2 files changed, 105 insertions, 90 deletions
diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c
index 1c2c6800949d..b4674f03d71a 100644
--- a/net/tls/tls_main.c
+++ b/net/tls/tls_main.c
@@ -1003,7 +1003,7 @@ static u16 tls_user_config(struct tls_context *ctx, bool tx)
 	return 0;
 }
 
-static int tls_get_info(const struct sock *sk, struct sk_buff *skb)
+static int tls_get_info(struct sock *sk, struct sk_buff *skb)
 {
 	u16 version, cipher_type;
 	struct tls_context *ctx;
diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c
index 31e8a94dfc11..211f57164cb6 100644
--- a/net/tls/tls_sw.c
+++ b/net/tls/tls_sw.c
@@ -52,6 +52,7 @@ struct tls_decrypt_arg {
 	struct_group(inargs,
 	bool zc;
 	bool async;
+	bool async_done;
 	u8 tail;
 	);
 
@@ -63,6 +64,7 @@ struct tls_decrypt_ctx {
 	u8 iv[TLS_MAX_IV_SIZE];
 	u8 aad[TLS_MAX_AAD_SIZE];
 	u8 tail;
+	bool free_sgout;
 	struct scatterlist sg[];
 };
 
@@ -187,7 +189,6 @@ static void tls_decrypt_done(void *data, int err)
 	struct aead_request *aead_req = data;
 	struct crypto_aead *aead = crypto_aead_reqtfm(aead_req);
 	struct scatterlist *sgout = aead_req->dst;
-	struct scatterlist *sgin = aead_req->src;
 	struct tls_sw_context_rx *ctx;
 	struct tls_decrypt_ctx *dctx;
 	struct tls_context *tls_ctx;
@@ -196,6 +197,17 @@ static void tls_decrypt_done(void *data, int err)
 	struct sock *sk;
 	int aead_size;
 
+	/* If requests get too backlogged crypto API returns -EBUSY and calls
+	 * ->complete(-EINPROGRESS) immediately followed by ->complete(0)
+	 * to make waiting for backlog to flush with crypto_wait_req() easier.
+	 * First wait converts -EBUSY -> -EINPROGRESS, and the second one
+	 * -EINPROGRESS -> 0.
+	 * We have a single struct crypto_async_request per direction, this
+	 * scheme doesn't help us, so just ignore the first ->complete().
+	 */
+	if (err == -EINPROGRESS)
+		return;
+
 	aead_size = sizeof(*aead_req) + crypto_aead_reqsize(aead);
 	aead_size = ALIGN(aead_size, __alignof__(*dctx));
 	dctx = (void *)((u8 *)aead_req + aead_size);
@@ -213,7 +225,7 @@ static void tls_decrypt_done(void *data, int err)
 	}
 
 	/* Free the destination pages if skb was not decrypted inplace */
-	if (sgout != sgin) {
+	if (dctx->free_sgout) {
 		/* Skip the first S/G entry as it points to AAD */
 		for_each_sg(sg_next(sgout), sg, UINT_MAX, pages) {
 			if (!sg)
@@ -224,10 +236,17 @@ static void tls_decrypt_done(void *data, int err)
 
 	kfree(aead_req);
 
-	spin_lock_bh(&ctx->decrypt_compl_lock);
-	if (!atomic_dec_return(&ctx->decrypt_pending))
+	if (atomic_dec_and_test(&ctx->decrypt_pending))
 		complete(&ctx->async_wait.completion);
-	spin_unlock_bh(&ctx->decrypt_compl_lock);
+}
+
+static int tls_decrypt_async_wait(struct tls_sw_context_rx *ctx)
+{
+	if (!atomic_dec_and_test(&ctx->decrypt_pending))
+		crypto_wait_req(-EINPROGRESS, &ctx->async_wait);
+	atomic_inc(&ctx->decrypt_pending);
+
+	return ctx->async_wait.err;
 }
 
 static int tls_do_decryption(struct sock *sk,
@@ -253,20 +272,33 @@ static int tls_do_decryption(struct sock *sk,
 		aead_request_set_callback(aead_req,
 					  CRYPTO_TFM_REQ_MAY_BACKLOG,
 					  tls_decrypt_done, aead_req);
+		DEBUG_NET_WARN_ON_ONCE(atomic_read(&ctx->decrypt_pending) < 1);
 		atomic_inc(&ctx->decrypt_pending);
 	} else {
+		DECLARE_CRYPTO_WAIT(wait);
+
 		aead_request_set_callback(aead_req,
 					  CRYPTO_TFM_REQ_MAY_BACKLOG,
-					  crypto_req_done, &ctx->async_wait);
+					  crypto_req_done, &wait);
+		ret = crypto_aead_decrypt(aead_req);
+		if (ret == -EINPROGRESS || ret == -EBUSY)
+			ret = crypto_wait_req(ret, &wait);
+		return ret;
 	}
 
 	ret = crypto_aead_decrypt(aead_req);
-	if (ret == -EINPROGRESS) {
-		if (darg->async)
-			return 0;
+	if (ret == -EINPROGRESS)
+		return 0;
 
-		ret = crypto_wait_req(ret, &ctx->async_wait);
+	if (ret == -EBUSY) {
+		ret = tls_decrypt_async_wait(ctx);
+		darg->async_done = true;
+		/* all completions have run, we're not doing async anymore */
+		darg->async = false;
+		return ret;
 	}
+
+	atomic_dec(&ctx->decrypt_pending);
 	darg->async = false;
 
 	return ret;
@@ -439,9 +471,10 @@ static void tls_encrypt_done(void *data, int err)
 	struct tls_rec *rec = data;
 	struct scatterlist *sge;
 	struct sk_msg *msg_en;
-	bool ready = false;
 	struct sock *sk;
-	int pending;
+
+	if (err == -EINPROGRESS) /* see the comment in tls_decrypt_done() */
+		return;
 
 	msg_en = &rec->msg_encrypted;
 
@@ -476,23 +509,25 @@ static void tls_encrypt_done(void *data, int err)
 		/* If received record is at head of tx_list, schedule tx */
 		first_rec = list_first_entry(&ctx->tx_list,
 					     struct tls_rec, list);
-		if (rec == first_rec)
-			ready = true;
+		if (rec == first_rec) {
+			/* Schedule the transmission */
+			if (!test_and_set_bit(BIT_TX_SCHEDULED,
+					      &ctx->tx_bitmask))
+				schedule_delayed_work(&ctx->tx_work.work, 1);
+		}
 	}
 
-	spin_lock_bh(&ctx->encrypt_compl_lock);
-	pending = atomic_dec_return(&ctx->encrypt_pending);
-
-	if (!pending && ctx->async_notify)
+	if (atomic_dec_and_test(&ctx->encrypt_pending))
 		complete(&ctx->async_wait.completion);
-	spin_unlock_bh(&ctx->encrypt_compl_lock);
+}
 
-	if (!ready)
-		return;
+static int tls_encrypt_async_wait(struct tls_sw_context_tx *ctx)
+{
+	if (!atomic_dec_and_test(&ctx->encrypt_pending))
+		crypto_wait_req(-EINPROGRESS, &ctx->async_wait);
+	atomic_inc(&ctx->encrypt_pending);
 
-	/* Schedule the transmission */
-	if (!test_and_set_bit(BIT_TX_SCHEDULED, &ctx->tx_bitmask))
-		schedule_delayed_work(&ctx->tx_work.work, 1);
+	return ctx->async_wait.err;
 }
 
 static int tls_do_encryption(struct sock *sk,
@@ -541,9 +576,14 @@ static int tls_do_encryption(struct sock *sk,
 
 	/* Add the record in tx_list */
 	list_add_tail((struct list_head *)&rec->list, &ctx->tx_list);
+	DEBUG_NET_WARN_ON_ONCE(atomic_read(&ctx->encrypt_pending) < 1);
 	atomic_inc(&ctx->encrypt_pending);
 
 	rc = crypto_aead_encrypt(aead_req);
+	if (rc == -EBUSY) {
+		rc = tls_encrypt_async_wait(ctx);
+		rc = rc ?: -EINPROGRESS;
+	}
 	if (!rc || rc != -EINPROGRESS) {
 		atomic_dec(&ctx->encrypt_pending);
 		sge->offset -= prot->prepend_size;
@@ -984,7 +1024,6 @@ static int tls_sw_sendmsg_locked(struct sock *sk, struct msghdr *msg,
 	int num_zc = 0;
 	int orig_size;
 	int ret = 0;
-	int pending;
 
 	if (!eor && (msg->msg_flags & MSG_EOR))
 		return -EINVAL;
@@ -1163,24 +1202,12 @@ trim_sgl:
 	if (!num_async) {
 		goto send_end;
 	} else if (num_zc) {
-		/* Wait for pending encryptions to get completed */
-		spin_lock_bh(&ctx->encrypt_compl_lock);
-		ctx->async_notify = true;
-
-		pending = atomic_read(&ctx->encrypt_pending);
-		spin_unlock_bh(&ctx->encrypt_compl_lock);
-		if (pending)
-			crypto_wait_req(-EINPROGRESS, &ctx->async_wait);
-		else
-			reinit_completion(&ctx->async_wait.completion);
-
-		/* There can be no concurrent accesses, since we have no
-		 * pending encrypt operations
-		 */
-		WRITE_ONCE(ctx->async_notify, false);
+		int err;
 
-		if (ctx->async_wait.err) {
-			ret = ctx->async_wait.err;
+		/* Wait for pending encryptions to get completed */
+		err = tls_encrypt_async_wait(ctx);
+		if (err) {
+			ret = err;
 			copied = 0;
 		}
 	}
@@ -1229,7 +1256,6 @@ void tls_sw_splice_eof(struct socket *sock)
 	ssize_t copied = 0;
 	bool retrying = false;
 	int ret = 0;
-	int pending;
 
 	if (!ctx->open_rec)
 		return;
@@ -1264,22 +1290,7 @@ retry:
 	}
 
 	/* Wait for pending encryptions to get completed */
-	spin_lock_bh(&ctx->encrypt_compl_lock);
-	ctx->async_notify = true;
-
-	pending = atomic_read(&ctx->encrypt_pending);
-	spin_unlock_bh(&ctx->encrypt_compl_lock);
-	if (pending)
-		crypto_wait_req(-EINPROGRESS, &ctx->async_wait);
-	else
-		reinit_completion(&ctx->async_wait.completion);
-
-	/* There can be no concurrent accesses, since we have no pending
-	 * encrypt operations
-	 */
-	WRITE_ONCE(ctx->async_notify, false);
-
-	if (ctx->async_wait.err)
+	if (tls_encrypt_async_wait(ctx))
 		goto unlock;
 
 	/* Transmit if any encryptions have completed */
@@ -1581,12 +1592,16 @@ static int tls_decrypt_sg(struct sock *sk, struct iov_iter *out_iov,
 	} else if (out_sg) {
 		memcpy(sgout, out_sg, n_sgout * sizeof(*sgout));
 	}
+	dctx->free_sgout = !!pages;
 
 	/* Prepare and submit AEAD request */
 	err = tls_do_decryption(sk, sgin, sgout, dctx->iv,
 				data_len + prot->tail_size, aead_req, darg);
-	if (err)
+	if (err) {
+		if (darg->async_done)
+			goto exit_free_skb;
 		goto exit_free_pages;
+	}
 
 	darg->skb = clear_skb ?: tls_strp_msg(ctx);
 	clear_skb = NULL;
@@ -1598,6 +1613,9 @@ static int tls_decrypt_sg(struct sock *sk, struct iov_iter *out_iov,
 		return err;
 	}
 
+	if (unlikely(darg->async_done))
+		return 0;
+
 	if (prot->tail_size)
 		darg->tail = dctx->tail;
 
@@ -1769,7 +1787,8 @@ static int process_rx_list(struct tls_sw_context_rx *ctx,
 			   u8 *control,
 			   size_t skip,
 			   size_t len,
-			   bool is_peek)
+			   bool is_peek,
+			   bool *more)
 {
 	struct sk_buff *skb = skb_peek(&ctx->rx_list);
 	struct tls_msg *tlm;
@@ -1782,7 +1801,7 @@ static int process_rx_list(struct tls_sw_context_rx *ctx,
 
 		err = tls_record_content_type(msg, tlm, control);
 		if (err <= 0)
-			goto out;
+			goto more;
 
 		if (skip < rxm->full_len)
 			break;
@@ -1800,12 +1819,12 @@ static int process_rx_list(struct tls_sw_context_rx *ctx,
 
 		err = tls_record_content_type(msg, tlm, control);
 		if (err <= 0)
-			goto out;
+			goto more;
 
 		err = skb_copy_datagram_msg(skb, rxm->offset + skip,
 					    msg, chunk);
 		if (err < 0)
-			goto out;
+			goto more;
 
 		len = len - chunk;
 		copied = copied + chunk;
@@ -1841,6 +1860,10 @@ static int process_rx_list(struct tls_sw_context_rx *ctx,
 
 out:
 	return copied ? : err;
+more:
+	if (more)
+		*more = true;
+	goto out;
 }
 
 static bool
@@ -1940,10 +1963,12 @@ int tls_sw_recvmsg(struct sock *sk,
 	struct strp_msg *rxm;
 	struct tls_msg *tlm;
 	ssize_t copied = 0;
+	ssize_t peeked = 0;
 	bool async = false;
 	int target, err;
 	bool is_kvec = iov_iter_is_kvec(&msg->msg_iter);
 	bool is_peek = flags & MSG_PEEK;
+	bool rx_more = false;
 	bool released = true;
 	bool bpf_strp_enabled;
 	bool zc_capable;
@@ -1963,12 +1988,12 @@ int tls_sw_recvmsg(struct sock *sk,
 		goto end;
 
 	/* Process pending decrypted records. It must be non-zero-copy */
-	err = process_rx_list(ctx, msg, &control, 0, len, is_peek);
+	err = process_rx_list(ctx, msg, &control, 0, len, is_peek, &rx_more);
 	if (err < 0)
 		goto end;
 
 	copied = err;
-	if (len <= copied)
+	if (len <= copied || (copied && control != TLS_RECORD_TYPE_DATA) || rx_more)
 		goto end;
 
 	target = sock_rcvlowat(sk, flags & MSG_WAITALL, len);
@@ -2061,6 +2086,8 @@ put_on_rx_list:
 				decrypted += chunk;
 				len -= chunk;
 				__skb_queue_tail(&ctx->rx_list, skb);
+				if (unlikely(control != TLS_RECORD_TYPE_DATA))
+					break;
 				continue;
 			}
 
@@ -2084,8 +2111,10 @@ put_on_rx_list:
 			if (err < 0)
 				goto put_on_rx_list_err;
 
-			if (is_peek)
+			if (is_peek) {
+				peeked += chunk;
 				goto put_on_rx_list;
+			}
 
 			if (partially_consumed) {
 				rxm->offset += chunk;
@@ -2109,16 +2138,10 @@ put_on_rx_list:
 
 recv_end:
 	if (async) {
-		int ret, pending;
+		int ret;
 
 		/* Wait for all previously submitted records to be decrypted */
-		spin_lock_bh(&ctx->decrypt_compl_lock);
-		reinit_completion(&ctx->async_wait.completion);
-		pending = atomic_read(&ctx->decrypt_pending);
-		spin_unlock_bh(&ctx->decrypt_compl_lock);
-		ret = 0;
-		if (pending)
-			ret = crypto_wait_req(-EINPROGRESS, &ctx->async_wait);
+		ret = tls_decrypt_async_wait(ctx);
 		__skb_queue_purge(&ctx->async_hold);
 
 		if (ret) {
@@ -2130,12 +2153,11 @@ recv_end:
 
 		/* Drain records from the rx_list & copy if required */
 		if (is_peek || is_kvec)
-			err = process_rx_list(ctx, msg, &control, copied,
-					      decrypted, is_peek);
+			err = process_rx_list(ctx, msg, &control, copied + peeked,
+					      decrypted - peeked, is_peek, NULL);
 		else
 			err = process_rx_list(ctx, msg, &control, 0,
-					      async_copy_bytes, is_peek);
-		decrypted += max(err, 0);
+					      async_copy_bytes, is_peek, NULL);
 	}
 
 	copied += decrypted;
@@ -2435,16 +2457,9 @@ 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);
 	struct tls_rec *rec, *tmp;
-	int pending;
 
 	/* Wait for any pending async encryptions to complete */
-	spin_lock_bh(&ctx->encrypt_compl_lock);
-	ctx->async_notify = true;
-	pending = atomic_read(&ctx->encrypt_pending);
-	spin_unlock_bh(&ctx->encrypt_compl_lock);
-
-	if (pending)
-		crypto_wait_req(-EINPROGRESS, &ctx->async_wait);
+	tls_encrypt_async_wait(ctx);
 
 	tls_tx_records(sk, -1);
 
@@ -2607,7 +2622,7 @@ static struct tls_sw_context_tx *init_ctx_tx(struct tls_context *ctx, struct soc
 	}
 
 	crypto_init_wait(&sw_ctx_tx->async_wait);
-	spin_lock_init(&sw_ctx_tx->encrypt_compl_lock);
+	atomic_set(&sw_ctx_tx->encrypt_pending, 1);
 	INIT_LIST_HEAD(&sw_ctx_tx->tx_list);
 	INIT_DELAYED_WORK(&sw_ctx_tx->tx_work.work, tx_work_handler);
 	sw_ctx_tx->tx_work.sk = sk;
@@ -2628,7 +2643,7 @@ static struct tls_sw_context_rx *init_ctx_rx(struct tls_context *ctx)
 	}
 
 	crypto_init_wait(&sw_ctx_rx->async_wait);
-	spin_lock_init(&sw_ctx_rx->decrypt_compl_lock);
+	atomic_set(&sw_ctx_rx->decrypt_pending, 1);
 	init_waitqueue_head(&sw_ctx_rx->wq);
 	skb_queue_head_init(&sw_ctx_rx->rx_list);
 	skb_queue_head_init(&sw_ctx_rx->async_hold);