp_get_available_ulp(char *buf, size_t maxlen)
 	rcu_read_unlock();
 }
 
+void tcp_update_ulp(struct sock *sk, struct proto *proto)
+{
+	struct inet_connection_sock *icsk = inet_csk(sk);
+
+	if (!icsk->icsk_ulp_ops) {
+		sk->sk_prot = proto;
+		return;
+	}
+
+	if (icsk->icsk_ulp_ops->update)
+		icsk->icsk_ulp_ops->update(sk, proto);
+}
+
 void tcp_cleanup_ulp(struct sock *sk)
 {
 	struct inet_connection_sock *icsk = inet_csk(sk);
diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c
index 4674e57e66b0..f208f8455ef2 100644
--- a/net/tls/tls_main.c
+++ b/net/tls/tls_main.c
@@ -261,24 +261,36 @@ void tls_ctx_free(struct tls_context *ctx)
 	kfree(ctx);
 }
 
-static void tls_sk_proto_close(struct sock *sk, long timeout)
+static void tls_ctx_free_deferred(struct work_struct *gc)
 {
-	struct tls_context *ctx = tls_get_ctx(sk);
-	long timeo = sock_sndtimeo(sk, 0);
-	void (*sk_proto_close)(struct sock *sk, long timeout);
-	bool free_ctx = false;
-
-	lock_sock(sk);
-	sk_proto_close = ctx->sk_proto_close;
+	struct tls_context *ctx = container_of(gc, struct tls_context, gc);
 
-	if (ctx->tx_conf == TLS_HW_RECORD && ctx->rx_conf == TLS_HW_RECORD)
-		goto skip_tx_cleanup;
+	/* Ensure any remaining work items are completed. The sk will
+	 * already have lost its tls_ctx reference by the time we get
+	 * here so no xmit operation will actually be performed.
+	 */
+	if (ctx->tx_conf == TLS_SW) {
+		tls_sw_cancel_work_tx(ctx);
+		tls_sw_free_ctx_tx(ctx);
+	}
 
-	if (ctx->tx_conf == TLS_BASE && ctx->rx_conf == TLS_BASE) {
-		free_ctx = true;
-		goto skip_tx_cleanup;
+	if (ctx->rx_conf == TLS_SW) {
+		tls_sw_strparser_done(ctx);
+		tls_sw_free_ctx_rx(ctx);
 	}
 
+	tls_ctx_free(ctx);
+}
+
+static void tls_ctx_free_wq(struct tls_context *ctx)
+{
+	INIT_WORK(&ctx->gc, tls_ctx_free_deferred);
+	schedule_work(&ctx->gc);
+}
+
+static void tls_sk_proto_cleanup(struct sock *sk,
+				 struct tls_context *ctx, long timeo)
+{
 	if (unlikely(sk->sk_write_pending) &&
 	    !wait_on_pending_writer(sk, &timeo))
 		tls_handle_open_record(sk, 0);
@@ -287,7 +299,7 @@ static void tls_sk_proto_close(struct sock *sk, long timeout)
 	if (ctx->tx_conf == TLS_SW) {
 		kfree(ctx->tx.rec_seq);
 		kfree(ctx->tx.iv);
-		tls_sw_free_resources_tx(sk);
+		tls_sw_release_resources_tx(sk);
 #ifdef CONFIG_TLS_DEVICE
 	} else if (ctx->tx_conf == TLS_HW) {
 		tls_device_free_resources_tx(sk);
@@ -295,26 +307,67 @@ static void tls_sk_proto_close(struct sock *sk, long timeout)
 	}
 
 	if (ctx->rx_conf == TLS_SW)
-		tls_sw_free_resources_rx(sk);
+		tls_sw_release_resources_rx(sk);
 
 #ifdef CONFIG_TLS_DEVICE
 	if (ctx->rx_conf == TLS_HW)
 		tls_device_offload_cleanup_rx(sk);
-
-	if (ctx->tx_conf != TLS_HW && ctx->rx_conf != TLS_HW) {
-#else
-	{
 #endif
-		tls_ctx_free(ctx);
-		ctx = NULL;
+}
+
+static void tls_sk_proto_unhash(struct sock *sk)
+{
+	struct inet_connection_sock *icsk = inet_csk(sk);
+	long timeo = sock_sndtimeo(sk, 0);
+	struct tls_context *ctx;
+
+	if (unlikely(!icsk->icsk_ulp_data)) {
+		if (sk->sk_prot->unhash)
+			sk->sk_prot->unhash(sk);
 	}
 
-skip_tx_cleanup:
+	ctx = tls_get_ctx(sk);
+	tls_sk_proto_cleanup(sk, ctx, timeo);
+	write_lock_bh(&sk->sk_callback_lock);
+	icsk->icsk_ulp_data = NULL;
+	sk->sk_prot = ctx->sk_proto;
+	write_unlock_bh(&sk->sk_callback_lock);
+
+	if (ctx->sk_proto->unhash)
+		ctx->sk_proto->unhash(sk);
+	tls_ctx_free_wq(ctx);
+}
+
+static void tls_sk_proto_close(struct sock *sk, long timeout)
+{
+	struct inet_connection_sock *icsk = inet_csk(sk);
+	struct tls_context *ctx = tls_get_ctx(sk);
+	long timeo = sock_sndtimeo(sk, 0);
+	bool free_ctx;
+
+	if (ctx->tx_conf == TLS_SW)
+		tls_sw_cancel_work_tx(ctx);
+
+	lock_sock(sk);
+	free_ctx = ctx->tx_conf != TLS_HW && ctx->rx_conf != TLS_HW;
+
+	if (ctx->tx_conf != TLS_BASE || ctx->rx_conf != TLS_BASE)
+		tls_sk_proto_cleanup(sk, ctx, timeo);
+
+	write_lock_bh(&sk->sk_callback_lock);
+	if (free_ctx)
+		icsk->icsk_ulp_data = NULL;
+	sk->sk_prot = ctx->sk_proto;
+	write_unlock_bh(&sk->sk_callback_lock);
 	release_sock(sk);
-	sk_proto_close(sk, timeout);
-	/* free ctx for TLS_HW_RECORD, used by tcp_set_state
-	 * for sk->sk_prot->unhash [tls_hw_unhash]
-	 */
+	if (ctx->tx_conf == TLS_SW)
+		tls_sw_free_ctx_tx(ctx);
+	if (ctx->rx_conf == TLS_SW || ctx->rx_conf == TLS_HW)
+		tls_sw_strparser_done(ctx);
+	if (ctx->rx_conf == TLS_SW)
+		tls_sw_free_ctx_rx(ctx);
+	ctx->sk_proto_close(sk, timeout);
+
 	if (free_ctx)
 		tls_ctx_free(ctx);
 }
@@ -526,6 +579,8 @@ static int do_tls_setsockopt_conf(struct sock *sk, char __user *optval,
 		{
 #endif
 			rc = tls_set_sw_offload(sk, ctx, 1);
+			if (rc)
+				goto err_crypto_info;
 			conf = TLS_SW;
 		}
 	} else {
@@ -537,13 +592,13 @@ static int do_tls_setsockopt_conf(struct sock *sk, char __user *optval,
 		{
 #endif
 			rc = tls_set_sw_offload(sk, ctx, 0);
+			if (rc)
+				goto err_crypto_info;
 			conf = TLS_SW;
 		}
+		tls_sw_strparser_arm(sk, ctx);
 	}
 
-	if (rc)
-		goto err_crypto_info;
-
 	if (tx)
 		ctx->tx_conf = conf;
 	else
@@ -607,6 +662,7 @@ static struct tls_context *create_ctx(struct sock *sk)
 	ctx->setsockopt = sk->sk_prot->setsockopt;
 	ctx->getsockopt = sk->sk_prot->getsockopt;
 	ctx->sk_proto_close = sk->sk_prot->close;
+	ctx->unhash = sk->sk_prot->unhash;
 	return ctx;
 }
 
@@ -730,6 +786,7 @@ static void build_protos(struct proto prot[TLS_NUM_CONFIG][TLS_NUM_CONFIG],
 	prot[TLS_BASE][TLS_BASE].setsockopt	= tls_setsockopt;
 	prot[TLS_BASE][TLS_BASE].getsockopt	= tls_getsockopt;
 	prot[TLS_BASE][TLS_BASE].close		= tls_sk_proto_close;
+	prot[TLS_BASE][TLS_BASE].unhash		= tls_sk_proto_unhash;
 
 	prot[TLS_SW][TLS_BASE] = prot[TLS_BASE][TLS_BASE];
 	prot[TLS_SW][TLS_BASE].sendmsg		= tls_sw_sendmsg;
@@ -747,16 +804,20 @@ static void build_protos(struct proto prot[TLS_NUM_CONFIG][TLS_NUM_CONFIG],
 
 #ifdef CONFIG_TLS_DEVICE
 	prot[TLS_HW][TLS_BASE] = prot[TLS_BASE][TLS_BASE];
+	prot[TLS_HW][TLS_BASE].unhash		= base->unhash;
 	prot[TLS_HW][TLS_BASE].sendmsg		= tls_device_sendmsg;
 	prot[TLS_HW][TLS_BASE].sendpage		= tls_device_sendpage;
 
 	prot[TLS_HW][TLS_SW] = prot[TLS_BASE][TLS_SW];
+	prot[TLS_HW][TLS_SW].unhash		= base->unhash;
 	prot[TLS_HW][TLS_SW].sendmsg		= tls_device_sendmsg;
 	prot[TLS_HW][TLS_SW].sendpage		= tls_device_sendpage;
 
 	prot[TLS_BASE][TLS_HW] = prot[TLS_BASE][TLS_SW];
+	prot[TLS_BASE][TLS_HW].unhash		= base->unhash;
 
 	prot[TLS_SW][TLS_HW] = prot[TLS_SW][TLS_SW];
+	prot[TLS_SW][TLS_HW].unhash		= base->unhash;
 
 	prot[TLS_HW][TLS_HW] = prot[TLS_HW][TLS_SW];
 #endif
@@ -764,7 +825,6 @@ static void build_protos(struct proto prot[TLS_NUM_CONFIG][TLS_NUM_CONFIG],
 	prot[TLS_HW_RECORD][TLS_HW_RECORD] = *base;
 	prot[TLS_HW_RECORD][TLS_HW_RECORD].hash		= tls_hw_hash;
 	prot[TLS_HW_RECORD][TLS_HW_RECORD].unhash	= tls_hw_unhash;
-	prot[TLS_HW_RECORD][TLS_HW_RECORD].close	= tls_sk_proto_close;
 }
 
 static int tls_init(struct sock *sk)
@@ -773,7 +833,7 @@ static int tls_init(struct sock *sk)
 	int rc = 0;
 
 	if (tls_hw_prot(sk))
-		goto out;
+		return 0;
 
 	/* The TLS ulp is currently supported only for TCP sockets
 	 * in ESTABLISHED state.
@@ -784,21 +844,38 @@ static int tls_init(struct sock *sk)
 	if (sk->sk_state != TCP_ESTABLISHED)
 		return -ENOTSUPP;
 
+	tls_build_proto(sk);
+
 	/* allocate tls context */
+	write_lock_bh(&sk->sk_callback_lock);
 	ctx = create_ctx(sk);
 	if (!ctx) {
 		rc = -ENOMEM;
 		goto out;
 	}
 
-	tls_build_proto(sk);
 	ctx->tx_conf = TLS_BASE;
 	ctx->rx_conf = TLS_BASE;
+	ctx->sk_proto = sk->sk_prot;
 	update_sk_prot(sk, ctx);
 out:
+	write_unlock_bh(&sk->sk_callback_lock);
 	return rc;
 }
 
+static void tls_update(struct sock *sk, struct proto *p)
+{
+	struct tls_context *ctx;
+
+	ctx = tls_get_ctx(sk);
+	if (likely(ctx)) {
+		ctx->sk_proto_close = p->close;
+		ctx->sk_proto = p;
+	} else {
+		sk->sk_prot = p;
+	}
+}
+
 void tls_register_device(struct tls_device *device)
 {
 	spin_lock_bh(&device_spinlock);
@@ -819,6 +896,7 @@ static struct tcp_ulp_ops tcp_tls_ulp_ops __read_mostly = {
 	.name			= "tls",
 	.owner			= THIS_MODULE,
 	.init			= tls_init,
+	.update			= tls_update,
 };
 
 static int __init tls_register(void)
diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c
index 53b4ad94e74a..91d21b048a9b 100644
--- a/net/tls/tls_sw.c
+++ b/net/tls/tls_sw.c
@@ -2054,7 +2054,16 @@ static void tls_data_ready(struct sock *sk)
 	}
 }
 
-void tls_sw_free_resources_tx(struct sock *sk)
+void tls_sw_cancel_work_tx(struct tls_context *tls_ctx)
+{
+	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
+
+	set_bit(BIT_TX_CLOSING, &ctx->tx_bitmask);
+	set_bit(BIT_TX_SCHEDULED, &ctx->tx_bitmask);
+	cancel_delayed_work_sync(&ctx->tx_work.work);
+}
+
+void tls_sw_release_resources_tx(struct sock *sk)
 {
 	struct tls_context *tls_ctx = tls_get_ctx(sk);
 	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
@@ -2065,11 +2074,6 @@ void tls_sw_free_resources_tx(struct sock *sk)
 	if (atomic_read(&ctx->encrypt_pending))
 		crypto_wait_req(-EINPROGRESS, &ctx->async_wait);
 
-	release_sock(sk);
-	cancel_delayed_work_sync(&ctx->tx_work.work);
-	lock_sock(sk);
-
-	/* Tx whatever records we can transmit and abandon the rest */
 	tls_tx_records(sk, -1);
 
 	/* Free up un-sent records in tx_list. First, free
@@ -2092,6 +2096,11 @@ void tls_sw_free_resources_tx(struct sock *sk)
 
 	crypto_free_aead(ctx->aead_send);
 	tls_free_open_rec(sk);
+}
+
+void tls_sw_free_ctx_tx(struct tls_context *tls_ctx)
+{
+	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
 
 	kfree(ctx);
 }
@@ -2110,25 +2119,40 @@ void tls_sw_release_resources_rx(struct sock *sk)
 		skb_queue_purge(&ctx->rx_list);
 		crypto_free_aead(ctx->aead_recv);
 		strp_stop(&ctx->strp);
-		write_lock_bh(&sk->sk_callback_lock);
-		sk->sk_data_ready = ctx->saved_data_ready;
-		write_unlock_bh(&sk->sk_callback_lock);
-		release_sock(sk);
-		strp_done(&ctx->strp);
-		lock_sock(sk);
+		/* If tls_sw_strparser_arm() was not called (cleanup paths)
+		 * we still want to strp_stop(), but sk->sk_data_ready was
+		 * never swapped.
+		 */
+		if (ctx->saved_data_ready) {
+			write_lock_bh(&sk->sk_callback_lock);
+			sk->sk_data_ready = ctx->saved_data_ready;
+			write_unlock_bh(&sk->sk_callback_lock);
+		}
 	}
 }
 
-void tls_sw_free_resources_rx(struct sock *sk)
+void tls_sw_strparser_done(struct tls_context *tls_ctx)
 {
-	struct tls_context *tls_ctx = tls_get_ctx(sk);
 	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
 
-	tls_sw_release_resources_rx(sk);
+	strp_done(&ctx->strp);
+}
+
+void tls_sw_free_ctx_rx(struct tls_context *tls_ctx)
+{
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
 
 	kfree(ctx);
 }
 
+void tls_sw_free_resources_rx(struct sock *sk)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+
+	tls_sw_release_resources_rx(sk);
+	tls_sw_free_ctx_rx(tls_ctx);
+}
+
 /* The work handler to transmitt the encrypted records in tx_list */
 static void tx_work_handler(struct work_struct *work)
 {
@@ -2137,11 +2161,17 @@ static void tx_work_handler(struct work_struct *work)
 					       struct tx_work, work);
 	struct sock *sk = tx_work->sk;
 	struct tls_context *tls_ctx = tls_get_ctx(sk);
-	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
+	struct tls_sw_context_tx *ctx;
 
-	if (!test_and_clear_bit(BIT_TX_SCHEDULED, &ctx->tx_bitmask))
+	if (unlikely(!tls_ctx))
 		return;
 
+	ctx = tls_sw_ctx_tx(tls_ctx);
+	if (test_bit(BIT_TX_CLOSING, &ctx->tx_bitmask))
+		return;
+
+	if (!test_and_clear_bit(BIT_TX_SCHEDULED, &ctx->tx_bitmask))
+		return;
 	lock_sock(sk);
 	tls_tx_records(sk, -1);
 	release_sock(sk);
@@ -2160,6 +2190,18 @@ void tls_sw_write_space(struct sock *sk, struct tls_context *ctx)
 	}
 }
 
+void tls_sw_strparser_arm(struct sock *sk, struct tls_context *tls_ctx)
+{
+	struct tls_sw_context_rx *rx_ctx = tls_sw_ctx_rx(tls_ctx);
+
+	write_lock_bh(&sk->sk_callback_lock);
+	rx_ctx->saved_data_ready = sk->sk_data_ready;
+	sk->sk_data_ready = tls_data_ready;
+	write_unlock_bh(&sk->sk_callback_lock);
+
+	strp_check_rcv(&rx_ctx->strp);
+}
+
 int tls_set_sw_offload(struct sock *sk, struct tls_context *ctx, int tx)
 {
 	struct tls_context *tls_ctx = tls_get_ctx(sk);
@@ -2357,13 +2399,6 @@ int tls_set_sw_offload(struct sock *sk, struct tls_context *ctx, int tx)
 		cb.parse_msg = tls_read_size;
 
 		strp_init(&sw_ctx_rx->strp, sk, &cb);
-
-		write_lock_bh(&sk->sk_callback_lock);
-		sw_ctx_rx->saved_data_ready = sk->sk_data_ready;
-		sk->sk_data_ready = tls_data_ready;
-		write_unlock_bh(&sk->sk_callback_lock);
-
-		strp_check_rcv(&sw_ctx_rx->strp);
 	}
 
 	goto out;