Adding upstream version 5.10.209.upstream/5.10.209upstream

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

10 files changed, 5807 insertions, 0 deletions

diff --git a/net/tls/Kconfig b/net/tls/Kconfig
new file mode 100644
index 000000000..fa0724fd8
--- /dev/null
+++ b/net/tls/Kconfig
@@ -0,0 +1,38 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# TLS configuration
+#
+config TLS
+	tristate "Transport Layer Security support"
+	depends on INET
+	select CRYPTO
+	select CRYPTO_AES
+	select CRYPTO_GCM
+	select STREAM_PARSER
+	select NET_SOCK_MSG
+	default n
+	help
+	Enable kernel support for TLS protocol. This allows symmetric
+	encryption handling of the TLS protocol to be done in-kernel.
+
+	If unsure, say N.
+
+config TLS_DEVICE
+	bool "Transport Layer Security HW offload"
+	depends on TLS
+	select SOCK_VALIDATE_XMIT
+	default n
+	help
+	Enable kernel support for HW offload of the TLS protocol.
+
+	If unsure, say N.
+
+config TLS_TOE
+	bool "Transport Layer Security TCP stack bypass"
+	depends on TLS
+	default n
+	help
+	Enable kernel support for legacy HW offload of the TLS protocol,
+	which is incompatible with the Linux networking stack semantics.
+
+	If unsure, say N.
diff --git a/net/tls/Makefile b/net/tls/Makefile
new file mode 100644
index 000000000..f1ffbfe89
--- /dev/null
+++ b/net/tls/Makefile
@@ -0,0 +1,13 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# Makefile for the TLS subsystem.
+#
+
+CFLAGS_trace.o := -I$(src)
+
+obj-$(CONFIG_TLS) += tls.o
+
+tls-y := tls_main.o tls_sw.o tls_proc.o trace.o
+
+tls-$(CONFIG_TLS_TOE) += tls_toe.o
+tls-$(CONFIG_TLS_DEVICE) += tls_device.o tls_device_fallback.o
diff --git a/net/tls/tls_device.c b/net/tls/tls_device.c
new file mode 100644
index 000000000..5cb684654
--- /dev/null
+++ b/net/tls/tls_device.c
@@ -0,0 +1,1410 @@
+/* Copyright (c) 2018, Mellanox Technologies All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <crypto/aead.h>
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/netdevice.h>
+#include <net/dst.h>
+#include <net/inet_connection_sock.h>
+#include <net/tcp.h>
+#include <net/tls.h>
+
+#include "trace.h"
+
+/* device_offload_lock is used to synchronize tls_dev_add
+ * against NETDEV_DOWN notifications.
+ */
+static DECLARE_RWSEM(device_offload_lock);
+
+static void tls_device_gc_task(struct work_struct *work);
+
+static DECLARE_WORK(tls_device_gc_work, tls_device_gc_task);
+static LIST_HEAD(tls_device_gc_list);
+static LIST_HEAD(tls_device_list);
+static LIST_HEAD(tls_device_down_list);
+static DEFINE_SPINLOCK(tls_device_lock);
+
+static void tls_device_free_ctx(struct tls_context *ctx)
+{
+	if (ctx->tx_conf == TLS_HW) {
+		kfree(tls_offload_ctx_tx(ctx));
+		kfree(ctx->tx.rec_seq);
+		kfree(ctx->tx.iv);
+	}
+
+	if (ctx->rx_conf == TLS_HW)
+		kfree(tls_offload_ctx_rx(ctx));
+
+	tls_ctx_free(NULL, ctx);
+}
+
+static void tls_device_gc_task(struct work_struct *work)
+{
+	struct tls_context *ctx, *tmp;
+	unsigned long flags;
+	LIST_HEAD(gc_list);
+
+	spin_lock_irqsave(&tls_device_lock, flags);
+	list_splice_init(&tls_device_gc_list, &gc_list);
+	spin_unlock_irqrestore(&tls_device_lock, flags);
+
+	list_for_each_entry_safe(ctx, tmp, &gc_list, list) {
+		struct net_device *netdev = ctx->netdev;
+
+		if (netdev && ctx->tx_conf == TLS_HW) {
+			netdev->tlsdev_ops->tls_dev_del(netdev, ctx,
+							TLS_OFFLOAD_CTX_DIR_TX);
+			dev_put(netdev);
+			ctx->netdev = NULL;
+		}
+
+		list_del(&ctx->list);
+		tls_device_free_ctx(ctx);
+	}
+}
+
+static void tls_device_queue_ctx_destruction(struct tls_context *ctx)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&tls_device_lock, flags);
+	if (unlikely(!refcount_dec_and_test(&ctx->refcount)))
+		goto unlock;
+
+	list_move_tail(&ctx->list, &tls_device_gc_list);
+
+	/* schedule_work inside the spinlock
+	 * to make sure tls_device_down waits for that work.
+	 */
+	schedule_work(&tls_device_gc_work);
+unlock:
+	spin_unlock_irqrestore(&tls_device_lock, flags);
+}
+
+/* We assume that the socket is already connected */
+static struct net_device *get_netdev_for_sock(struct sock *sk)
+{
+	struct dst_entry *dst = sk_dst_get(sk);
+	struct net_device *netdev = NULL;
+
+	if (likely(dst)) {
+		netdev = dst->dev;
+		dev_hold(netdev);
+	}
+
+	dst_release(dst);
+
+	return netdev;
+}
+
+static void destroy_record(struct tls_record_info *record)
+{
+	int i;
+
+	for (i = 0; i < record->num_frags; i++)
+		__skb_frag_unref(&record->frags[i]);
+	kfree(record);
+}
+
+static void delete_all_records(struct tls_offload_context_tx *offload_ctx)
+{
+	struct tls_record_info *info, *temp;
+
+	list_for_each_entry_safe(info, temp, &offload_ctx->records_list, list) {
+		list_del(&info->list);
+		destroy_record(info);
+	}
+
+	offload_ctx->retransmit_hint = NULL;
+}
+
+static void tls_icsk_clean_acked(struct sock *sk, u32 acked_seq)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_record_info *info, *temp;
+	struct tls_offload_context_tx *ctx;
+	u64 deleted_records = 0;
+	unsigned long flags;
+
+	if (!tls_ctx)
+		return;
+
+	ctx = tls_offload_ctx_tx(tls_ctx);
+
+	spin_lock_irqsave(&ctx->lock, flags);
+	info = ctx->retransmit_hint;
+	if (info && !before(acked_seq, info->end_seq))
+		ctx->retransmit_hint = NULL;
+
+	list_for_each_entry_safe(info, temp, &ctx->records_list, list) {
+		if (before(acked_seq, info->end_seq))
+			break;
+		list_del(&info->list);
+
+		destroy_record(info);
+		deleted_records++;
+	}
+
+	ctx->unacked_record_sn += deleted_records;
+	spin_unlock_irqrestore(&ctx->lock, flags);
+}
+
+/* At this point, there should be no references on this
+ * socket and no in-flight SKBs associated with this
+ * socket, so it is safe to free all the resources.
+ */
+void tls_device_sk_destruct(struct sock *sk)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_offload_context_tx *ctx = tls_offload_ctx_tx(tls_ctx);
+
+	tls_ctx->sk_destruct(sk);
+
+	if (tls_ctx->tx_conf == TLS_HW) {
+		if (ctx->open_record)
+			destroy_record(ctx->open_record);
+		delete_all_records(ctx);
+		crypto_free_aead(ctx->aead_send);
+		clean_acked_data_disable(inet_csk(sk));
+	}
+
+	tls_device_queue_ctx_destruction(tls_ctx);
+}
+EXPORT_SYMBOL_GPL(tls_device_sk_destruct);
+
+void tls_device_free_resources_tx(struct sock *sk)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+
+	tls_free_partial_record(sk, tls_ctx);
+}
+
+void tls_offload_tx_resync_request(struct sock *sk, u32 got_seq, u32 exp_seq)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+
+	trace_tls_device_tx_resync_req(sk, got_seq, exp_seq);
+	WARN_ON(test_and_set_bit(TLS_TX_SYNC_SCHED, &tls_ctx->flags));
+}
+EXPORT_SYMBOL_GPL(tls_offload_tx_resync_request);
+
+static void tls_device_resync_tx(struct sock *sk, struct tls_context *tls_ctx,
+				 u32 seq)
+{
+	struct net_device *netdev;
+	struct sk_buff *skb;
+	int err = 0;
+	u8 *rcd_sn;
+
+	skb = tcp_write_queue_tail(sk);
+	if (skb)
+		TCP_SKB_CB(skb)->eor = 1;
+
+	rcd_sn = tls_ctx->tx.rec_seq;
+
+	trace_tls_device_tx_resync_send(sk, seq, rcd_sn);
+	down_read(&device_offload_lock);
+	netdev = tls_ctx->netdev;
+	if (netdev)
+		err = netdev->tlsdev_ops->tls_dev_resync(netdev, sk, seq,
+							 rcd_sn,
+							 TLS_OFFLOAD_CTX_DIR_TX);
+	up_read(&device_offload_lock);
+	if (err)
+		return;
+
+	clear_bit_unlock(TLS_TX_SYNC_SCHED, &tls_ctx->flags);
+}
+
+static void tls_append_frag(struct tls_record_info *record,
+			    struct page_frag *pfrag,
+			    int size)
+{
+	skb_frag_t *frag;
+
+	frag = &record->frags[record->num_frags - 1];
+	if (skb_frag_page(frag) == pfrag->page &&
+	    skb_frag_off(frag) + skb_frag_size(frag) == pfrag->offset) {
+		skb_frag_size_add(frag, size);
+	} else {
+		++frag;
+		__skb_frag_set_page(frag, pfrag->page);
+		skb_frag_off_set(frag, pfrag->offset);
+		skb_frag_size_set(frag, size);
+		++record->num_frags;
+		get_page(pfrag->page);
+	}
+
+	pfrag->offset += size;
+	record->len += size;
+}
+
+static int tls_push_record(struct sock *sk,
+			   struct tls_context *ctx,
+			   struct tls_offload_context_tx *offload_ctx,
+			   struct tls_record_info *record,
+			   int flags)
+{
+	struct tls_prot_info *prot = &ctx->prot_info;
+	struct tcp_sock *tp = tcp_sk(sk);
+	skb_frag_t *frag;
+	int i;
+
+	record->end_seq = tp->write_seq + record->len;
+	list_add_tail_rcu(&record->list, &offload_ctx->records_list);
+	offload_ctx->open_record = NULL;
+
+	if (test_bit(TLS_TX_SYNC_SCHED, &ctx->flags))
+		tls_device_resync_tx(sk, ctx, tp->write_seq);
+
+	tls_advance_record_sn(sk, prot, &ctx->tx);
+
+	for (i = 0; i < record->num_frags; i++) {
+		frag = &record->frags[i];
+		sg_unmark_end(&offload_ctx->sg_tx_data[i]);
+		sg_set_page(&offload_ctx->sg_tx_data[i], skb_frag_page(frag),
+			    skb_frag_size(frag), skb_frag_off(frag));
+		sk_mem_charge(sk, skb_frag_size(frag));
+		get_page(skb_frag_page(frag));
+	}
+	sg_mark_end(&offload_ctx->sg_tx_data[record->num_frags - 1]);
+
+	/* all ready, send */
+	return tls_push_sg(sk, ctx, offload_ctx->sg_tx_data, 0, flags);
+}
+
+static int tls_device_record_close(struct sock *sk,
+				   struct tls_context *ctx,
+				   struct tls_record_info *record,
+				   struct page_frag *pfrag,
+				   unsigned char record_type)
+{
+	struct tls_prot_info *prot = &ctx->prot_info;
+	int ret;
+
+	/* append tag
+	 * device will fill in the tag, we just need to append a placeholder
+	 * use socket memory to improve coalescing (re-using a single buffer
+	 * increases frag count)
+	 * if we can't allocate memory now, steal some back from data
+	 */
+	if (likely(skb_page_frag_refill(prot->tag_size, pfrag,
+					sk->sk_allocation))) {
+		ret = 0;
+		tls_append_frag(record, pfrag, prot->tag_size);
+	} else {
+		ret = prot->tag_size;
+		if (record->len <= prot->overhead_size)
+			return -ENOMEM;
+	}
+
+	/* fill prepend */
+	tls_fill_prepend(ctx, skb_frag_address(&record->frags[0]),
+			 record->len - prot->overhead_size,
+			 record_type, prot->version);
+	return ret;
+}
+
+static int tls_create_new_record(struct tls_offload_context_tx *offload_ctx,
+				 struct page_frag *pfrag,
+				 size_t prepend_size)
+{
+	struct tls_record_info *record;
+	skb_frag_t *frag;
+
+	record = kmalloc(sizeof(*record), GFP_KERNEL);
+	if (!record)
+		return -ENOMEM;
+
+	frag = &record->frags[0];
+	__skb_frag_set_page(frag, pfrag->page);
+	skb_frag_off_set(frag, pfrag->offset);
+	skb_frag_size_set(frag, prepend_size);
+
+	get_page(pfrag->page);
+	pfrag->offset += prepend_size;
+
+	record->num_frags = 1;
+	record->len = prepend_size;
+	offload_ctx->open_record = record;
+	return 0;
+}
+
+static int tls_do_allocation(struct sock *sk,
+			     struct tls_offload_context_tx *offload_ctx,
+			     struct page_frag *pfrag,
+			     size_t prepend_size)
+{
+	int ret;
+
+	if (!offload_ctx->open_record) {
+		if (unlikely(!skb_page_frag_refill(prepend_size, pfrag,
+						   sk->sk_allocation))) {
+			READ_ONCE(sk->sk_prot)->enter_memory_pressure(sk);
+			sk_stream_moderate_sndbuf(sk);
+			return -ENOMEM;
+		}
+
+		ret = tls_create_new_record(offload_ctx, pfrag, prepend_size);
+		if (ret)
+			return ret;
+
+		if (pfrag->size > pfrag->offset)
+			return 0;
+	}
+
+	if (!sk_page_frag_refill(sk, pfrag))
+		return -ENOMEM;
+
+	return 0;
+}
+
+static int tls_device_copy_data(void *addr, size_t bytes, struct iov_iter *i)
+{
+	size_t pre_copy, nocache;
+
+	pre_copy = ~((unsigned long)addr - 1) & (SMP_CACHE_BYTES - 1);
+	if (pre_copy) {
+		pre_copy = min(pre_copy, bytes);
+		if (copy_from_iter(addr, pre_copy, i) != pre_copy)
+			return -EFAULT;
+		bytes -= pre_copy;
+		addr += pre_copy;
+	}
+
+	nocache = round_down(bytes, SMP_CACHE_BYTES);
+	if (copy_from_iter_nocache(addr, nocache, i) != nocache)
+		return -EFAULT;
+	bytes -= nocache;
+	addr += nocache;
+
+	if (bytes && copy_from_iter(addr, bytes, i) != bytes)
+		return -EFAULT;
+
+	return 0;
+}
+
+static int tls_push_data(struct sock *sk,
+			 struct iov_iter *msg_iter,
+			 size_t size, int flags,
+			 unsigned char record_type)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_offload_context_tx *ctx = tls_offload_ctx_tx(tls_ctx);
+	struct tls_record_info *record = ctx->open_record;
+	int tls_push_record_flags;
+	struct page_frag *pfrag;
+	size_t orig_size = size;
+	u32 max_open_record_len;
+	bool more = false;
+	bool done = false;
+	int copy, rc = 0;
+	long timeo;
+
+	if (flags &
+	    ~(MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL | MSG_SENDPAGE_NOTLAST))
+		return -EOPNOTSUPP;
+
+	if (unlikely(sk->sk_err))
+		return -sk->sk_err;
+
+	flags |= MSG_SENDPAGE_DECRYPTED;
+	tls_push_record_flags = flags | MSG_SENDPAGE_NOTLAST;
+
+	timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);
+	if (tls_is_partially_sent_record(tls_ctx)) {
+		rc = tls_push_partial_record(sk, tls_ctx, flags);
+		if (rc < 0)
+			return rc;
+	}
+
+	pfrag = sk_page_frag(sk);
+
+	/* TLS_HEADER_SIZE is not counted as part of the TLS record, and
+	 * we need to leave room for an authentication tag.
+	 */
+	max_open_record_len = TLS_MAX_PAYLOAD_SIZE +
+			      prot->prepend_size;
+	do {
+		rc = tls_do_allocation(sk, ctx, pfrag, prot->prepend_size);
+		if (unlikely(rc)) {
+			rc = sk_stream_wait_memory(sk, &timeo);
+			if (!rc)
+				continue;
+
+			record = ctx->open_record;
+			if (!record)
+				break;
+handle_error:
+			if (record_type != TLS_RECORD_TYPE_DATA) {
+				/* avoid sending partial
+				 * record with type !=
+				 * application_data
+				 */
+				size = orig_size;
+				destroy_record(record);
+				ctx->open_record = NULL;
+			} else if (record->len > prot->prepend_size) {
+				goto last_record;
+			}
+
+			break;
+		}
+
+		record = ctx->open_record;
+		copy = min_t(size_t, size, (pfrag->size - pfrag->offset));
+		copy = min_t(size_t, copy, (max_open_record_len - record->len));
+
+		if (copy) {
+			rc = tls_device_copy_data(page_address(pfrag->page) +
+						  pfrag->offset, copy, msg_iter);
+			if (rc)
+				goto handle_error;
+			tls_append_frag(record, pfrag, copy);
+		}
+
+		size -= copy;
+		if (!size) {
+last_record:
+			tls_push_record_flags = flags;
+			if (flags & (MSG_SENDPAGE_NOTLAST | MSG_MORE)) {
+				more = true;
+				break;
+			}
+
+			done = true;
+		}
+
+		if (done || record->len >= max_open_record_len ||
+		    (record->num_frags >= MAX_SKB_FRAGS - 1)) {
+			rc = tls_device_record_close(sk, tls_ctx, record,
+						     pfrag, record_type);
+			if (rc) {
+				if (rc > 0) {
+					size += rc;
+				} else {
+					size = orig_size;
+					destroy_record(record);
+					ctx->open_record = NULL;
+					break;
+				}
+			}
+
+			rc = tls_push_record(sk,
+					     tls_ctx,
+					     ctx,
+					     record,
+					     tls_push_record_flags);
+			if (rc < 0)
+				break;
+		}
+	} while (!done);
+
+	tls_ctx->pending_open_record_frags = more;
+
+	if (orig_size - size > 0)
+		rc = orig_size - size;
+
+	return rc;
+}
+
+int tls_device_sendmsg(struct sock *sk, struct msghdr *msg, size_t size)
+{
+	unsigned char record_type = TLS_RECORD_TYPE_DATA;
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	int rc;
+
+	mutex_lock(&tls_ctx->tx_lock);
+	lock_sock(sk);
+
+	if (unlikely(msg->msg_controllen)) {
+		rc = tls_proccess_cmsg(sk, msg, &record_type);
+		if (rc)
+			goto out;
+	}
+
+	rc = tls_push_data(sk, &msg->msg_iter, size,
+			   msg->msg_flags, record_type);
+
+out:
+	release_sock(sk);
+	mutex_unlock(&tls_ctx->tx_lock);
+	return rc;
+}
+
+int tls_device_sendpage(struct sock *sk, struct page *page,
+			int offset, size_t size, int flags)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct iov_iter	msg_iter;
+	char *kaddr;
+	struct kvec iov;
+	int rc;
+
+	if (flags & MSG_SENDPAGE_NOTLAST)
+		flags |= MSG_MORE;
+
+	mutex_lock(&tls_ctx->tx_lock);
+	lock_sock(sk);
+
+	if (flags & MSG_OOB) {
+		rc = -EOPNOTSUPP;
+		goto out;
+	}
+
+	kaddr = kmap(page);
+	iov.iov_base = kaddr + offset;
+	iov.iov_len = size;
+	iov_iter_kvec(&msg_iter, WRITE, &iov, 1, size);
+	rc = tls_push_data(sk, &msg_iter, size,
+			   flags, TLS_RECORD_TYPE_DATA);
+	kunmap(page);
+
+out:
+	release_sock(sk);
+	mutex_unlock(&tls_ctx->tx_lock);
+	return rc;
+}
+
+struct tls_record_info *tls_get_record(struct tls_offload_context_tx *context,
+				       u32 seq, u64 *p_record_sn)
+{
+	u64 record_sn = context->hint_record_sn;
+	struct tls_record_info *info, *last;
+
+	info = context->retransmit_hint;
+	if (!info ||
+	    before(seq, info->end_seq - info->len)) {
+		/* if retransmit_hint is irrelevant start
+		 * from the beggining of the list
+		 */
+		info = list_first_entry_or_null(&context->records_list,
+						struct tls_record_info, list);
+		if (!info)
+			return NULL;
+		/* send the start_marker record if seq number is before the
+		 * tls offload start marker sequence number. This record is
+		 * required to handle TCP packets which are before TLS offload
+		 * started.
+		 *  And if it's not start marker, look if this seq number
+		 * belongs to the list.
+		 */
+		if (likely(!tls_record_is_start_marker(info))) {
+			/* we have the first record, get the last record to see
+			 * if this seq number belongs to the list.
+			 */
+			last = list_last_entry(&context->records_list,
+					       struct tls_record_info, list);
+
+			if (!between(seq, tls_record_start_seq(info),
+				     last->end_seq))
+				return NULL;
+		}
+		record_sn = context->unacked_record_sn;
+	}
+
+	/* We just need the _rcu for the READ_ONCE() */
+	rcu_read_lock();
+	list_for_each_entry_from_rcu(info, &context->records_list, list) {
+		if (before(seq, info->end_seq)) {
+			if (!context->retransmit_hint ||
+			    after(info->end_seq,
+				  context->retransmit_hint->end_seq)) {
+				context->hint_record_sn = record_sn;
+				context->retransmit_hint = info;
+			}
+			*p_record_sn = record_sn;
+			goto exit_rcu_unlock;
+		}
+		record_sn++;
+	}
+	info = NULL;
+
+exit_rcu_unlock:
+	rcu_read_unlock();
+	return info;
+}
+EXPORT_SYMBOL(tls_get_record);
+
+static int tls_device_push_pending_record(struct sock *sk, int flags)
+{
+	struct iov_iter	msg_iter;
+
+	iov_iter_kvec(&msg_iter, WRITE, NULL, 0, 0);
+	return tls_push_data(sk, &msg_iter, 0, flags, TLS_RECORD_TYPE_DATA);
+}
+
+void tls_device_write_space(struct sock *sk, struct tls_context *ctx)
+{
+	if (tls_is_partially_sent_record(ctx)) {
+		gfp_t sk_allocation = sk->sk_allocation;
+
+		WARN_ON_ONCE(sk->sk_write_pending);
+
+		sk->sk_allocation = GFP_ATOMIC;
+		tls_push_partial_record(sk, ctx,
+					MSG_DONTWAIT | MSG_NOSIGNAL |
+					MSG_SENDPAGE_DECRYPTED);
+		sk->sk_allocation = sk_allocation;
+	}
+}
+
+static void tls_device_resync_rx(struct tls_context *tls_ctx,
+				 struct sock *sk, u32 seq, u8 *rcd_sn)
+{
+	struct tls_offload_context_rx *rx_ctx = tls_offload_ctx_rx(tls_ctx);
+	struct net_device *netdev;
+
+	trace_tls_device_rx_resync_send(sk, seq, rcd_sn, rx_ctx->resync_type);
+	rcu_read_lock();
+	netdev = READ_ONCE(tls_ctx->netdev);
+	if (netdev)
+		netdev->tlsdev_ops->tls_dev_resync(netdev, sk, seq, rcd_sn,
+						   TLS_OFFLOAD_CTX_DIR_RX);
+	rcu_read_unlock();
+	TLS_INC_STATS(sock_net(sk), LINUX_MIB_TLSRXDEVICERESYNC);
+}
+
+static bool
+tls_device_rx_resync_async(struct tls_offload_resync_async *resync_async,
+			   s64 resync_req, u32 *seq, u16 *rcd_delta)
+{
+	u32 is_async = resync_req & RESYNC_REQ_ASYNC;
+	u32 req_seq = resync_req >> 32;
+	u32 req_end = req_seq + ((resync_req >> 16) & 0xffff);
+	u16 i;
+
+	*rcd_delta = 0;
+
+	if (is_async) {
+		/* shouldn't get to wraparound:
+		 * too long in async stage, something bad happened
+		 */
+		if (WARN_ON_ONCE(resync_async->rcd_delta == USHRT_MAX))
+			return false;
+
+		/* asynchronous stage: log all headers seq such that
+		 * req_seq <= seq <= end_seq, and wait for real resync request
+		 */
+		if (before(*seq, req_seq))
+			return false;
+		if (!after(*seq, req_end) &&
+		    resync_async->loglen < TLS_DEVICE_RESYNC_ASYNC_LOGMAX)
+			resync_async->log[resync_async->loglen++] = *seq;
+
+		resync_async->rcd_delta++;
+
+		return false;
+	}
+
+	/* synchronous stage: check against the logged entries and
+	 * proceed to check the next entries if no match was found
+	 */
+	for (i = 0; i < resync_async->loglen; i++)
+		if (req_seq == resync_async->log[i] &&
+		    atomic64_try_cmpxchg(&resync_async->req, &resync_req, 0)) {
+			*rcd_delta = resync_async->rcd_delta - i;
+			*seq = req_seq;
+			resync_async->loglen = 0;
+			resync_async->rcd_delta = 0;
+			return true;
+		}
+
+	resync_async->loglen = 0;
+	resync_async->rcd_delta = 0;
+
+	if (req_seq == *seq &&
+	    atomic64_try_cmpxchg(&resync_async->req,
+				 &resync_req, 0))
+		return true;
+
+	return false;
+}
+
+void tls_device_rx_resync_new_rec(struct sock *sk, u32 rcd_len, u32 seq)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_offload_context_rx *rx_ctx;
+	u8 rcd_sn[TLS_MAX_REC_SEQ_SIZE];
+	u32 sock_data, is_req_pending;
+	struct tls_prot_info *prot;
+	s64 resync_req;
+	u16 rcd_delta;
+	u32 req_seq;
+
+	if (tls_ctx->rx_conf != TLS_HW)
+		return;
+	if (unlikely(test_bit(TLS_RX_DEV_DEGRADED, &tls_ctx->flags)))
+		return;
+
+	prot = &tls_ctx->prot_info;
+	rx_ctx = tls_offload_ctx_rx(tls_ctx);
+	memcpy(rcd_sn, tls_ctx->rx.rec_seq, prot->rec_seq_size);
+
+	switch (rx_ctx->resync_type) {
+	case TLS_OFFLOAD_SYNC_TYPE_DRIVER_REQ:
+		resync_req = atomic64_read(&rx_ctx->resync_req);
+		req_seq = resync_req >> 32;
+		seq += TLS_HEADER_SIZE - 1;
+		is_req_pending = resync_req;
+
+		if (likely(!is_req_pending) || req_seq != seq ||
+		    !atomic64_try_cmpxchg(&rx_ctx->resync_req, &resync_req, 0))
+			return;
+		break;
+	case TLS_OFFLOAD_SYNC_TYPE_CORE_NEXT_HINT:
+		if (likely(!rx_ctx->resync_nh_do_now))
+			return;
+
+		/* head of next rec is already in, note that the sock_inq will
+		 * include the currently parsed message when called from parser
+		 */
+		sock_data = tcp_inq(sk);
+		if (sock_data > rcd_len) {
+			trace_tls_device_rx_resync_nh_delay(sk, sock_data,
+							    rcd_len);
+			return;
+		}
+
+		rx_ctx->resync_nh_do_now = 0;
+		seq += rcd_len;
+		tls_bigint_increment(rcd_sn, prot->rec_seq_size);
+		break;
+	case TLS_OFFLOAD_SYNC_TYPE_DRIVER_REQ_ASYNC:
+		resync_req = atomic64_read(&rx_ctx->resync_async->req);
+		is_req_pending = resync_req;
+		if (likely(!is_req_pending))
+			return;
+
+		if (!tls_device_rx_resync_async(rx_ctx->resync_async,
+						resync_req, &seq, &rcd_delta))
+			return;
+		tls_bigint_subtract(rcd_sn, rcd_delta);
+		break;
+	}
+
+	tls_device_resync_rx(tls_ctx, sk, seq, rcd_sn);
+}
+
+static void tls_device_core_ctrl_rx_resync(struct tls_context *tls_ctx,
+					   struct tls_offload_context_rx *ctx,
+					   struct sock *sk, struct sk_buff *skb)
+{
+	struct strp_msg *rxm;
+
+	/* device will request resyncs by itself based on stream scan */
+	if (ctx->resync_type != TLS_OFFLOAD_SYNC_TYPE_CORE_NEXT_HINT)
+		return;
+	/* already scheduled */
+	if (ctx->resync_nh_do_now)
+		return;
+	/* seen decrypted fragments since last fully-failed record */
+	if (ctx->resync_nh_reset) {
+		ctx->resync_nh_reset = 0;
+		ctx->resync_nh.decrypted_failed = 1;
+		ctx->resync_nh.decrypted_tgt = TLS_DEVICE_RESYNC_NH_START_IVAL;
+		return;
+	}
+
+	if (++ctx->resync_nh.decrypted_failed <= ctx->resync_nh.decrypted_tgt)
+		return;
+
+	/* doing resync, bump the next target in case it fails */
+	if (ctx->resync_nh.decrypted_tgt < TLS_DEVICE_RESYNC_NH_MAX_IVAL)
+		ctx->resync_nh.decrypted_tgt *= 2;
+	else
+		ctx->resync_nh.decrypted_tgt += TLS_DEVICE_RESYNC_NH_MAX_IVAL;
+
+	rxm = strp_msg(skb);
+
+	/* head of next rec is already in, parser will sync for us */
+	if (tcp_inq(sk) > rxm->full_len) {
+		trace_tls_device_rx_resync_nh_schedule(sk);
+		ctx->resync_nh_do_now = 1;
+	} else {
+		struct tls_prot_info *prot = &tls_ctx->prot_info;
+		u8 rcd_sn[TLS_MAX_REC_SEQ_SIZE];
+
+		memcpy(rcd_sn, tls_ctx->rx.rec_seq, prot->rec_seq_size);
+		tls_bigint_increment(rcd_sn, prot->rec_seq_size);
+
+		tls_device_resync_rx(tls_ctx, sk, tcp_sk(sk)->copied_seq,
+				     rcd_sn);
+	}
+}
+
+static int tls_device_reencrypt(struct sock *sk, struct sk_buff *skb)
+{
+	struct strp_msg *rxm = strp_msg(skb);
+	int err = 0, offset = rxm->offset, copy, nsg, data_len, pos;
+	struct sk_buff *skb_iter, *unused;
+	struct scatterlist sg[1];
+	char *orig_buf, *buf;
+
+	orig_buf = kmalloc(rxm->full_len + TLS_HEADER_SIZE +
+			   TLS_CIPHER_AES_GCM_128_IV_SIZE, sk->sk_allocation);
+	if (!orig_buf)
+		return -ENOMEM;
+	buf = orig_buf;
+
+	nsg = skb_cow_data(skb, 0, &unused);
+	if (unlikely(nsg < 0)) {
+		err = nsg;
+		goto free_buf;
+	}
+
+	sg_init_table(sg, 1);
+	sg_set_buf(&sg[0], buf,
+		   rxm->full_len + TLS_HEADER_SIZE +
+		   TLS_CIPHER_AES_GCM_128_IV_SIZE);
+	err = skb_copy_bits(skb, offset, buf,
+			    TLS_HEADER_SIZE + TLS_CIPHER_AES_GCM_128_IV_SIZE);
+	if (err)
+		goto free_buf;
+
+	/* We are interested only in the decrypted data not the auth */
+	err = decrypt_skb(sk, skb, sg);
+	if (err != -EBADMSG)
+		goto free_buf;
+	else
+		err = 0;
+
+	data_len = rxm->full_len - TLS_CIPHER_AES_GCM_128_TAG_SIZE;
+
+	if (skb_pagelen(skb) > offset) {
+		copy = min_t(int, skb_pagelen(skb) - offset, data_len);
+
+		if (skb->decrypted) {
+			err = skb_store_bits(skb, offset, buf, copy);
+			if (err)
+				goto free_buf;
+		}
+
+		offset += copy;
+		buf += copy;
+	}
+
+	pos = skb_pagelen(skb);
+	skb_walk_frags(skb, skb_iter) {
+		int frag_pos;
+
+		/* Practically all frags must belong to msg if reencrypt
+		 * is needed with current strparser and coalescing logic,
+		 * but strparser may "get optimized", so let's be safe.
+		 */
+		if (pos + skb_iter->len <= offset)
+			goto done_with_frag;
+		if (pos >= data_len + rxm->offset)
+			break;
+
+		frag_pos = offset - pos;
+		copy = min_t(int, skb_iter->len - frag_pos,
+			     data_len + rxm->offset - offset);
+
+		if (skb_iter->decrypted) {
+			err = skb_store_bits(skb_iter, frag_pos, buf, copy);
+			if (err)
+				goto free_buf;
+		}
+
+		offset += copy;
+		buf += copy;
+done_with_frag:
+		pos += skb_iter->len;
+	}
+
+free_buf:
+	kfree(orig_buf);
+	return err;
+}
+
+int tls_device_decrypted(struct sock *sk, struct tls_context *tls_ctx,
+			 struct sk_buff *skb, struct strp_msg *rxm)
+{
+	struct tls_offload_context_rx *ctx = tls_offload_ctx_rx(tls_ctx);
+	int is_decrypted = skb->decrypted;
+	int is_encrypted = !is_decrypted;
+	struct sk_buff *skb_iter;
+
+	/* Check if all the data is decrypted already */
+	skb_walk_frags(skb, skb_iter) {
+		is_decrypted &= skb_iter->decrypted;
+		is_encrypted &= !skb_iter->decrypted;
+	}
+
+	trace_tls_device_decrypted(sk, tcp_sk(sk)->copied_seq - rxm->full_len,
+				   tls_ctx->rx.rec_seq, rxm->full_len,
+				   is_encrypted, is_decrypted);
+
+	ctx->sw.decrypted |= is_decrypted;
+
+	if (unlikely(test_bit(TLS_RX_DEV_DEGRADED, &tls_ctx->flags))) {
+		if (likely(is_encrypted || is_decrypted))
+			return 0;
+
+		/* After tls_device_down disables the offload, the next SKB will
+		 * likely have initial fragments decrypted, and final ones not
+		 * decrypted. We need to reencrypt that single SKB.
+		 */
+		return tls_device_reencrypt(sk, skb);
+	}
+
+	/* Return immediately if the record is either entirely plaintext or
+	 * entirely ciphertext. Otherwise handle reencrypt partially decrypted
+	 * record.
+	 */
+	if (is_decrypted) {
+		ctx->resync_nh_reset = 1;
+		return 0;
+	}
+	if (is_encrypted) {
+		tls_device_core_ctrl_rx_resync(tls_ctx, ctx, sk, skb);
+		return 0;
+	}
+
+	ctx->resync_nh_reset = 1;
+	return tls_device_reencrypt(sk, skb);
+}
+
+static void tls_device_attach(struct tls_context *ctx, struct sock *sk,
+			      struct net_device *netdev)
+{
+	if (sk->sk_destruct != tls_device_sk_destruct) {
+		refcount_set(&ctx->refcount, 1);
+		dev_hold(netdev);
+		ctx->netdev = netdev;
+		spin_lock_irq(&tls_device_lock);
+		list_add_tail(&ctx->list, &tls_device_list);
+		spin_unlock_irq(&tls_device_lock);
+
+		ctx->sk_destruct = sk->sk_destruct;
+		smp_store_release(&sk->sk_destruct, tls_device_sk_destruct);
+	}
+}
+
+int tls_set_device_offload(struct sock *sk, struct tls_context *ctx)
+{
+	u16 nonce_size, tag_size, iv_size, rec_seq_size;
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_record_info *start_marker_record;
+	struct tls_offload_context_tx *offload_ctx;
+	struct tls_crypto_info *crypto_info;
+	struct net_device *netdev;
+	char *iv, *rec_seq;
+	struct sk_buff *skb;
+	__be64 rcd_sn;
+	int rc;
+
+	if (!ctx)
+		return -EINVAL;
+
+	if (ctx->priv_ctx_tx)
+		return -EEXIST;
+
+	start_marker_record = kmalloc(sizeof(*start_marker_record), GFP_KERNEL);
+	if (!start_marker_record)
+		return -ENOMEM;
+
+	offload_ctx = kzalloc(TLS_OFFLOAD_CONTEXT_SIZE_TX, GFP_KERNEL);
+	if (!offload_ctx) {
+		rc = -ENOMEM;
+		goto free_marker_record;
+	}
+
+	crypto_info = &ctx->crypto_send.info;
+	if (crypto_info->version != TLS_1_2_VERSION) {
+		rc = -EOPNOTSUPP;
+		goto free_offload_ctx;
+	}
+
+	switch (crypto_info->cipher_type) {
+	case TLS_CIPHER_AES_GCM_128:
+		nonce_size = TLS_CIPHER_AES_GCM_128_IV_SIZE;
+		tag_size = TLS_CIPHER_AES_GCM_128_TAG_SIZE;
+		iv_size = TLS_CIPHER_AES_GCM_128_IV_SIZE;
+		iv = ((struct tls12_crypto_info_aes_gcm_128 *)crypto_info)->iv;
+		rec_seq_size = TLS_CIPHER_AES_GCM_128_REC_SEQ_SIZE;
+		rec_seq =
+		 ((struct tls12_crypto_info_aes_gcm_128 *)crypto_info)->rec_seq;
+		break;
+	default:
+		rc = -EINVAL;
+		goto free_offload_ctx;
+	}
+
+	/* Sanity-check the rec_seq_size for stack allocations */
+	if (rec_seq_size > TLS_MAX_REC_SEQ_SIZE) {
+		rc = -EINVAL;
+		goto free_offload_ctx;
+	}
+
+	prot->version = crypto_info->version;
+	prot->cipher_type = crypto_info->cipher_type;
+	prot->prepend_size = TLS_HEADER_SIZE + nonce_size;
+	prot->tag_size = tag_size;
+	prot->overhead_size = prot->prepend_size + prot->tag_size;
+	prot->iv_size = iv_size;
+	ctx->tx.iv = kmalloc(iv_size + TLS_CIPHER_AES_GCM_128_SALT_SIZE,
+			     GFP_KERNEL);
+	if (!ctx->tx.iv) {
+		rc = -ENOMEM;
+		goto free_offload_ctx;
+	}
+
+	memcpy(ctx->tx.iv + TLS_CIPHER_AES_GCM_128_SALT_SIZE, iv, iv_size);
+
+	prot->rec_seq_size = rec_seq_size;
+	ctx->tx.rec_seq = kmemdup(rec_seq, rec_seq_size, GFP_KERNEL);
+	if (!ctx->tx.rec_seq) {
+		rc = -ENOMEM;
+		goto free_iv;
+	}
+
+	rc = tls_sw_fallback_init(sk, offload_ctx, crypto_info);
+	if (rc)
+		goto free_rec_seq;
+
+	/* start at rec_seq - 1 to account for the start marker record */
+	memcpy(&rcd_sn, ctx->tx.rec_seq, sizeof(rcd_sn));
+	offload_ctx->unacked_record_sn = be64_to_cpu(rcd_sn) - 1;
+
+	start_marker_record->end_seq = tcp_sk(sk)->write_seq;
+	start_marker_record->len = 0;
+	start_marker_record->num_frags = 0;
+
+	INIT_LIST_HEAD(&offload_ctx->records_list);
+	list_add_tail(&start_marker_record->list, &offload_ctx->records_list);
+	spin_lock_init(&offload_ctx->lock);
+	sg_init_table(offload_ctx->sg_tx_data,
+		      ARRAY_SIZE(offload_ctx->sg_tx_data));
+
+	clean_acked_data_enable(inet_csk(sk), &tls_icsk_clean_acked);
+	ctx->push_pending_record = tls_device_push_pending_record;
+
+	/* TLS offload is greatly simplified if we don't send
+	 * SKBs where only part of the payload needs to be encrypted.
+	 * So mark the last skb in the write queue as end of record.
+	 */
+	skb = tcp_write_queue_tail(sk);
+	if (skb)
+		TCP_SKB_CB(skb)->eor = 1;
+
+	netdev = get_netdev_for_sock(sk);
+	if (!netdev) {
+		pr_err_ratelimited("%s: netdev not found\n", __func__);
+		rc = -EINVAL;
+		goto disable_cad;
+	}
+
+	if (!(netdev->features & NETIF_F_HW_TLS_TX)) {
+		rc = -EOPNOTSUPP;
+		goto release_netdev;
+	}
+
+	/* Avoid offloading if the device is down
+	 * We don't want to offload new flows after
+	 * the NETDEV_DOWN event
+	 *
+	 * device_offload_lock is taken in tls_devices's NETDEV_DOWN
+	 * handler thus protecting from the device going down before
+	 * ctx was added to tls_device_list.
+	 */
+	down_read(&device_offload_lock);
+	if (!(netdev->flags & IFF_UP)) {
+		rc = -EINVAL;
+		goto release_lock;
+	}
+
+	ctx->priv_ctx_tx = offload_ctx;
+	rc = netdev->tlsdev_ops->tls_dev_add(netdev, sk, TLS_OFFLOAD_CTX_DIR_TX,
+					     &ctx->crypto_send.info,
+					     tcp_sk(sk)->write_seq);
+	trace_tls_device_offload_set(sk, TLS_OFFLOAD_CTX_DIR_TX,
+				     tcp_sk(sk)->write_seq, rec_seq, rc);
+	if (rc)
+		goto release_lock;
+
+	tls_device_attach(ctx, sk, netdev);
+	up_read(&device_offload_lock);
+
+	/* following this assignment tls_is_sk_tx_device_offloaded
+	 * will return true and the context might be accessed
+	 * by the netdev's xmit function.
+	 */
+	smp_store_release(&sk->sk_validate_xmit_skb, tls_validate_xmit_skb);
+	dev_put(netdev);
+
+	return 0;
+
+release_lock:
+	up_read(&device_offload_lock);
+release_netdev:
+	dev_put(netdev);
+disable_cad:
+	clean_acked_data_disable(inet_csk(sk));
+	crypto_free_aead(offload_ctx->aead_send);
+free_rec_seq:
+	kfree(ctx->tx.rec_seq);
+free_iv:
+	kfree(ctx->tx.iv);
+free_offload_ctx:
+	kfree(offload_ctx);
+	ctx->priv_ctx_tx = NULL;
+free_marker_record:
+	kfree(start_marker_record);
+	return rc;
+}
+
+int tls_set_device_offload_rx(struct sock *sk, struct tls_context *ctx)
+{
+	struct tls12_crypto_info_aes_gcm_128 *info;
+	struct tls_offload_context_rx *context;
+	struct net_device *netdev;
+	int rc = 0;
+
+	if (ctx->crypto_recv.info.version != TLS_1_2_VERSION)
+		return -EOPNOTSUPP;
+
+	netdev = get_netdev_for_sock(sk);
+	if (!netdev) {
+		pr_err_ratelimited("%s: netdev not found\n", __func__);
+		return -EINVAL;
+	}
+
+	if (!(netdev->features & NETIF_F_HW_TLS_RX)) {
+		rc = -EOPNOTSUPP;
+		goto release_netdev;
+	}
+
+	/* Avoid offloading if the device is down
+	 * We don't want to offload new flows after
+	 * the NETDEV_DOWN event
+	 *
+	 * device_offload_lock is taken in tls_devices's NETDEV_DOWN
+	 * handler thus protecting from the device going down before
+	 * ctx was added to tls_device_list.
+	 */
+	down_read(&device_offload_lock);
+	if (!(netdev->flags & IFF_UP)) {
+		rc = -EINVAL;
+		goto release_lock;
+	}
+
+	context = kzalloc(TLS_OFFLOAD_CONTEXT_SIZE_RX, GFP_KERNEL);
+	if (!context) {
+		rc = -ENOMEM;
+		goto release_lock;
+	}
+	context->resync_nh_reset = 1;
+
+	ctx->priv_ctx_rx = context;
+	rc = tls_set_sw_offload(sk, ctx, 0);
+	if (rc)
+		goto release_ctx;
+
+	rc = netdev->tlsdev_ops->tls_dev_add(netdev, sk, TLS_OFFLOAD_CTX_DIR_RX,
+					     &ctx->crypto_recv.info,
+					     tcp_sk(sk)->copied_seq);
+	info = (void *)&ctx->crypto_recv.info;
+	trace_tls_device_offload_set(sk, TLS_OFFLOAD_CTX_DIR_RX,
+				     tcp_sk(sk)->copied_seq, info->rec_seq, rc);
+	if (rc)
+		goto free_sw_resources;
+
+	tls_device_attach(ctx, sk, netdev);
+	up_read(&device_offload_lock);
+
+	dev_put(netdev);
+
+	return 0;
+
+free_sw_resources:
+	up_read(&device_offload_lock);
+	tls_sw_free_resources_rx(sk);
+	down_read(&device_offload_lock);
+release_ctx:
+	ctx->priv_ctx_rx = NULL;
+release_lock:
+	up_read(&device_offload_lock);
+release_netdev:
+	dev_put(netdev);
+	return rc;
+}
+
+void tls_device_offload_cleanup_rx(struct sock *sk)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct net_device *netdev;
+
+	down_read(&device_offload_lock);
+	netdev = tls_ctx->netdev;
+	if (!netdev)
+		goto out;
+
+	netdev->tlsdev_ops->tls_dev_del(netdev, tls_ctx,
+					TLS_OFFLOAD_CTX_DIR_RX);
+
+	if (tls_ctx->tx_conf != TLS_HW) {
+		dev_put(netdev);
+		tls_ctx->netdev = NULL;
+	} else {
+		set_bit(TLS_RX_DEV_CLOSED, &tls_ctx->flags);
+	}
+out:
+	up_read(&device_offload_lock);
+	tls_sw_release_resources_rx(sk);
+}
+
+static int tls_device_down(struct net_device *netdev)
+{
+	struct tls_context *ctx, *tmp;
+	unsigned long flags;
+	LIST_HEAD(list);
+
+	/* Request a write lock to block new offload attempts */
+	down_write(&device_offload_lock);
+
+	spin_lock_irqsave(&tls_device_lock, flags);
+	list_for_each_entry_safe(ctx, tmp, &tls_device_list, list) {
+		if (ctx->netdev != netdev ||
+		    !refcount_inc_not_zero(&ctx->refcount))
+			continue;
+
+		list_move(&ctx->list, &list);
+	}
+	spin_unlock_irqrestore(&tls_device_lock, flags);
+
+	list_for_each_entry_safe(ctx, tmp, &list, list)	{
+		/* Stop offloaded TX and switch to the fallback.
+		 * tls_is_sk_tx_device_offloaded will return false.
+		 */
+		WRITE_ONCE(ctx->sk->sk_validate_xmit_skb, tls_validate_xmit_skb_sw);
+
+		/* Stop the RX and TX resync.
+		 * tls_dev_resync must not be called after tls_dev_del.
+		 */
+		WRITE_ONCE(ctx->netdev, NULL);
+
+		/* Start skipping the RX resync logic completely. */
+		set_bit(TLS_RX_DEV_DEGRADED, &ctx->flags);
+
+		/* Sync with inflight packets. After this point:
+		 * TX: no non-encrypted packets will be passed to the driver.
+		 * RX: resync requests from the driver will be ignored.
+		 */
+		synchronize_net();
+
+		/* Release the offload context on the driver side. */
+		if (ctx->tx_conf == TLS_HW)
+			netdev->tlsdev_ops->tls_dev_del(netdev, ctx,
+							TLS_OFFLOAD_CTX_DIR_TX);
+		if (ctx->rx_conf == TLS_HW &&
+		    !test_bit(TLS_RX_DEV_CLOSED, &ctx->flags))
+			netdev->tlsdev_ops->tls_dev_del(netdev, ctx,
+							TLS_OFFLOAD_CTX_DIR_RX);
+
+		dev_put(netdev);
+
+		/* Move the context to a separate list for two reasons:
+		 * 1. When the context is deallocated, list_del is called.
+		 * 2. It's no longer an offloaded context, so we don't want to
+		 *    run offload-specific code on this context.
+		 */
+		spin_lock_irqsave(&tls_device_lock, flags);
+		list_move_tail(&ctx->list, &tls_device_down_list);
+		spin_unlock_irqrestore(&tls_device_lock, flags);
+
+		/* Device contexts for RX and TX will be freed in on sk_destruct
+		 * by tls_device_free_ctx. rx_conf and tx_conf stay in TLS_HW.
+		 * Now release the ref taken above.
+		 */
+		if (refcount_dec_and_test(&ctx->refcount)) {
+			/* sk_destruct ran after tls_device_down took a ref, and
+			 * it returned early. Complete the destruction here.
+			 */
+			list_del(&ctx->list);
+			tls_device_free_ctx(ctx);
+		}
+	}
+
+	up_write(&device_offload_lock);
+
+	flush_work(&tls_device_gc_work);
+
+	return NOTIFY_DONE;
+}
+
+static int tls_dev_event(struct notifier_block *this, unsigned long event,
+			 void *ptr)
+{
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
+
+	if (!dev->tlsdev_ops &&
+	    !(dev->features & (NETIF_F_HW_TLS_RX | NETIF_F_HW_TLS_TX)))
+		return NOTIFY_DONE;
+
+	switch (event) {
+	case NETDEV_REGISTER:
+	case NETDEV_FEAT_CHANGE:
+		if ((dev->features & NETIF_F_HW_TLS_RX) &&
+		    !dev->tlsdev_ops->tls_dev_resync)
+			return NOTIFY_BAD;
+
+		if  (dev->tlsdev_ops &&
+		     dev->tlsdev_ops->tls_dev_add &&
+		     dev->tlsdev_ops->tls_dev_del)
+			return NOTIFY_DONE;
+		else
+			return NOTIFY_BAD;
+	case NETDEV_DOWN:
+		return tls_device_down(dev);
+	}
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block tls_dev_notifier = {
+	.notifier_call	= tls_dev_event,
+};
+
+int __init tls_device_init(void)
+{
+	return register_netdevice_notifier(&tls_dev_notifier);
+}
+
+void __exit tls_device_cleanup(void)
+{
+	unregister_netdevice_notifier(&tls_dev_notifier);
+	flush_work(&tls_device_gc_work);
+	clean_acked_data_flush();
+}
diff --git a/net/tls/tls_device_fallback.c b/net/tls/tls_device_fallback.c
new file mode 100644
index 000000000..0d40016bf
--- /dev/null
+++ b/net/tls/tls_device_fallback.c
@@ -0,0 +1,479 @@
+/* Copyright (c) 2018, Mellanox Technologies All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <net/tls.h>
+#include <crypto/aead.h>
+#include <crypto/scatterwalk.h>
+#include <net/ip6_checksum.h>
+
+static void chain_to_walk(struct scatterlist *sg, struct scatter_walk *walk)
+{
+	struct scatterlist *src = walk->sg;
+	int diff = walk->offset - src->offset;
+
+	sg_set_page(sg, sg_page(src),
+		    src->length - diff, walk->offset);
+
+	scatterwalk_crypto_chain(sg, sg_next(src), 2);
+}
+
+static int tls_enc_record(struct aead_request *aead_req,
+			  struct crypto_aead *aead, char *aad,
+			  char *iv, __be64 rcd_sn,
+			  struct scatter_walk *in,
+			  struct scatter_walk *out, int *in_len)
+{
+	unsigned char buf[TLS_HEADER_SIZE + TLS_CIPHER_AES_GCM_128_IV_SIZE];
+	struct scatterlist sg_in[3];
+	struct scatterlist sg_out[3];
+	u16 len;
+	int rc;
+
+	len = min_t(int, *in_len, ARRAY_SIZE(buf));
+
+	scatterwalk_copychunks(buf, in, len, 0);
+	scatterwalk_copychunks(buf, out, len, 1);
+
+	*in_len -= len;
+	if (!*in_len)
+		return 0;
+
+	scatterwalk_pagedone(in, 0, 1);
+	scatterwalk_pagedone(out, 1, 1);
+
+	len = buf[4] | (buf[3] << 8);
+	len -= TLS_CIPHER_AES_GCM_128_IV_SIZE;
+
+	tls_make_aad(aad, len - TLS_CIPHER_AES_GCM_128_TAG_SIZE,
+		(char *)&rcd_sn, sizeof(rcd_sn), buf[0],
+		TLS_1_2_VERSION);
+
+	memcpy(iv + TLS_CIPHER_AES_GCM_128_SALT_SIZE, buf + TLS_HEADER_SIZE,
+	       TLS_CIPHER_AES_GCM_128_IV_SIZE);
+
+	sg_init_table(sg_in, ARRAY_SIZE(sg_in));
+	sg_init_table(sg_out, ARRAY_SIZE(sg_out));
+	sg_set_buf(sg_in, aad, TLS_AAD_SPACE_SIZE);
+	sg_set_buf(sg_out, aad, TLS_AAD_SPACE_SIZE);
+	chain_to_walk(sg_in + 1, in);
+	chain_to_walk(sg_out + 1, out);
+
+	*in_len -= len;
+	if (*in_len < 0) {
+		*in_len += TLS_CIPHER_AES_GCM_128_TAG_SIZE;
+		/* the input buffer doesn't contain the entire record.
+		 * trim len accordingly. The resulting authentication tag
+		 * will contain garbage, but we don't care, so we won't
+		 * include any of it in the output skb
+		 * Note that we assume the output buffer length
+		 * is larger then input buffer length + tag size
+		 */
+		if (*in_len < 0)
+			len += *in_len;
+
+		*in_len = 0;
+	}
+
+	if (*in_len) {
+		scatterwalk_copychunks(NULL, in, len, 2);
+		scatterwalk_pagedone(in, 0, 1);
+		scatterwalk_copychunks(NULL, out, len, 2);
+		scatterwalk_pagedone(out, 1, 1);
+	}
+
+	len -= TLS_CIPHER_AES_GCM_128_TAG_SIZE;
+	aead_request_set_crypt(aead_req, sg_in, sg_out, len, iv);
+
+	rc = crypto_aead_encrypt(aead_req);
+
+	return rc;
+}
+
+static void tls_init_aead_request(struct aead_request *aead_req,
+				  struct crypto_aead *aead)
+{
+	aead_request_set_tfm(aead_req, aead);
+	aead_request_set_ad(aead_req, TLS_AAD_SPACE_SIZE);
+}
+
+static struct aead_request *tls_alloc_aead_request(struct crypto_aead *aead,
+						   gfp_t flags)
+{
+	unsigned int req_size = sizeof(struct aead_request) +
+		crypto_aead_reqsize(aead);
+	struct aead_request *aead_req;
+
+	aead_req = kzalloc(req_size, flags);
+	if (aead_req)
+		tls_init_aead_request(aead_req, aead);
+	return aead_req;
+}
+
+static int tls_enc_records(struct aead_request *aead_req,
+			   struct crypto_aead *aead, struct scatterlist *sg_in,
+			   struct scatterlist *sg_out, char *aad, char *iv,
+			   u64 rcd_sn, int len)
+{
+	struct scatter_walk out, in;
+	int rc;
+
+	scatterwalk_start(&in, sg_in);
+	scatterwalk_start(&out, sg_out);
+
+	do {
+		rc = tls_enc_record(aead_req, aead, aad, iv,
+				    cpu_to_be64(rcd_sn), &in, &out, &len);
+		rcd_sn++;
+
+	} while (rc == 0 && len);
+
+	scatterwalk_done(&in, 0, 0);
+	scatterwalk_done(&out, 1, 0);
+
+	return rc;
+}
+
+/* Can't use icsk->icsk_af_ops->send_check here because the ip addresses
+ * might have been changed by NAT.
+ */
+static void update_chksum(struct sk_buff *skb, int headln)
+{
+	struct tcphdr *th = tcp_hdr(skb);
+	int datalen = skb->len - headln;
+	const struct ipv6hdr *ipv6h;
+	const struct iphdr *iph;
+
+	/* We only changed the payload so if we are using partial we don't
+	 * need to update anything.
+	 */
+	if (likely(skb->ip_summed == CHECKSUM_PARTIAL))
+		return;
+
+	skb->ip_summed = CHECKSUM_PARTIAL;
+	skb->csum_start = skb_transport_header(skb) - skb->head;
+	skb->csum_offset = offsetof(struct tcphdr, check);
+
+	if (skb->sk->sk_family == AF_INET6) {
+		ipv6h = ipv6_hdr(skb);
+		th->check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr,
+					     datalen, IPPROTO_TCP, 0);
+	} else {
+		iph = ip_hdr(skb);
+		th->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr, datalen,
+					       IPPROTO_TCP, 0);
+	}
+}
+
+static void complete_skb(struct sk_buff *nskb, struct sk_buff *skb, int headln)
+{
+	struct sock *sk = skb->sk;
+	int delta;
+
+	skb_copy_header(nskb, skb);
+
+	skb_put(nskb, skb->len);
+	memcpy(nskb->data, skb->data, headln);
+
+	nskb->destructor = skb->destructor;
+	nskb->sk = sk;
+	skb->destructor = NULL;
+	skb->sk = NULL;
+
+	update_chksum(nskb, headln);
+
+	/* sock_efree means skb must gone through skb_orphan_partial() */
+	if (nskb->destructor == sock_efree)
+		return;
+
+	delta = nskb->truesize - skb->truesize;
+	if (likely(delta < 0))
+		WARN_ON_ONCE(refcount_sub_and_test(-delta, &sk->sk_wmem_alloc));
+	else if (delta)
+		refcount_add(delta, &sk->sk_wmem_alloc);
+}
+
+/* This function may be called after the user socket is already
+ * closed so make sure we don't use anything freed during
+ * tls_sk_proto_close here
+ */
+
+static int fill_sg_in(struct scatterlist *sg_in,
+		      struct sk_buff *skb,
+		      struct tls_offload_context_tx *ctx,
+		      u64 *rcd_sn,
+		      s32 *sync_size,
+		      int *resync_sgs)
+{
+	int tcp_payload_offset = skb_transport_offset(skb) + tcp_hdrlen(skb);
+	int payload_len = skb->len - tcp_payload_offset;
+	u32 tcp_seq = ntohl(tcp_hdr(skb)->seq);
+	struct tls_record_info *record;
+	unsigned long flags;
+	int remaining;
+	int i;
+
+	spin_lock_irqsave(&ctx->lock, flags);
+	record = tls_get_record(ctx, tcp_seq, rcd_sn);
+	if (!record) {
+		spin_unlock_irqrestore(&ctx->lock, flags);
+		return -EINVAL;
+	}
+
+	*sync_size = tcp_seq - tls_record_start_seq(record);
+	if (*sync_size < 0) {
+		int is_start_marker = tls_record_is_start_marker(record);
+
+		spin_unlock_irqrestore(&ctx->lock, flags);
+		/* This should only occur if the relevant record was
+		 * already acked. In that case it should be ok
+		 * to drop the packet and avoid retransmission.
+		 *
+		 * There is a corner case where the packet contains
+		 * both an acked and a non-acked record.
+		 * We currently don't handle that case and rely
+		 * on TCP to retranmit a packet that doesn't contain
+		 * already acked payload.
+		 */
+		if (!is_start_marker)
+			*sync_size = 0;
+		return -EINVAL;
+	}
+
+	remaining = *sync_size;
+	for (i = 0; remaining > 0; i++) {
+		skb_frag_t *frag = &record->frags[i];
+
+		__skb_frag_ref(frag);
+		sg_set_page(sg_in + i, skb_frag_page(frag),
+			    skb_frag_size(frag), skb_frag_off(frag));
+
+		remaining -= skb_frag_size(frag);
+
+		if (remaining < 0)
+			sg_in[i].length += remaining;
+	}
+	*resync_sgs = i;
+
+	spin_unlock_irqrestore(&ctx->lock, flags);
+	if (skb_to_sgvec(skb, &sg_in[i], tcp_payload_offset, payload_len) < 0)
+		return -EINVAL;
+
+	return 0;
+}
+
+static void fill_sg_out(struct scatterlist sg_out[3], void *buf,
+			struct tls_context *tls_ctx,
+			struct sk_buff *nskb,
+			int tcp_payload_offset,
+			int payload_len,
+			int sync_size,
+			void *dummy_buf)
+{
+	sg_set_buf(&sg_out[0], dummy_buf, sync_size);
+	sg_set_buf(&sg_out[1], nskb->data + tcp_payload_offset, payload_len);
+	/* Add room for authentication tag produced by crypto */
+	dummy_buf += sync_size;
+	sg_set_buf(&sg_out[2], dummy_buf, TLS_CIPHER_AES_GCM_128_TAG_SIZE);
+}
+
+static struct sk_buff *tls_enc_skb(struct tls_context *tls_ctx,
+				   struct scatterlist sg_out[3],
+				   struct scatterlist *sg_in,
+				   struct sk_buff *skb,
+				   s32 sync_size, u64 rcd_sn)
+{
+	int tcp_payload_offset = skb_transport_offset(skb) + tcp_hdrlen(skb);
+	struct tls_offload_context_tx *ctx = tls_offload_ctx_tx(tls_ctx);
+	int payload_len = skb->len - tcp_payload_offset;
+	void *buf, *iv, *aad, *dummy_buf;
+	struct aead_request *aead_req;
+	struct sk_buff *nskb = NULL;
+	int buf_len;
+
+	aead_req = tls_alloc_aead_request(ctx->aead_send, GFP_ATOMIC);
+	if (!aead_req)
+		return NULL;
+
+	buf_len = TLS_CIPHER_AES_GCM_128_SALT_SIZE +
+		  TLS_CIPHER_AES_GCM_128_IV_SIZE +
+		  TLS_AAD_SPACE_SIZE +
+		  sync_size +
+		  TLS_CIPHER_AES_GCM_128_TAG_SIZE;
+	buf = kmalloc(buf_len, GFP_ATOMIC);
+	if (!buf)
+		goto free_req;
+
+	iv = buf;
+	memcpy(iv, tls_ctx->crypto_send.aes_gcm_128.salt,
+	       TLS_CIPHER_AES_GCM_128_SALT_SIZE);
+	aad = buf + TLS_CIPHER_AES_GCM_128_SALT_SIZE +
+	      TLS_CIPHER_AES_GCM_128_IV_SIZE;
+	dummy_buf = aad + TLS_AAD_SPACE_SIZE;
+
+	nskb = alloc_skb(skb_headroom(skb) + skb->len, GFP_ATOMIC);
+	if (!nskb)
+		goto free_buf;
+
+	skb_reserve(nskb, skb_headroom(skb));
+
+	fill_sg_out(sg_out, buf, tls_ctx, nskb, tcp_payload_offset,
+		    payload_len, sync_size, dummy_buf);
+
+	if (tls_enc_records(aead_req, ctx->aead_send, sg_in, sg_out, aad, iv,
+			    rcd_sn, sync_size + payload_len) < 0)
+		goto free_nskb;
+
+	complete_skb(nskb, skb, tcp_payload_offset);
+
+	/* validate_xmit_skb_list assumes that if the skb wasn't segmented
+	 * nskb->prev will point to the skb itself
+	 */
+	nskb->prev = nskb;
+
+free_buf:
+	kfree(buf);
+free_req:
+	kfree(aead_req);
+	return nskb;
+free_nskb:
+	kfree_skb(nskb);
+	nskb = NULL;
+	goto free_buf;
+}
+
+static struct sk_buff *tls_sw_fallback(struct sock *sk, struct sk_buff *skb)
+{
+	int tcp_payload_offset = skb_transport_offset(skb) + tcp_hdrlen(skb);
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_offload_context_tx *ctx = tls_offload_ctx_tx(tls_ctx);
+	int payload_len = skb->len - tcp_payload_offset;
+	struct scatterlist *sg_in, sg_out[3];
+	struct sk_buff *nskb = NULL;
+	int sg_in_max_elements;
+	int resync_sgs = 0;
+	s32 sync_size = 0;
+	u64 rcd_sn;
+
+	/* worst case is:
+	 * MAX_SKB_FRAGS in tls_record_info
+	 * MAX_SKB_FRAGS + 1 in SKB head and frags.
+	 */
+	sg_in_max_elements = 2 * MAX_SKB_FRAGS + 1;
+
+	if (!payload_len)
+		return skb;
+
+	sg_in = kmalloc_array(sg_in_max_elements, sizeof(*sg_in), GFP_ATOMIC);
+	if (!sg_in)
+		goto free_orig;
+
+	sg_init_table(sg_in, sg_in_max_elements);
+	sg_init_table(sg_out, ARRAY_SIZE(sg_out));
+
+	if (fill_sg_in(sg_in, skb, ctx, &rcd_sn, &sync_size, &resync_sgs)) {
+		/* bypass packets before kernel TLS socket option was set */
+		if (sync_size < 0 && payload_len <= -sync_size)
+			nskb = skb_get(skb);
+		goto put_sg;
+	}
+
+	nskb = tls_enc_skb(tls_ctx, sg_out, sg_in, skb, sync_size, rcd_sn);
+
+put_sg:
+	while (resync_sgs)
+		put_page(sg_page(&sg_in[--resync_sgs]));
+	kfree(sg_in);
+free_orig:
+	if (nskb)
+		consume_skb(skb);
+	else
+		kfree_skb(skb);
+	return nskb;
+}
+
+struct sk_buff *tls_validate_xmit_skb(struct sock *sk,
+				      struct net_device *dev,
+				      struct sk_buff *skb)
+{
+	if (dev == tls_get_ctx(sk)->netdev)
+		return skb;
+
+	return tls_sw_fallback(sk, skb);
+}
+EXPORT_SYMBOL_GPL(tls_validate_xmit_skb);
+
+struct sk_buff *tls_validate_xmit_skb_sw(struct sock *sk,
+					 struct net_device *dev,
+					 struct sk_buff *skb)
+{
+	return tls_sw_fallback(sk, skb);
+}
+
+struct sk_buff *tls_encrypt_skb(struct sk_buff *skb)
+{
+	return tls_sw_fallback(skb->sk, skb);
+}
+EXPORT_SYMBOL_GPL(tls_encrypt_skb);
+
+int tls_sw_fallback_init(struct sock *sk,
+			 struct tls_offload_context_tx *offload_ctx,
+			 struct tls_crypto_info *crypto_info)
+{
+	const u8 *key;
+	int rc;
+
+	offload_ctx->aead_send =
+	    crypto_alloc_aead("gcm(aes)", 0, CRYPTO_ALG_ASYNC);
+	if (IS_ERR(offload_ctx->aead_send)) {
+		rc = PTR_ERR(offload_ctx->aead_send);
+		pr_err_ratelimited("crypto_alloc_aead failed rc=%d\n", rc);
+		offload_ctx->aead_send = NULL;
+		goto err_out;
+	}
+
+	key = ((struct tls12_crypto_info_aes_gcm_128 *)crypto_info)->key;
+
+	rc = crypto_aead_setkey(offload_ctx->aead_send, key,
+				TLS_CIPHER_AES_GCM_128_KEY_SIZE);
+	if (rc)
+		goto free_aead;
+
+	rc = crypto_aead_setauthsize(offload_ctx->aead_send,
+				     TLS_CIPHER_AES_GCM_128_TAG_SIZE);
+	if (rc)
+		goto free_aead;
+
+	return 0;
+free_aead:
+	crypto_free_aead(offload_ctx->aead_send);
+err_out:
+	return rc;
+}
diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c
new file mode 100644
index 000000000..7ee3c8b03
--- /dev/null
+++ b/net/tls/tls_main.c
@@ -0,0 +1,929 @@
+/*
+ * Copyright (c) 2016-2017, Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2016-2017, Dave Watson <davejwatson@fb.com>. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/module.h>
+
+#include <net/tcp.h>
+#include <net/inet_common.h>
+#include <linux/highmem.h>
+#include <linux/netdevice.h>
+#include <linux/sched/signal.h>
+#include <linux/inetdevice.h>
+#include <linux/inet_diag.h>
+
+#include <net/snmp.h>
+#include <net/tls.h>
+#include <net/tls_toe.h>
+
+MODULE_AUTHOR("Mellanox Technologies");
+MODULE_DESCRIPTION("Transport Layer Security Support");
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_ALIAS_TCP_ULP("tls");
+
+enum {
+	TLSV4,
+	TLSV6,
+	TLS_NUM_PROTS,
+};
+
+static const struct proto *saved_tcpv6_prot;
+static DEFINE_MUTEX(tcpv6_prot_mutex);
+static const struct proto *saved_tcpv4_prot;
+static DEFINE_MUTEX(tcpv4_prot_mutex);
+static struct proto tls_prots[TLS_NUM_PROTS][TLS_NUM_CONFIG][TLS_NUM_CONFIG];
+static struct proto_ops tls_proto_ops[TLS_NUM_PROTS][TLS_NUM_CONFIG][TLS_NUM_CONFIG];
+static void build_protos(struct proto prot[TLS_NUM_CONFIG][TLS_NUM_CONFIG],
+			 const struct proto *base);
+
+void update_sk_prot(struct sock *sk, struct tls_context *ctx)
+{
+	int ip_ver = sk->sk_family == AF_INET6 ? TLSV6 : TLSV4;
+
+	WRITE_ONCE(sk->sk_prot,
+		   &tls_prots[ip_ver][ctx->tx_conf][ctx->rx_conf]);
+	WRITE_ONCE(sk->sk_socket->ops,
+		   &tls_proto_ops[ip_ver][ctx->tx_conf][ctx->rx_conf]);
+}
+
+int wait_on_pending_writer(struct sock *sk, long *timeo)
+{
+	int rc = 0;
+	DEFINE_WAIT_FUNC(wait, woken_wake_function);
+
+	add_wait_queue(sk_sleep(sk), &wait);
+	while (1) {
+		if (!*timeo) {
+			rc = -EAGAIN;
+			break;
+		}
+
+		if (signal_pending(current)) {
+			rc = sock_intr_errno(*timeo);
+			break;
+		}
+
+		if (sk_wait_event(sk, timeo,
+				  !READ_ONCE(sk->sk_write_pending), &wait))
+			break;
+	}
+	remove_wait_queue(sk_sleep(sk), &wait);
+	return rc;
+}
+
+int tls_push_sg(struct sock *sk,
+		struct tls_context *ctx,
+		struct scatterlist *sg,
+		u16 first_offset,
+		int flags)
+{
+	int sendpage_flags = flags | MSG_SENDPAGE_NOTLAST;
+	int ret = 0;
+	struct page *p;
+	size_t size;
+	int offset = first_offset;
+
+	size = sg->length - offset;
+	offset += sg->offset;
+
+	ctx->in_tcp_sendpages = true;
+	while (1) {
+		if (sg_is_last(sg))
+			sendpage_flags = flags;
+
+		/* is sending application-limited? */
+		tcp_rate_check_app_limited(sk);
+		p = sg_page(sg);
+retry:
+		ret = do_tcp_sendpages(sk, p, offset, size, sendpage_flags);
+
+		if (ret != size) {
+			if (ret > 0) {
+				offset += ret;
+				size -= ret;
+				goto retry;
+			}
+
+			offset -= sg->offset;
+			ctx->partially_sent_offset = offset;
+			ctx->partially_sent_record = (void *)sg;
+			ctx->in_tcp_sendpages = false;
+			return ret;
+		}
+
+		put_page(p);
+		sk_mem_uncharge(sk, sg->length);
+		sg = sg_next(sg);
+		if (!sg)
+			break;
+
+		offset = sg->offset;
+		size = sg->length;
+	}
+
+	ctx->in_tcp_sendpages = false;
+
+	return 0;
+}
+
+static int tls_handle_open_record(struct sock *sk, int flags)
+{
+	struct tls_context *ctx = tls_get_ctx(sk);
+
+	if (tls_is_pending_open_record(ctx))
+		return ctx->push_pending_record(sk, flags);
+
+	return 0;
+}
+
+int tls_proccess_cmsg(struct sock *sk, struct msghdr *msg,
+		      unsigned char *record_type)
+{
+	struct cmsghdr *cmsg;
+	int rc = -EINVAL;
+
+	for_each_cmsghdr(cmsg, msg) {
+		if (!CMSG_OK(msg, cmsg))
+			return -EINVAL;
+		if (cmsg->cmsg_level != SOL_TLS)
+			continue;
+
+		switch (cmsg->cmsg_type) {
+		case TLS_SET_RECORD_TYPE:
+			if (cmsg->cmsg_len < CMSG_LEN(sizeof(*record_type)))
+				return -EINVAL;
+
+			if (msg->msg_flags & MSG_MORE)
+				return -EINVAL;
+
+			rc = tls_handle_open_record(sk, msg->msg_flags);
+			if (rc)
+				return rc;
+
+			*record_type = *(unsigned char *)CMSG_DATA(cmsg);
+			rc = 0;
+			break;
+		default:
+			return -EINVAL;
+		}
+	}
+
+	return rc;
+}
+
+int tls_push_partial_record(struct sock *sk, struct tls_context *ctx,
+			    int flags)
+{
+	struct scatterlist *sg;
+	u16 offset;
+
+	sg = ctx->partially_sent_record;
+	offset = ctx->partially_sent_offset;
+
+	ctx->partially_sent_record = NULL;
+	return tls_push_sg(sk, ctx, sg, offset, flags);
+}
+
+void tls_free_partial_record(struct sock *sk, struct tls_context *ctx)
+{
+	struct scatterlist *sg;
+
+	for (sg = ctx->partially_sent_record; sg; sg = sg_next(sg)) {
+		put_page(sg_page(sg));
+		sk_mem_uncharge(sk, sg->length);
+	}
+	ctx->partially_sent_record = NULL;
+}
+
+static void tls_write_space(struct sock *sk)
+{
+	struct tls_context *ctx = tls_get_ctx(sk);
+
+	/* If in_tcp_sendpages call lower protocol write space handler
+	 * to ensure we wake up any waiting operations there. For example
+	 * if do_tcp_sendpages where to call sk_wait_event.
+	 */
+	if (ctx->in_tcp_sendpages) {
+		ctx->sk_write_space(sk);
+		return;
+	}
+
+#ifdef CONFIG_TLS_DEVICE
+	if (ctx->tx_conf == TLS_HW)
+		tls_device_write_space(sk, ctx);
+	else
+#endif
+		tls_sw_write_space(sk, ctx);
+
+	ctx->sk_write_space(sk);
+}
+
+/**
+ * tls_ctx_free() - free TLS ULP context
+ * @sk:  socket to with @ctx is attached
+ * @ctx: TLS context structure
+ *
+ * Free TLS context. If @sk is %NULL caller guarantees that the socket
+ * to which @ctx was attached has no outstanding references.
+ */
+void tls_ctx_free(struct sock *sk, struct tls_context *ctx)
+{
+	if (!ctx)
+		return;
+
+	memzero_explicit(&ctx->crypto_send, sizeof(ctx->crypto_send));
+	memzero_explicit(&ctx->crypto_recv, sizeof(ctx->crypto_recv));
+	mutex_destroy(&ctx->tx_lock);
+
+	if (sk)
+		kfree_rcu(ctx, rcu);
+	else
+		kfree(ctx);
+}
+
+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);
+
+	/* We need these for tls_sw_fallback handling of other packets */
+	if (ctx->tx_conf == TLS_SW) {
+		kfree(ctx->tx.rec_seq);
+		kfree(ctx->tx.iv);
+		tls_sw_release_resources_tx(sk);
+		TLS_DEC_STATS(sock_net(sk), LINUX_MIB_TLSCURRTXSW);
+	} else if (ctx->tx_conf == TLS_HW) {
+		tls_device_free_resources_tx(sk);
+		TLS_DEC_STATS(sock_net(sk), LINUX_MIB_TLSCURRTXDEVICE);
+	}
+
+	if (ctx->rx_conf == TLS_SW) {
+		tls_sw_release_resources_rx(sk);
+		TLS_DEC_STATS(sock_net(sk), LINUX_MIB_TLSCURRRXSW);
+	} else if (ctx->rx_conf == TLS_HW) {
+		tls_device_offload_cleanup_rx(sk);
+		TLS_DEC_STATS(sock_net(sk), LINUX_MIB_TLSCURRRXDEVICE);
+	}
+}
+
+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)
+		rcu_assign_pointer(icsk->icsk_ulp_data, NULL);
+	WRITE_ONCE(sk->sk_prot, ctx->sk_proto);
+	if (sk->sk_write_space == tls_write_space)
+		sk->sk_write_space = ctx->sk_write_space;
+	write_unlock_bh(&sk->sk_callback_lock);
+	release_sock(sk);
+	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(sk, ctx);
+}
+
+static int do_tls_getsockopt_conf(struct sock *sk, char __user *optval,
+				  int __user *optlen, int tx)
+{
+	int rc = 0;
+	struct tls_context *ctx = tls_get_ctx(sk);
+	struct tls_crypto_info *crypto_info;
+	struct cipher_context *cctx;
+	int len;
+
+	if (get_user(len, optlen))
+		return -EFAULT;
+
+	if (!optval || (len < sizeof(*crypto_info))) {
+		rc = -EINVAL;
+		goto out;
+	}
+
+	if (!ctx) {
+		rc = -EBUSY;
+		goto out;
+	}
+
+	/* get user crypto info */
+	if (tx) {
+		crypto_info = &ctx->crypto_send.info;
+		cctx = &ctx->tx;
+	} else {
+		crypto_info = &ctx->crypto_recv.info;
+		cctx = &ctx->rx;
+	}
+
+	if (!TLS_CRYPTO_INFO_READY(crypto_info)) {
+		rc = -EBUSY;
+		goto out;
+	}
+
+	if (len == sizeof(*crypto_info)) {
+		if (copy_to_user(optval, crypto_info, sizeof(*crypto_info)))
+			rc = -EFAULT;
+		goto out;
+	}
+
+	switch (crypto_info->cipher_type) {
+	case TLS_CIPHER_AES_GCM_128: {
+		struct tls12_crypto_info_aes_gcm_128 *
+		  crypto_info_aes_gcm_128 =
+		  container_of(crypto_info,
+			       struct tls12_crypto_info_aes_gcm_128,
+			       info);
+
+		if (len != sizeof(*crypto_info_aes_gcm_128)) {
+			rc = -EINVAL;
+			goto out;
+		}
+		memcpy(crypto_info_aes_gcm_128->iv,
+		       cctx->iv + TLS_CIPHER_AES_GCM_128_SALT_SIZE,
+		       TLS_CIPHER_AES_GCM_128_IV_SIZE);
+		memcpy(crypto_info_aes_gcm_128->rec_seq, cctx->rec_seq,
+		       TLS_CIPHER_AES_GCM_128_REC_SEQ_SIZE);
+		if (copy_to_user(optval,
+				 crypto_info_aes_gcm_128,
+				 sizeof(*crypto_info_aes_gcm_128)))
+			rc = -EFAULT;
+		break;
+	}
+	case TLS_CIPHER_AES_GCM_256: {
+		struct tls12_crypto_info_aes_gcm_256 *
+		  crypto_info_aes_gcm_256 =
+		  container_of(crypto_info,
+			       struct tls12_crypto_info_aes_gcm_256,
+			       info);
+
+		if (len != sizeof(*crypto_info_aes_gcm_256)) {
+			rc = -EINVAL;
+			goto out;
+		}
+		memcpy(crypto_info_aes_gcm_256->iv,
+		       cctx->iv + TLS_CIPHER_AES_GCM_256_SALT_SIZE,
+		       TLS_CIPHER_AES_GCM_256_IV_SIZE);
+		memcpy(crypto_info_aes_gcm_256->rec_seq, cctx->rec_seq,
+		       TLS_CIPHER_AES_GCM_256_REC_SEQ_SIZE);
+		if (copy_to_user(optval,
+				 crypto_info_aes_gcm_256,
+				 sizeof(*crypto_info_aes_gcm_256)))
+			rc = -EFAULT;
+		break;
+	}
+	default:
+		rc = -EINVAL;
+	}
+
+out:
+	return rc;
+}
+
+static int do_tls_getsockopt(struct sock *sk, int optname,
+			     char __user *optval, int __user *optlen)
+{
+	int rc = 0;
+
+	lock_sock(sk);
+
+	switch (optname) {
+	case TLS_TX:
+	case TLS_RX:
+		rc = do_tls_getsockopt_conf(sk, optval, optlen,
+					    optname == TLS_TX);
+		break;
+	default:
+		rc = -ENOPROTOOPT;
+		break;
+	}
+
+	release_sock(sk);
+
+	return rc;
+}
+
+static int tls_getsockopt(struct sock *sk, int level, int optname,
+			  char __user *optval, int __user *optlen)
+{
+	struct tls_context *ctx = tls_get_ctx(sk);
+
+	if (level != SOL_TLS)
+		return ctx->sk_proto->getsockopt(sk, level,
+						 optname, optval, optlen);
+
+	return do_tls_getsockopt(sk, optname, optval, optlen);
+}
+
+static int do_tls_setsockopt_conf(struct sock *sk, sockptr_t optval,
+				  unsigned int optlen, int tx)
+{
+	struct tls_crypto_info *crypto_info;
+	struct tls_crypto_info *alt_crypto_info;
+	struct tls_context *ctx = tls_get_ctx(sk);
+	size_t optsize;
+	int rc = 0;
+	int conf;
+
+	if (sockptr_is_null(optval) || (optlen < sizeof(*crypto_info))) {
+		rc = -EINVAL;
+		goto out;
+	}
+
+	if (tx) {
+		crypto_info = &ctx->crypto_send.info;
+		alt_crypto_info = &ctx->crypto_recv.info;
+	} else {
+		crypto_info = &ctx->crypto_recv.info;
+		alt_crypto_info = &ctx->crypto_send.info;
+	}
+
+	/* Currently we don't support set crypto info more than one time */
+	if (TLS_CRYPTO_INFO_READY(crypto_info)) {
+		rc = -EBUSY;
+		goto out;
+	}
+
+	rc = copy_from_sockptr(crypto_info, optval, sizeof(*crypto_info));
+	if (rc) {
+		rc = -EFAULT;
+		goto err_crypto_info;
+	}
+
+	/* check version */
+	if (crypto_info->version != TLS_1_2_VERSION &&
+	    crypto_info->version != TLS_1_3_VERSION) {
+		rc = -EINVAL;
+		goto err_crypto_info;
+	}
+
+	/* Ensure that TLS version and ciphers are same in both directions */
+	if (TLS_CRYPTO_INFO_READY(alt_crypto_info)) {
+		if (alt_crypto_info->version != crypto_info->version ||
+		    alt_crypto_info->cipher_type != crypto_info->cipher_type) {
+			rc = -EINVAL;
+			goto err_crypto_info;
+		}
+	}
+
+	switch (crypto_info->cipher_type) {
+	case TLS_CIPHER_AES_GCM_128:
+		optsize = sizeof(struct tls12_crypto_info_aes_gcm_128);
+		break;
+	case TLS_CIPHER_AES_GCM_256: {
+		optsize = sizeof(struct tls12_crypto_info_aes_gcm_256);
+		break;
+	}
+	case TLS_CIPHER_AES_CCM_128:
+		optsize = sizeof(struct tls12_crypto_info_aes_ccm_128);
+		break;
+	default:
+		rc = -EINVAL;
+		goto err_crypto_info;
+	}
+
+	if (optlen != optsize) {
+		rc = -EINVAL;
+		goto err_crypto_info;
+	}
+
+	rc = copy_from_sockptr_offset(crypto_info + 1, optval,
+				      sizeof(*crypto_info),
+				      optlen - sizeof(*crypto_info));
+	if (rc) {
+		rc = -EFAULT;
+		goto err_crypto_info;
+	}
+
+	if (tx) {
+		rc = tls_set_device_offload(sk, ctx);
+		conf = TLS_HW;
+		if (!rc) {
+			TLS_INC_STATS(sock_net(sk), LINUX_MIB_TLSTXDEVICE);
+			TLS_INC_STATS(sock_net(sk), LINUX_MIB_TLSCURRTXDEVICE);
+		} else {
+			rc = tls_set_sw_offload(sk, ctx, 1);
+			if (rc)
+				goto err_crypto_info;
+			TLS_INC_STATS(sock_net(sk), LINUX_MIB_TLSTXSW);
+			TLS_INC_STATS(sock_net(sk), LINUX_MIB_TLSCURRTXSW);
+			conf = TLS_SW;
+		}
+	} else {
+		rc = tls_set_device_offload_rx(sk, ctx);
+		conf = TLS_HW;
+		if (!rc) {
+			TLS_INC_STATS(sock_net(sk), LINUX_MIB_TLSRXDEVICE);
+			TLS_INC_STATS(sock_net(sk), LINUX_MIB_TLSCURRRXDEVICE);
+		} else {
+			rc = tls_set_sw_offload(sk, ctx, 0);
+			if (rc)
+				goto err_crypto_info;
+			TLS_INC_STATS(sock_net(sk), LINUX_MIB_TLSRXSW);
+			TLS_INC_STATS(sock_net(sk), LINUX_MIB_TLSCURRRXSW);
+			conf = TLS_SW;
+		}
+		tls_sw_strparser_arm(sk, ctx);
+	}
+
+	if (tx)
+		ctx->tx_conf = conf;
+	else
+		ctx->rx_conf = conf;
+	update_sk_prot(sk, ctx);
+	if (tx) {
+		ctx->sk_write_space = sk->sk_write_space;
+		sk->sk_write_space = tls_write_space;
+	}
+	goto out;
+
+err_crypto_info:
+	memzero_explicit(crypto_info, sizeof(union tls_crypto_context));
+out:
+	return rc;
+}
+
+static int do_tls_setsockopt(struct sock *sk, int optname, sockptr_t optval,
+			     unsigned int optlen)
+{
+	int rc = 0;
+
+	switch (optname) {
+	case TLS_TX:
+	case TLS_RX:
+		lock_sock(sk);
+		rc = do_tls_setsockopt_conf(sk, optval, optlen,
+					    optname == TLS_TX);
+		release_sock(sk);
+		break;
+	default:
+		rc = -ENOPROTOOPT;
+		break;
+	}
+	return rc;
+}
+
+static int tls_setsockopt(struct sock *sk, int level, int optname,
+			  sockptr_t optval, unsigned int optlen)
+{
+	struct tls_context *ctx = tls_get_ctx(sk);
+
+	if (level != SOL_TLS)
+		return ctx->sk_proto->setsockopt(sk, level, optname, optval,
+						 optlen);
+
+	return do_tls_setsockopt(sk, optname, optval, optlen);
+}
+
+struct tls_context *tls_ctx_create(struct sock *sk)
+{
+	struct inet_connection_sock *icsk = inet_csk(sk);
+	struct tls_context *ctx;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_ATOMIC);
+	if (!ctx)
+		return NULL;
+
+	mutex_init(&ctx->tx_lock);
+	rcu_assign_pointer(icsk->icsk_ulp_data, ctx);
+	ctx->sk_proto = READ_ONCE(sk->sk_prot);
+	ctx->sk = sk;
+	return ctx;
+}
+
+static void build_proto_ops(struct proto_ops ops[TLS_NUM_CONFIG][TLS_NUM_CONFIG],
+			    const struct proto_ops *base)
+{
+	ops[TLS_BASE][TLS_BASE] = *base;
+
+	ops[TLS_SW  ][TLS_BASE] = ops[TLS_BASE][TLS_BASE];
+	ops[TLS_SW  ][TLS_BASE].sendpage_locked	= tls_sw_sendpage_locked;
+
+	ops[TLS_BASE][TLS_SW  ] = ops[TLS_BASE][TLS_BASE];
+	ops[TLS_BASE][TLS_SW  ].splice_read	= tls_sw_splice_read;
+
+	ops[TLS_SW  ][TLS_SW  ] = ops[TLS_SW  ][TLS_BASE];
+	ops[TLS_SW  ][TLS_SW  ].splice_read	= tls_sw_splice_read;
+
+#ifdef CONFIG_TLS_DEVICE
+	ops[TLS_HW  ][TLS_BASE] = ops[TLS_BASE][TLS_BASE];
+	ops[TLS_HW  ][TLS_BASE].sendpage_locked	= NULL;
+
+	ops[TLS_HW  ][TLS_SW  ] = ops[TLS_BASE][TLS_SW  ];
+	ops[TLS_HW  ][TLS_SW  ].sendpage_locked	= NULL;
+
+	ops[TLS_BASE][TLS_HW  ] = ops[TLS_BASE][TLS_SW  ];
+
+	ops[TLS_SW  ][TLS_HW  ] = ops[TLS_SW  ][TLS_SW  ];
+
+	ops[TLS_HW  ][TLS_HW  ] = ops[TLS_HW  ][TLS_SW  ];
+	ops[TLS_HW  ][TLS_HW  ].sendpage_locked	= NULL;
+#endif
+#ifdef CONFIG_TLS_TOE
+	ops[TLS_HW_RECORD][TLS_HW_RECORD] = *base;
+#endif
+}
+
+static void tls_build_proto(struct sock *sk)
+{
+	int ip_ver = sk->sk_family == AF_INET6 ? TLSV6 : TLSV4;
+	struct proto *prot = READ_ONCE(sk->sk_prot);
+
+	/* Build IPv6 TLS whenever the address of tcpv6 _prot changes */
+	if (ip_ver == TLSV6 &&
+	    unlikely(prot != smp_load_acquire(&saved_tcpv6_prot))) {
+		mutex_lock(&tcpv6_prot_mutex);
+		if (likely(prot != saved_tcpv6_prot)) {
+			build_protos(tls_prots[TLSV6], prot);
+			build_proto_ops(tls_proto_ops[TLSV6],
+					sk->sk_socket->ops);
+			smp_store_release(&saved_tcpv6_prot, prot);
+		}
+		mutex_unlock(&tcpv6_prot_mutex);
+	}
+
+	if (ip_ver == TLSV4 &&
+	    unlikely(prot != smp_load_acquire(&saved_tcpv4_prot))) {
+		mutex_lock(&tcpv4_prot_mutex);
+		if (likely(prot != saved_tcpv4_prot)) {
+			build_protos(tls_prots[TLSV4], prot);
+			build_proto_ops(tls_proto_ops[TLSV4],
+					sk->sk_socket->ops);
+			smp_store_release(&saved_tcpv4_prot, prot);
+		}
+		mutex_unlock(&tcpv4_prot_mutex);
+	}
+}
+
+static void build_protos(struct proto prot[TLS_NUM_CONFIG][TLS_NUM_CONFIG],
+			 const struct proto *base)
+{
+	prot[TLS_BASE][TLS_BASE] = *base;
+	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_SW][TLS_BASE] = prot[TLS_BASE][TLS_BASE];
+	prot[TLS_SW][TLS_BASE].sendmsg		= tls_sw_sendmsg;
+	prot[TLS_SW][TLS_BASE].sendpage		= tls_sw_sendpage;
+
+	prot[TLS_BASE][TLS_SW] = prot[TLS_BASE][TLS_BASE];
+	prot[TLS_BASE][TLS_SW].recvmsg		  = tls_sw_recvmsg;
+	prot[TLS_BASE][TLS_SW].stream_memory_read = tls_sw_stream_read;
+	prot[TLS_BASE][TLS_SW].close		  = tls_sk_proto_close;
+
+	prot[TLS_SW][TLS_SW] = prot[TLS_SW][TLS_BASE];
+	prot[TLS_SW][TLS_SW].recvmsg		= tls_sw_recvmsg;
+	prot[TLS_SW][TLS_SW].stream_memory_read	= tls_sw_stream_read;
+	prot[TLS_SW][TLS_SW].close		= tls_sk_proto_close;
+
+#ifdef CONFIG_TLS_DEVICE
+	prot[TLS_HW][TLS_BASE] = prot[TLS_BASE][TLS_BASE];
+	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].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_SW][TLS_HW] = prot[TLS_SW][TLS_SW];
+
+	prot[TLS_HW][TLS_HW] = prot[TLS_HW][TLS_SW];
+#endif
+#ifdef CONFIG_TLS_TOE
+	prot[TLS_HW_RECORD][TLS_HW_RECORD] = *base;
+	prot[TLS_HW_RECORD][TLS_HW_RECORD].hash		= tls_toe_hash;
+	prot[TLS_HW_RECORD][TLS_HW_RECORD].unhash	= tls_toe_unhash;
+#endif
+}
+
+static int tls_init(struct sock *sk)
+{
+	struct tls_context *ctx;
+	int rc = 0;
+
+	tls_build_proto(sk);
+
+#ifdef CONFIG_TLS_TOE
+	if (tls_toe_bypass(sk))
+		return 0;
+#endif
+
+	/* The TLS ulp is currently supported only for TCP sockets
+	 * in ESTABLISHED state.
+	 * Supporting sockets in LISTEN state will require us
+	 * to modify the accept implementation to clone rather then
+	 * share the ulp context.
+	 */
+	if (sk->sk_state != TCP_ESTABLISHED)
+		return -ENOTCONN;
+
+	/* allocate tls context */
+	write_lock_bh(&sk->sk_callback_lock);
+	ctx = tls_ctx_create(sk);
+	if (!ctx) {
+		rc = -ENOMEM;
+		goto out;
+	}
+
+	ctx->tx_conf = TLS_BASE;
+	ctx->rx_conf = TLS_BASE;
+	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,
+		       void (*write_space)(struct sock *sk))
+{
+	struct tls_context *ctx;
+
+	ctx = tls_get_ctx(sk);
+	if (likely(ctx)) {
+		ctx->sk_write_space = write_space;
+		ctx->sk_proto = p;
+	} else {
+		/* Pairs with lockless read in sk_clone_lock(). */
+		WRITE_ONCE(sk->sk_prot, p);
+		sk->sk_write_space = write_space;
+	}
+}
+
+static int tls_get_info(const struct sock *sk, struct sk_buff *skb)
+{
+	u16 version, cipher_type;
+	struct tls_context *ctx;
+	struct nlattr *start;
+	int err;
+
+	start = nla_nest_start_noflag(skb, INET_ULP_INFO_TLS);
+	if (!start)
+		return -EMSGSIZE;
+
+	rcu_read_lock();
+	ctx = rcu_dereference(inet_csk(sk)->icsk_ulp_data);
+	if (!ctx) {
+		err = 0;
+		goto nla_failure;
+	}
+	version = ctx->prot_info.version;
+	if (version) {
+		err = nla_put_u16(skb, TLS_INFO_VERSION, version);
+		if (err)
+			goto nla_failure;
+	}
+	cipher_type = ctx->prot_info.cipher_type;
+	if (cipher_type) {
+		err = nla_put_u16(skb, TLS_INFO_CIPHER, cipher_type);
+		if (err)
+			goto nla_failure;
+	}
+	err = nla_put_u16(skb, TLS_INFO_TXCONF, tls_user_config(ctx, true));
+	if (err)
+		goto nla_failure;
+
+	err = nla_put_u16(skb, TLS_INFO_RXCONF, tls_user_config(ctx, false));
+	if (err)
+		goto nla_failure;
+
+	rcu_read_unlock();
+	nla_nest_end(skb, start);
+	return 0;
+
+nla_failure:
+	rcu_read_unlock();
+	nla_nest_cancel(skb, start);
+	return err;
+}
+
+static size_t tls_get_info_size(const struct sock *sk)
+{
+	size_t size = 0;
+
+	size += nla_total_size(0) +		/* INET_ULP_INFO_TLS */
+		nla_total_size(sizeof(u16)) +	/* TLS_INFO_VERSION */
+		nla_total_size(sizeof(u16)) +	/* TLS_INFO_CIPHER */
+		nla_total_size(sizeof(u16)) +	/* TLS_INFO_RXCONF */
+		nla_total_size(sizeof(u16)) +	/* TLS_INFO_TXCONF */
+		0;
+
+	return size;
+}
+
+static int __net_init tls_init_net(struct net *net)
+{
+	int err;
+
+	net->mib.tls_statistics = alloc_percpu(struct linux_tls_mib);
+	if (!net->mib.tls_statistics)
+		return -ENOMEM;
+
+	err = tls_proc_init(net);
+	if (err)
+		goto err_free_stats;
+
+	return 0;
+err_free_stats:
+	free_percpu(net->mib.tls_statistics);
+	return err;
+}
+
+static void __net_exit tls_exit_net(struct net *net)
+{
+	tls_proc_fini(net);
+	free_percpu(net->mib.tls_statistics);
+}
+
+static struct pernet_operations tls_proc_ops = {
+	.init = tls_init_net,
+	.exit = tls_exit_net,
+};
+
+static struct tcp_ulp_ops tcp_tls_ulp_ops __read_mostly = {
+	.name			= "tls",
+	.owner			= THIS_MODULE,
+	.init			= tls_init,
+	.update			= tls_update,
+	.get_info		= tls_get_info,
+	.get_info_size		= tls_get_info_size,
+};
+
+static int __init tls_register(void)
+{
+	int err;
+
+	err = register_pernet_subsys(&tls_proc_ops);
+	if (err)
+		return err;
+
+	err = tls_device_init();
+	if (err) {
+		unregister_pernet_subsys(&tls_proc_ops);
+		return err;
+	}
+
+	tcp_register_ulp(&tcp_tls_ulp_ops);
+
+	return 0;
+}
+
+static void __exit tls_unregister(void)
+{
+	tcp_unregister_ulp(&tcp_tls_ulp_ops);
+	tls_device_cleanup();
+	unregister_pernet_subsys(&tls_proc_ops);
+}
+
+module_init(tls_register);
+module_exit(tls_unregister);
diff --git a/net/tls/tls_proc.c b/net/tls/tls_proc.c
new file mode 100644
index 000000000..feeceb0e4
--- /dev/null
+++ b/net/tls/tls_proc.c
@@ -0,0 +1,52 @@
+// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+/* Copyright (C) 2019 Netronome Systems, Inc. */
+
+#include <linux/proc_fs.h>
+#include <linux/seq_file.h>
+#include <net/snmp.h>
+#include <net/tls.h>
+
+#ifdef CONFIG_PROC_FS
+static const struct snmp_mib tls_mib_list[] = {
+	SNMP_MIB_ITEM("TlsCurrTxSw", LINUX_MIB_TLSCURRTXSW),
+	SNMP_MIB_ITEM("TlsCurrRxSw", LINUX_MIB_TLSCURRRXSW),
+	SNMP_MIB_ITEM("TlsCurrTxDevice", LINUX_MIB_TLSCURRTXDEVICE),
+	SNMP_MIB_ITEM("TlsCurrRxDevice", LINUX_MIB_TLSCURRRXDEVICE),
+	SNMP_MIB_ITEM("TlsTxSw", LINUX_MIB_TLSTXSW),
+	SNMP_MIB_ITEM("TlsRxSw", LINUX_MIB_TLSRXSW),
+	SNMP_MIB_ITEM("TlsTxDevice", LINUX_MIB_TLSTXDEVICE),
+	SNMP_MIB_ITEM("TlsRxDevice", LINUX_MIB_TLSRXDEVICE),
+	SNMP_MIB_ITEM("TlsDecryptError", LINUX_MIB_TLSDECRYPTERROR),
+	SNMP_MIB_ITEM("TlsRxDeviceResync", LINUX_MIB_TLSRXDEVICERESYNC),
+	SNMP_MIB_SENTINEL
+};
+
+static int tls_statistics_seq_show(struct seq_file *seq, void *v)
+{
+	unsigned long buf[LINUX_MIB_TLSMAX] = {};
+	struct net *net = seq->private;
+	int i;
+
+	snmp_get_cpu_field_batch(buf, tls_mib_list, net->mib.tls_statistics);
+	for (i = 0; tls_mib_list[i].name; i++)
+		seq_printf(seq, "%-32s\t%lu\n", tls_mib_list[i].name, buf[i]);
+
+	return 0;
+}
+#endif
+
+int __net_init tls_proc_init(struct net *net)
+{
+#ifdef CONFIG_PROC_FS
+	if (!proc_create_net_single("tls_stat", 0444, net->proc_net,
+				    tls_statistics_seq_show, NULL))
+		return -ENOMEM;
+#endif /* CONFIG_PROC_FS */
+
+	return 0;
+}
+
+void __net_exit tls_proc_fini(struct net *net)
+{
+	remove_proc_entry("tls_stat", net->proc_net);
+}
diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c
new file mode 100644
index 000000000..dd980438f
--- /dev/null
+++ b/net/tls/tls_sw.c
@@ -0,0 +1,2535 @@
+/*
+ * Copyright (c) 2016-2017, Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2016-2017, Dave Watson <davejwatson@fb.com>. All rights reserved.
+ * Copyright (c) 2016-2017, Lance Chao <lancerchao@fb.com>. All rights reserved.
+ * Copyright (c) 2016, Fridolin Pokorny <fridolin.pokorny@gmail.com>. All rights reserved.
+ * Copyright (c) 2016, Nikos Mavrogiannopoulos <nmav@gnutls.org>. All rights reserved.
+ * Copyright (c) 2018, Covalent IO, Inc. http://covalent.io
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/bug.h>
+#include <linux/sched/signal.h>
+#include <linux/module.h>
+#include <linux/splice.h>
+#include <crypto/aead.h>
+
+#include <net/strparser.h>
+#include <net/tls.h>
+
+noinline void tls_err_abort(struct sock *sk, int err)
+{
+	WARN_ON_ONCE(err >= 0);
+	/* sk->sk_err should contain a positive error code. */
+	sk->sk_err = -err;
+	sk->sk_error_report(sk);
+}
+
+static int __skb_nsg(struct sk_buff *skb, int offset, int len,
+                     unsigned int recursion_level)
+{
+        int start = skb_headlen(skb);
+        int i, chunk = start - offset;
+        struct sk_buff *frag_iter;
+        int elt = 0;
+
+        if (unlikely(recursion_level >= 24))
+                return -EMSGSIZE;
+
+        if (chunk > 0) {
+                if (chunk > len)
+                        chunk = len;
+                elt++;
+                len -= chunk;
+                if (len == 0)
+                        return elt;
+                offset += chunk;
+        }
+
+        for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
+                int end;
+
+                WARN_ON(start > offset + len);
+
+                end = start + skb_frag_size(&skb_shinfo(skb)->frags[i]);
+                chunk = end - offset;
+                if (chunk > 0) {
+                        if (chunk > len)
+                                chunk = len;
+                        elt++;
+                        len -= chunk;
+                        if (len == 0)
+                                return elt;
+                        offset += chunk;
+                }
+                start = end;
+        }
+
+        if (unlikely(skb_has_frag_list(skb))) {
+                skb_walk_frags(skb, frag_iter) {
+                        int end, ret;
+
+                        WARN_ON(start > offset + len);
+
+                        end = start + frag_iter->len;
+                        chunk = end - offset;
+                        if (chunk > 0) {
+                                if (chunk > len)
+                                        chunk = len;
+                                ret = __skb_nsg(frag_iter, offset - start, chunk,
+                                                recursion_level + 1);
+                                if (unlikely(ret < 0))
+                                        return ret;
+                                elt += ret;
+                                len -= chunk;
+                                if (len == 0)
+                                        return elt;
+                                offset += chunk;
+                        }
+                        start = end;
+                }
+        }
+        BUG_ON(len);
+        return elt;
+}
+
+/* Return the number of scatterlist elements required to completely map the
+ * skb, or -EMSGSIZE if the recursion depth is exceeded.
+ */
+static int skb_nsg(struct sk_buff *skb, int offset, int len)
+{
+        return __skb_nsg(skb, offset, len, 0);
+}
+
+static int padding_length(struct tls_sw_context_rx *ctx,
+			  struct tls_prot_info *prot, struct sk_buff *skb)
+{
+	struct strp_msg *rxm = strp_msg(skb);
+	int sub = 0;
+
+	/* Determine zero-padding length */
+	if (prot->version == TLS_1_3_VERSION) {
+		char content_type = 0;
+		int err;
+		int back = 17;
+
+		while (content_type == 0) {
+			if (back > rxm->full_len - prot->prepend_size)
+				return -EBADMSG;
+			err = skb_copy_bits(skb,
+					    rxm->offset + rxm->full_len - back,
+					    &content_type, 1);
+			if (err)
+				return err;
+			if (content_type)
+				break;
+			sub++;
+			back++;
+		}
+		ctx->control = content_type;
+	}
+	return sub;
+}
+
+static void tls_decrypt_done(struct crypto_async_request *req, int err)
+{
+	struct aead_request *aead_req = (struct aead_request *)req;
+	struct scatterlist *sgout = aead_req->dst;
+	struct scatterlist *sgin = aead_req->src;
+	struct tls_sw_context_rx *ctx;
+	struct tls_context *tls_ctx;
+	struct tls_prot_info *prot;
+	struct scatterlist *sg;
+	struct sk_buff *skb;
+	unsigned int pages;
+	int pending;
+
+	skb = (struct sk_buff *)req->data;
+	tls_ctx = tls_get_ctx(skb->sk);
+	ctx = tls_sw_ctx_rx(tls_ctx);
+	prot = &tls_ctx->prot_info;
+
+	/* Propagate if there was an err */
+	if (err) {
+		if (err == -EBADMSG)
+			TLS_INC_STATS(sock_net(skb->sk),
+				      LINUX_MIB_TLSDECRYPTERROR);
+		ctx->async_wait.err = err;
+		tls_err_abort(skb->sk, err);
+	} else {
+		struct strp_msg *rxm = strp_msg(skb);
+		int pad;
+
+		pad = padding_length(ctx, prot, skb);
+		if (pad < 0) {
+			ctx->async_wait.err = pad;
+			tls_err_abort(skb->sk, pad);
+		} else {
+			rxm->full_len -= pad;
+			rxm->offset += prot->prepend_size;
+			rxm->full_len -= prot->overhead_size;
+		}
+	}
+
+	/* After using skb->sk to propagate sk through crypto async callback
+	 * we need to NULL it again.
+	 */
+	skb->sk = NULL;
+
+
+	/* Free the destination pages if skb was not decrypted inplace */
+	if (sgout != sgin) {
+		/* Skip the first S/G entry as it points to AAD */
+		for_each_sg(sg_next(sgout), sg, UINT_MAX, pages) {
+			if (!sg)
+				break;
+			put_page(sg_page(sg));
+		}
+	}
+
+	kfree(aead_req);
+
+	spin_lock_bh(&ctx->decrypt_compl_lock);
+	pending = atomic_dec_return(&ctx->decrypt_pending);
+
+	if (!pending && ctx->async_notify)
+		complete(&ctx->async_wait.completion);
+	spin_unlock_bh(&ctx->decrypt_compl_lock);
+}
+
+static int tls_do_decryption(struct sock *sk,
+			     struct sk_buff *skb,
+			     struct scatterlist *sgin,
+			     struct scatterlist *sgout,
+			     char *iv_recv,
+			     size_t data_len,
+			     struct aead_request *aead_req,
+			     bool async)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+	int ret;
+
+	aead_request_set_tfm(aead_req, ctx->aead_recv);
+	aead_request_set_ad(aead_req, prot->aad_size);
+	aead_request_set_crypt(aead_req, sgin, sgout,
+			       data_len + prot->tag_size,
+			       (u8 *)iv_recv);
+
+	if (async) {
+		/* Using skb->sk to push sk through to crypto async callback
+		 * handler. This allows propagating errors up to the socket
+		 * if needed. It _must_ be cleared in the async handler
+		 * before consume_skb is called. We _know_ skb->sk is NULL
+		 * because it is a clone from strparser.
+		 */
+		skb->sk = sk;
+		aead_request_set_callback(aead_req,
+					  CRYPTO_TFM_REQ_MAY_BACKLOG,
+					  tls_decrypt_done, skb);
+		atomic_inc(&ctx->decrypt_pending);
+	} else {
+		aead_request_set_callback(aead_req,
+					  CRYPTO_TFM_REQ_MAY_BACKLOG,
+					  crypto_req_done, &ctx->async_wait);
+	}
+
+	ret = crypto_aead_decrypt(aead_req);
+	if (ret == -EINPROGRESS) {
+		if (async)
+			return ret;
+
+		ret = crypto_wait_req(ret, &ctx->async_wait);
+	}
+
+	if (async)
+		atomic_dec(&ctx->decrypt_pending);
+
+	return ret;
+}
+
+static void tls_trim_both_msgs(struct sock *sk, int target_size)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
+	struct tls_rec *rec = ctx->open_rec;
+
+	sk_msg_trim(sk, &rec->msg_plaintext, target_size);
+	if (target_size > 0)
+		target_size += prot->overhead_size;
+	sk_msg_trim(sk, &rec->msg_encrypted, target_size);
+}
+
+static int tls_alloc_encrypted_msg(struct sock *sk, int len)
+{
+	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 = ctx->open_rec;
+	struct sk_msg *msg_en = &rec->msg_encrypted;
+
+	return sk_msg_alloc(sk, msg_en, len, 0);
+}
+
+static int tls_clone_plaintext_msg(struct sock *sk, int required)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
+	struct tls_rec *rec = ctx->open_rec;
+	struct sk_msg *msg_pl = &rec->msg_plaintext;
+	struct sk_msg *msg_en = &rec->msg_encrypted;
+	int skip, len;
+
+	/* We add page references worth len bytes from encrypted sg
+	 * at the end of plaintext sg. It is guaranteed that msg_en
+	 * has enough required room (ensured by caller).
+	 */
+	len = required - msg_pl->sg.size;
+
+	/* Skip initial bytes in msg_en's data to be able to use
+	 * same offset of both plain and encrypted data.
+	 */
+	skip = prot->prepend_size + msg_pl->sg.size;
+
+	return sk_msg_clone(sk, msg_pl, msg_en, skip, len);
+}
+
+static struct tls_rec *tls_get_rec(struct sock *sk)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
+	struct sk_msg *msg_pl, *msg_en;
+	struct tls_rec *rec;
+	int mem_size;
+
+	mem_size = sizeof(struct tls_rec) + crypto_aead_reqsize(ctx->aead_send);
+
+	rec = kzalloc(mem_size, sk->sk_allocation);
+	if (!rec)
+		return NULL;
+
+	msg_pl = &rec->msg_plaintext;
+	msg_en = &rec->msg_encrypted;
+
+	sk_msg_init(msg_pl);
+	sk_msg_init(msg_en);
+
+	sg_init_table(rec->sg_aead_in, 2);
+	sg_set_buf(&rec->sg_aead_in[0], rec->aad_space, prot->aad_size);
+	sg_unmark_end(&rec->sg_aead_in[1]);
+
+	sg_init_table(rec->sg_aead_out, 2);
+	sg_set_buf(&rec->sg_aead_out[0], rec->aad_space, prot->aad_size);
+	sg_unmark_end(&rec->sg_aead_out[1]);
+
+	return rec;
+}
+
+static void tls_free_rec(struct sock *sk, struct tls_rec *rec)
+{
+	sk_msg_free(sk, &rec->msg_encrypted);
+	sk_msg_free(sk, &rec->msg_plaintext);
+	kfree(rec);
+}
+
+static void tls_free_open_rec(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 = ctx->open_rec;
+
+	if (rec) {
+		tls_free_rec(sk, rec);
+		ctx->open_rec = NULL;
+	}
+}
+
+int tls_tx_records(struct sock *sk, int flags)
+{
+	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;
+	struct sk_msg *msg_en;
+	int tx_flags, rc = 0;
+
+	if (tls_is_partially_sent_record(tls_ctx)) {
+		rec = list_first_entry(&ctx->tx_list,
+				       struct tls_rec, list);
+
+		if (flags == -1)
+			tx_flags = rec->tx_flags;
+		else
+			tx_flags = flags;
+
+		rc = tls_push_partial_record(sk, tls_ctx, tx_flags);
+		if (rc)
+			goto tx_err;
+
+		/* Full record has been transmitted.
+		 * Remove the head of tx_list
+		 */
+		list_del(&rec->list);
+		sk_msg_free(sk, &rec->msg_plaintext);
+		kfree(rec);
+	}
+
+	/* Tx all ready records */
+	list_for_each_entry_safe(rec, tmp, &ctx->tx_list, list) {
+		if (READ_ONCE(rec->tx_ready)) {
+			if (flags == -1)
+				tx_flags = rec->tx_flags;
+			else
+				tx_flags = flags;
+
+			msg_en = &rec->msg_encrypted;
+			rc = tls_push_sg(sk, tls_ctx,
+					 &msg_en->sg.data[msg_en->sg.curr],
+					 0, tx_flags);
+			if (rc)
+				goto tx_err;
+
+			list_del(&rec->list);
+			sk_msg_free(sk, &rec->msg_plaintext);
+			kfree(rec);
+		} else {
+			break;
+		}
+	}
+
+tx_err:
+	if (rc < 0 && rc != -EAGAIN)
+		tls_err_abort(sk, -EBADMSG);
+
+	return rc;
+}
+
+static void tls_encrypt_done(struct crypto_async_request *req, int err)
+{
+	struct aead_request *aead_req = (struct aead_request *)req;
+	struct sock *sk = req->data;
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
+	struct scatterlist *sge;
+	struct sk_msg *msg_en;
+	struct tls_rec *rec;
+	bool ready = false;
+	int pending;
+
+	rec = container_of(aead_req, struct tls_rec, aead_req);
+	msg_en = &rec->msg_encrypted;
+
+	sge = sk_msg_elem(msg_en, msg_en->sg.curr);
+	sge->offset -= prot->prepend_size;
+	sge->length += prot->prepend_size;
+
+	/* Check if error is previously set on socket */
+	if (err || sk->sk_err) {
+		rec = NULL;
+
+		/* If err is already set on socket, return the same code */
+		if (sk->sk_err) {
+			ctx->async_wait.err = -sk->sk_err;
+		} else {
+			ctx->async_wait.err = err;
+			tls_err_abort(sk, err);
+		}
+	}
+
+	if (rec) {
+		struct tls_rec *first_rec;
+
+		/* Mark the record as ready for transmission */
+		smp_store_mb(rec->tx_ready, true);
+
+		/* 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;
+	}
+
+	spin_lock_bh(&ctx->encrypt_compl_lock);
+	pending = atomic_dec_return(&ctx->encrypt_pending);
+
+	if (!pending && ctx->async_notify)
+		complete(&ctx->async_wait.completion);
+	spin_unlock_bh(&ctx->encrypt_compl_lock);
+
+	if (!ready)
+		return;
+
+	/* Schedule the transmission */
+	if (!test_and_set_bit(BIT_TX_SCHEDULED, &ctx->tx_bitmask))
+		schedule_delayed_work(&ctx->tx_work.work, 1);
+}
+
+static int tls_do_encryption(struct sock *sk,
+			     struct tls_context *tls_ctx,
+			     struct tls_sw_context_tx *ctx,
+			     struct aead_request *aead_req,
+			     size_t data_len, u32 start)
+{
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_rec *rec = ctx->open_rec;
+	struct sk_msg *msg_en = &rec->msg_encrypted;
+	struct scatterlist *sge = sk_msg_elem(msg_en, start);
+	int rc, iv_offset = 0;
+
+	/* For CCM based ciphers, first byte of IV is a constant */
+	if (prot->cipher_type == TLS_CIPHER_AES_CCM_128) {
+		rec->iv_data[0] = TLS_AES_CCM_IV_B0_BYTE;
+		iv_offset = 1;
+	}
+
+	memcpy(&rec->iv_data[iv_offset], tls_ctx->tx.iv,
+	       prot->iv_size + prot->salt_size);
+
+	xor_iv_with_seq(prot->version, rec->iv_data + iv_offset, tls_ctx->tx.rec_seq);
+
+	sge->offset += prot->prepend_size;
+	sge->length -= prot->prepend_size;
+
+	msg_en->sg.curr = start;
+
+	aead_request_set_tfm(aead_req, ctx->aead_send);
+	aead_request_set_ad(aead_req, prot->aad_size);
+	aead_request_set_crypt(aead_req, rec->sg_aead_in,
+			       rec->sg_aead_out,
+			       data_len, rec->iv_data);
+
+	aead_request_set_callback(aead_req, CRYPTO_TFM_REQ_MAY_BACKLOG,
+				  tls_encrypt_done, sk);
+
+	/* Add the record in tx_list */
+	list_add_tail((struct list_head *)&rec->list, &ctx->tx_list);
+	atomic_inc(&ctx->encrypt_pending);
+
+	rc = crypto_aead_encrypt(aead_req);
+	if (!rc || rc != -EINPROGRESS) {
+		atomic_dec(&ctx->encrypt_pending);
+		sge->offset -= prot->prepend_size;
+		sge->length += prot->prepend_size;
+	}
+
+	if (!rc) {
+		WRITE_ONCE(rec->tx_ready, true);
+	} else if (rc != -EINPROGRESS) {
+		list_del(&rec->list);
+		return rc;
+	}
+
+	/* Unhook the record from context if encryption is not failure */
+	ctx->open_rec = NULL;
+	tls_advance_record_sn(sk, prot, &tls_ctx->tx);
+	return rc;
+}
+
+static int tls_split_open_record(struct sock *sk, struct tls_rec *from,
+				 struct tls_rec **to, struct sk_msg *msg_opl,
+				 struct sk_msg *msg_oen, u32 split_point,
+				 u32 tx_overhead_size, u32 *orig_end)
+{
+	u32 i, j, bytes = 0, apply = msg_opl->apply_bytes;
+	struct scatterlist *sge, *osge, *nsge;
+	u32 orig_size = msg_opl->sg.size;
+	struct scatterlist tmp = { };
+	struct sk_msg *msg_npl;
+	struct tls_rec *new;
+	int ret;
+
+	new = tls_get_rec(sk);
+	if (!new)
+		return -ENOMEM;
+	ret = sk_msg_alloc(sk, &new->msg_encrypted, msg_opl->sg.size +
+			   tx_overhead_size, 0);
+	if (ret < 0) {
+		tls_free_rec(sk, new);
+		return ret;
+	}
+
+	*orig_end = msg_opl->sg.end;
+	i = msg_opl->sg.start;
+	sge = sk_msg_elem(msg_opl, i);
+	while (apply && sge->length) {
+		if (sge->length > apply) {
+			u32 len = sge->length - apply;
+
+			get_page(sg_page(sge));
+			sg_set_page(&tmp, sg_page(sge), len,
+				    sge->offset + apply);
+			sge->length = apply;
+			bytes += apply;
+			apply = 0;
+		} else {
+			apply -= sge->length;
+			bytes += sge->length;
+		}
+
+		sk_msg_iter_var_next(i);
+		if (i == msg_opl->sg.end)
+			break;
+		sge = sk_msg_elem(msg_opl, i);
+	}
+
+	msg_opl->sg.end = i;
+	msg_opl->sg.curr = i;
+	msg_opl->sg.copybreak = 0;
+	msg_opl->apply_bytes = 0;
+	msg_opl->sg.size = bytes;
+
+	msg_npl = &new->msg_plaintext;
+	msg_npl->apply_bytes = apply;
+	msg_npl->sg.size = orig_size - bytes;
+
+	j = msg_npl->sg.start;
+	nsge = sk_msg_elem(msg_npl, j);
+	if (tmp.length) {
+		memcpy(nsge, &tmp, sizeof(*nsge));
+		sk_msg_iter_var_next(j);
+		nsge = sk_msg_elem(msg_npl, j);
+	}
+
+	osge = sk_msg_elem(msg_opl, i);
+	while (osge->length) {
+		memcpy(nsge, osge, sizeof(*nsge));
+		sg_unmark_end(nsge);
+		sk_msg_iter_var_next(i);
+		sk_msg_iter_var_next(j);
+		if (i == *orig_end)
+			break;
+		osge = sk_msg_elem(msg_opl, i);
+		nsge = sk_msg_elem(msg_npl, j);
+	}
+
+	msg_npl->sg.end = j;
+	msg_npl->sg.curr = j;
+	msg_npl->sg.copybreak = 0;
+
+	*to = new;
+	return 0;
+}
+
+static void tls_merge_open_record(struct sock *sk, struct tls_rec *to,
+				  struct tls_rec *from, u32 orig_end)
+{
+	struct sk_msg *msg_npl = &from->msg_plaintext;
+	struct sk_msg *msg_opl = &to->msg_plaintext;
+	struct scatterlist *osge, *nsge;
+	u32 i, j;
+
+	i = msg_opl->sg.end;
+	sk_msg_iter_var_prev(i);
+	j = msg_npl->sg.start;
+
+	osge = sk_msg_elem(msg_opl, i);
+	nsge = sk_msg_elem(msg_npl, j);
+
+	if (sg_page(osge) == sg_page(nsge) &&
+	    osge->offset + osge->length == nsge->offset) {
+		osge->length += nsge->length;
+		put_page(sg_page(nsge));
+	}
+
+	msg_opl->sg.end = orig_end;
+	msg_opl->sg.curr = orig_end;
+	msg_opl->sg.copybreak = 0;
+	msg_opl->apply_bytes = msg_opl->sg.size + msg_npl->sg.size;
+	msg_opl->sg.size += msg_npl->sg.size;
+
+	sk_msg_free(sk, &to->msg_encrypted);
+	sk_msg_xfer_full(&to->msg_encrypted, &from->msg_encrypted);
+
+	kfree(from);
+}
+
+static int tls_push_record(struct sock *sk, int flags,
+			   unsigned char record_type)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
+	struct tls_rec *rec = ctx->open_rec, *tmp = NULL;
+	u32 i, split_point, orig_end;
+	struct sk_msg *msg_pl, *msg_en;
+	struct aead_request *req;
+	bool split;
+	int rc;
+
+	if (!rec)
+		return 0;
+
+	msg_pl = &rec->msg_plaintext;
+	msg_en = &rec->msg_encrypted;
+
+	split_point = msg_pl->apply_bytes;
+	split = split_point && split_point < msg_pl->sg.size;
+	if (unlikely((!split &&
+		      msg_pl->sg.size +
+		      prot->overhead_size > msg_en->sg.size) ||
+		     (split &&
+		      split_point +
+		      prot->overhead_size > msg_en->sg.size))) {
+		split = true;
+		split_point = msg_en->sg.size;
+	}
+	if (split) {
+		rc = tls_split_open_record(sk, rec, &tmp, msg_pl, msg_en,
+					   split_point, prot->overhead_size,
+					   &orig_end);
+		if (rc < 0)
+			return rc;
+		/* This can happen if above tls_split_open_record allocates
+		 * a single large encryption buffer instead of two smaller
+		 * ones. In this case adjust pointers and continue without
+		 * split.
+		 */
+		if (!msg_pl->sg.size) {
+			tls_merge_open_record(sk, rec, tmp, orig_end);
+			msg_pl = &rec->msg_plaintext;
+			msg_en = &rec->msg_encrypted;
+			split = false;
+		}
+		sk_msg_trim(sk, msg_en, msg_pl->sg.size +
+			    prot->overhead_size);
+	}
+
+	rec->tx_flags = flags;
+	req = &rec->aead_req;
+
+	i = msg_pl->sg.end;
+	sk_msg_iter_var_prev(i);
+
+	rec->content_type = record_type;
+	if (prot->version == TLS_1_3_VERSION) {
+		/* Add content type to end of message.  No padding added */
+		sg_set_buf(&rec->sg_content_type, &rec->content_type, 1);
+		sg_mark_end(&rec->sg_content_type);
+		sg_chain(msg_pl->sg.data, msg_pl->sg.end + 1,
+			 &rec->sg_content_type);
+	} else {
+		sg_mark_end(sk_msg_elem(msg_pl, i));
+	}
+
+	if (msg_pl->sg.end < msg_pl->sg.start) {
+		sg_chain(&msg_pl->sg.data[msg_pl->sg.start],
+			 MAX_SKB_FRAGS - msg_pl->sg.start + 1,
+			 msg_pl->sg.data);
+	}
+
+	i = msg_pl->sg.start;
+	sg_chain(rec->sg_aead_in, 2, &msg_pl->sg.data[i]);
+
+	i = msg_en->sg.end;
+	sk_msg_iter_var_prev(i);
+	sg_mark_end(sk_msg_elem(msg_en, i));
+
+	i = msg_en->sg.start;
+	sg_chain(rec->sg_aead_out, 2, &msg_en->sg.data[i]);
+
+	tls_make_aad(rec->aad_space, msg_pl->sg.size + prot->tail_size,
+		     tls_ctx->tx.rec_seq, prot->rec_seq_size,
+		     record_type, prot->version);
+
+	tls_fill_prepend(tls_ctx,
+			 page_address(sg_page(&msg_en->sg.data[i])) +
+			 msg_en->sg.data[i].offset,
+			 msg_pl->sg.size + prot->tail_size,
+			 record_type, prot->version);
+
+	tls_ctx->pending_open_record_frags = false;
+
+	rc = tls_do_encryption(sk, tls_ctx, ctx, req,
+			       msg_pl->sg.size + prot->tail_size, i);
+	if (rc < 0) {
+		if (rc != -EINPROGRESS) {
+			tls_err_abort(sk, -EBADMSG);
+			if (split) {
+				tls_ctx->pending_open_record_frags = true;
+				tls_merge_open_record(sk, rec, tmp, orig_end);
+			}
+		}
+		ctx->async_capable = 1;
+		return rc;
+	} else if (split) {
+		msg_pl = &tmp->msg_plaintext;
+		msg_en = &tmp->msg_encrypted;
+		sk_msg_trim(sk, msg_en, msg_pl->sg.size + prot->overhead_size);
+		tls_ctx->pending_open_record_frags = true;
+		ctx->open_rec = tmp;
+	}
+
+	return tls_tx_records(sk, flags);
+}
+
+static int bpf_exec_tx_verdict(struct sk_msg *msg, struct sock *sk,
+			       bool full_record, u8 record_type,
+			       ssize_t *copied, int flags)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
+	struct sk_msg msg_redir = { };
+	struct sk_psock *psock;
+	struct sock *sk_redir;
+	struct tls_rec *rec;
+	bool enospc, policy;
+	int err = 0, send;
+	u32 delta = 0;
+
+	policy = !(flags & MSG_SENDPAGE_NOPOLICY);
+	psock = sk_psock_get(sk);
+	if (!psock || !policy) {
+		err = tls_push_record(sk, flags, record_type);
+		if (err && err != -EINPROGRESS && sk->sk_err == EBADMSG) {
+			*copied -= sk_msg_free(sk, msg);
+			tls_free_open_rec(sk);
+			err = -sk->sk_err;
+		}
+		if (psock)
+			sk_psock_put(sk, psock);
+		return err;
+	}
+more_data:
+	enospc = sk_msg_full(msg);
+	if (psock->eval == __SK_NONE) {
+		delta = msg->sg.size;
+		psock->eval = sk_psock_msg_verdict(sk, psock, msg);
+		delta -= msg->sg.size;
+	}
+	if (msg->cork_bytes && msg->cork_bytes > msg->sg.size &&
+	    !enospc && !full_record) {
+		err = -ENOSPC;
+		goto out_err;
+	}
+	msg->cork_bytes = 0;
+	send = msg->sg.size;
+	if (msg->apply_bytes && msg->apply_bytes < send)
+		send = msg->apply_bytes;
+
+	switch (psock->eval) {
+	case __SK_PASS:
+		err = tls_push_record(sk, flags, record_type);
+		if (err && err != -EINPROGRESS && sk->sk_err == EBADMSG) {
+			*copied -= sk_msg_free(sk, msg);
+			tls_free_open_rec(sk);
+			err = -sk->sk_err;
+			goto out_err;
+		}
+		break;
+	case __SK_REDIRECT:
+		sk_redir = psock->sk_redir;
+		memcpy(&msg_redir, msg, sizeof(*msg));
+		if (msg->apply_bytes < send)
+			msg->apply_bytes = 0;
+		else
+			msg->apply_bytes -= send;
+		sk_msg_return_zero(sk, msg, send);
+		msg->sg.size -= send;
+		release_sock(sk);
+		err = tcp_bpf_sendmsg_redir(sk_redir, &msg_redir, send, flags);
+		lock_sock(sk);
+		if (err < 0) {
+			*copied -= sk_msg_free_nocharge(sk, &msg_redir);
+			msg->sg.size = 0;
+		}
+		if (msg->sg.size == 0)
+			tls_free_open_rec(sk);
+		break;
+	case __SK_DROP:
+	default:
+		sk_msg_free_partial(sk, msg, send);
+		if (msg->apply_bytes < send)
+			msg->apply_bytes = 0;
+		else
+			msg->apply_bytes -= send;
+		if (msg->sg.size == 0)
+			tls_free_open_rec(sk);
+		*copied -= (send + delta);
+		err = -EACCES;
+	}
+
+	if (likely(!err)) {
+		bool reset_eval = !ctx->open_rec;
+
+		rec = ctx->open_rec;
+		if (rec) {
+			msg = &rec->msg_plaintext;
+			if (!msg->apply_bytes)
+				reset_eval = true;
+		}
+		if (reset_eval) {
+			psock->eval = __SK_NONE;
+			if (psock->sk_redir) {
+				sock_put(psock->sk_redir);
+				psock->sk_redir = NULL;
+			}
+		}
+		if (rec)
+			goto more_data;
+	}
+ out_err:
+	sk_psock_put(sk, psock);
+	return err;
+}
+
+static int tls_sw_push_pending_record(struct sock *sk, int flags)
+{
+	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 = ctx->open_rec;
+	struct sk_msg *msg_pl;
+	size_t copied;
+
+	if (!rec)
+		return 0;
+
+	msg_pl = &rec->msg_plaintext;
+	copied = msg_pl->sg.size;
+	if (!copied)
+		return 0;
+
+	return bpf_exec_tx_verdict(msg_pl, sk, true, TLS_RECORD_TYPE_DATA,
+				   &copied, flags);
+}
+
+int tls_sw_sendmsg(struct sock *sk, struct msghdr *msg, size_t size)
+{
+	long timeo = sock_sndtimeo(sk, msg->msg_flags & MSG_DONTWAIT);
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
+	bool async_capable = ctx->async_capable;
+	unsigned char record_type = TLS_RECORD_TYPE_DATA;
+	bool is_kvec = iov_iter_is_kvec(&msg->msg_iter);
+	bool eor = !(msg->msg_flags & MSG_MORE);
+	size_t try_to_copy;
+	ssize_t copied = 0;
+	struct sk_msg *msg_pl, *msg_en;
+	struct tls_rec *rec;
+	int required_size;
+	int num_async = 0;
+	bool full_record;
+	int record_room;
+	int num_zc = 0;
+	int orig_size;
+	int ret = 0;
+	int pending;
+
+	if (msg->msg_flags & ~(MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL |
+			       MSG_CMSG_COMPAT))
+		return -EOPNOTSUPP;
+
+	ret = mutex_lock_interruptible(&tls_ctx->tx_lock);
+	if (ret)
+		return ret;
+	lock_sock(sk);
+
+	if (unlikely(msg->msg_controllen)) {
+		ret = tls_proccess_cmsg(sk, msg, &record_type);
+		if (ret) {
+			if (ret == -EINPROGRESS)
+				num_async++;
+			else if (ret != -EAGAIN)
+				goto send_end;
+		}
+	}
+
+	while (msg_data_left(msg)) {
+		if (sk->sk_err) {
+			ret = -sk->sk_err;
+			goto send_end;
+		}
+
+		if (ctx->open_rec)
+			rec = ctx->open_rec;
+		else
+			rec = ctx->open_rec = tls_get_rec(sk);
+		if (!rec) {
+			ret = -ENOMEM;
+			goto send_end;
+		}
+
+		msg_pl = &rec->msg_plaintext;
+		msg_en = &rec->msg_encrypted;
+
+		orig_size = msg_pl->sg.size;
+		full_record = false;
+		try_to_copy = msg_data_left(msg);
+		record_room = TLS_MAX_PAYLOAD_SIZE - msg_pl->sg.size;
+		if (try_to_copy >= record_room) {
+			try_to_copy = record_room;
+			full_record = true;
+		}
+
+		required_size = msg_pl->sg.size + try_to_copy +
+				prot->overhead_size;
+
+		if (!sk_stream_memory_free(sk))
+			goto wait_for_sndbuf;
+
+alloc_encrypted:
+		ret = tls_alloc_encrypted_msg(sk, required_size);
+		if (ret) {
+			if (ret != -ENOSPC)
+				goto wait_for_memory;
+
+			/* Adjust try_to_copy according to the amount that was
+			 * actually allocated. The difference is due
+			 * to max sg elements limit
+			 */
+			try_to_copy -= required_size - msg_en->sg.size;
+			full_record = true;
+		}
+
+		if (!is_kvec && (full_record || eor) && !async_capable) {
+			u32 first = msg_pl->sg.end;
+
+			ret = sk_msg_zerocopy_from_iter(sk, &msg->msg_iter,
+							msg_pl, try_to_copy);
+			if (ret)
+				goto fallback_to_reg_send;
+
+			num_zc++;
+			copied += try_to_copy;
+
+			sk_msg_sg_copy_set(msg_pl, first);
+			ret = bpf_exec_tx_verdict(msg_pl, sk, full_record,
+						  record_type, &copied,
+						  msg->msg_flags);
+			if (ret) {
+				if (ret == -EINPROGRESS)
+					num_async++;
+				else if (ret == -ENOMEM)
+					goto wait_for_memory;
+				else if (ctx->open_rec && ret == -ENOSPC)
+					goto rollback_iter;
+				else if (ret != -EAGAIN)
+					goto send_end;
+			}
+			continue;
+rollback_iter:
+			copied -= try_to_copy;
+			sk_msg_sg_copy_clear(msg_pl, first);
+			iov_iter_revert(&msg->msg_iter,
+					msg_pl->sg.size - orig_size);
+fallback_to_reg_send:
+			sk_msg_trim(sk, msg_pl, orig_size);
+		}
+
+		required_size = msg_pl->sg.size + try_to_copy;
+
+		ret = tls_clone_plaintext_msg(sk, required_size);
+		if (ret) {
+			if (ret != -ENOSPC)
+				goto send_end;
+
+			/* Adjust try_to_copy according to the amount that was
+			 * actually allocated. The difference is due
+			 * to max sg elements limit
+			 */
+			try_to_copy -= required_size - msg_pl->sg.size;
+			full_record = true;
+			sk_msg_trim(sk, msg_en,
+				    msg_pl->sg.size + prot->overhead_size);
+		}
+
+		if (try_to_copy) {
+			ret = sk_msg_memcopy_from_iter(sk, &msg->msg_iter,
+						       msg_pl, try_to_copy);
+			if (ret < 0)
+				goto trim_sgl;
+		}
+
+		/* Open records defined only if successfully copied, otherwise
+		 * we would trim the sg but not reset the open record frags.
+		 */
+		tls_ctx->pending_open_record_frags = true;
+		copied += try_to_copy;
+		if (full_record || eor) {
+			ret = bpf_exec_tx_verdict(msg_pl, sk, full_record,
+						  record_type, &copied,
+						  msg->msg_flags);
+			if (ret) {
+				if (ret == -EINPROGRESS)
+					num_async++;
+				else if (ret == -ENOMEM)
+					goto wait_for_memory;
+				else if (ret != -EAGAIN) {
+					if (ret == -ENOSPC)
+						ret = 0;
+					goto send_end;
+				}
+			}
+		}
+
+		continue;
+
+wait_for_sndbuf:
+		set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
+wait_for_memory:
+		ret = sk_stream_wait_memory(sk, &timeo);
+		if (ret) {
+trim_sgl:
+			if (ctx->open_rec)
+				tls_trim_both_msgs(sk, orig_size);
+			goto send_end;
+		}
+
+		if (ctx->open_rec && msg_en->sg.size < required_size)
+			goto alloc_encrypted;
+	}
+
+	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);
+
+		if (ctx->async_wait.err) {
+			ret = ctx->async_wait.err;
+			copied = 0;
+		}
+	}
+
+	/* Transmit if any encryptions have completed */
+	if (test_and_clear_bit(BIT_TX_SCHEDULED, &ctx->tx_bitmask)) {
+		cancel_delayed_work(&ctx->tx_work.work);
+		tls_tx_records(sk, msg->msg_flags);
+	}
+
+send_end:
+	ret = sk_stream_error(sk, msg->msg_flags, ret);
+
+	release_sock(sk);
+	mutex_unlock(&tls_ctx->tx_lock);
+	return copied > 0 ? copied : ret;
+}
+
+static int tls_sw_do_sendpage(struct sock *sk, struct page *page,
+			      int offset, size_t size, int flags)
+{
+	long timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_tx *ctx = tls_sw_ctx_tx(tls_ctx);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	unsigned char record_type = TLS_RECORD_TYPE_DATA;
+	struct sk_msg *msg_pl;
+	struct tls_rec *rec;
+	int num_async = 0;
+	ssize_t copied = 0;
+	bool full_record;
+	int record_room;
+	int ret = 0;
+	bool eor;
+
+	eor = !(flags & MSG_SENDPAGE_NOTLAST);
+	sk_clear_bit(SOCKWQ_ASYNC_NOSPACE, sk);
+
+	/* Call the sk_stream functions to manage the sndbuf mem. */
+	while (size > 0) {
+		size_t copy, required_size;
+
+		if (sk->sk_err) {
+			ret = -sk->sk_err;
+			goto sendpage_end;
+		}
+
+		if (ctx->open_rec)
+			rec = ctx->open_rec;
+		else
+			rec = ctx->open_rec = tls_get_rec(sk);
+		if (!rec) {
+			ret = -ENOMEM;
+			goto sendpage_end;
+		}
+
+		msg_pl = &rec->msg_plaintext;
+
+		full_record = false;
+		record_room = TLS_MAX_PAYLOAD_SIZE - msg_pl->sg.size;
+		copy = size;
+		if (copy >= record_room) {
+			copy = record_room;
+			full_record = true;
+		}
+
+		required_size = msg_pl->sg.size + copy + prot->overhead_size;
+
+		if (!sk_stream_memory_free(sk))
+			goto wait_for_sndbuf;
+alloc_payload:
+		ret = tls_alloc_encrypted_msg(sk, required_size);
+		if (ret) {
+			if (ret != -ENOSPC)
+				goto wait_for_memory;
+
+			/* Adjust copy according to the amount that was
+			 * actually allocated. The difference is due
+			 * to max sg elements limit
+			 */
+			copy -= required_size - msg_pl->sg.size;
+			full_record = true;
+		}
+
+		sk_msg_page_add(msg_pl, page, copy, offset);
+		msg_pl->sg.copybreak = 0;
+		msg_pl->sg.curr = msg_pl->sg.end;
+		sk_mem_charge(sk, copy);
+
+		offset += copy;
+		size -= copy;
+		copied += copy;
+
+		tls_ctx->pending_open_record_frags = true;
+		if (full_record || eor || sk_msg_full(msg_pl)) {
+			ret = bpf_exec_tx_verdict(msg_pl, sk, full_record,
+						  record_type, &copied, flags);
+			if (ret) {
+				if (ret == -EINPROGRESS)
+					num_async++;
+				else if (ret == -ENOMEM)
+					goto wait_for_memory;
+				else if (ret != -EAGAIN) {
+					if (ret == -ENOSPC)
+						ret = 0;
+					goto sendpage_end;
+				}
+			}
+		}
+		continue;
+wait_for_sndbuf:
+		set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
+wait_for_memory:
+		ret = sk_stream_wait_memory(sk, &timeo);
+		if (ret) {
+			if (ctx->open_rec)
+				tls_trim_both_msgs(sk, msg_pl->sg.size);
+			goto sendpage_end;
+		}
+
+		if (ctx->open_rec)
+			goto alloc_payload;
+	}
+
+	if (num_async) {
+		/* Transmit if any encryptions have completed */
+		if (test_and_clear_bit(BIT_TX_SCHEDULED, &ctx->tx_bitmask)) {
+			cancel_delayed_work(&ctx->tx_work.work);
+			tls_tx_records(sk, flags);
+		}
+	}
+sendpage_end:
+	ret = sk_stream_error(sk, flags, ret);
+	return copied > 0 ? copied : ret;
+}
+
+int tls_sw_sendpage_locked(struct sock *sk, struct page *page,
+			   int offset, size_t size, int flags)
+{
+	if (flags & ~(MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL |
+		      MSG_SENDPAGE_NOTLAST | MSG_SENDPAGE_NOPOLICY |
+		      MSG_NO_SHARED_FRAGS))
+		return -EOPNOTSUPP;
+
+	return tls_sw_do_sendpage(sk, page, offset, size, flags);
+}
+
+int tls_sw_sendpage(struct sock *sk, struct page *page,
+		    int offset, size_t size, int flags)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	int ret;
+
+	if (flags & ~(MSG_MORE | MSG_DONTWAIT | MSG_NOSIGNAL |
+		      MSG_SENDPAGE_NOTLAST | MSG_SENDPAGE_NOPOLICY))
+		return -EOPNOTSUPP;
+
+	ret = mutex_lock_interruptible(&tls_ctx->tx_lock);
+	if (ret)
+		return ret;
+	lock_sock(sk);
+	ret = tls_sw_do_sendpage(sk, page, offset, size, flags);
+	release_sock(sk);
+	mutex_unlock(&tls_ctx->tx_lock);
+	return ret;
+}
+
+static struct sk_buff *tls_wait_data(struct sock *sk, struct sk_psock *psock,
+				     bool nonblock, long timeo, int *err)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+	struct sk_buff *skb;
+	DEFINE_WAIT_FUNC(wait, woken_wake_function);
+
+	while (!(skb = ctx->recv_pkt) && sk_psock_queue_empty(psock)) {
+		if (sk->sk_err) {
+			*err = sock_error(sk);
+			return NULL;
+		}
+
+		if (!skb_queue_empty(&sk->sk_receive_queue)) {
+			__strp_unpause(&ctx->strp);
+			if (ctx->recv_pkt)
+				return ctx->recv_pkt;
+		}
+
+		if (sk->sk_shutdown & RCV_SHUTDOWN)
+			return NULL;
+
+		if (sock_flag(sk, SOCK_DONE))
+			return NULL;
+
+		if (nonblock || !timeo) {
+			*err = -EAGAIN;
+			return NULL;
+		}
+
+		add_wait_queue(sk_sleep(sk), &wait);
+		sk_set_bit(SOCKWQ_ASYNC_WAITDATA, sk);
+		sk_wait_event(sk, &timeo,
+			      ctx->recv_pkt != skb ||
+			      !sk_psock_queue_empty(psock),
+			      &wait);
+		sk_clear_bit(SOCKWQ_ASYNC_WAITDATA, sk);
+		remove_wait_queue(sk_sleep(sk), &wait);
+
+		/* Handle signals */
+		if (signal_pending(current)) {
+			*err = sock_intr_errno(timeo);
+			return NULL;
+		}
+	}
+
+	return skb;
+}
+
+static int tls_setup_from_iter(struct sock *sk, struct iov_iter *from,
+			       int length, int *pages_used,
+			       unsigned int *size_used,
+			       struct scatterlist *to,
+			       int to_max_pages)
+{
+	int rc = 0, i = 0, num_elem = *pages_used, maxpages;
+	struct page *pages[MAX_SKB_FRAGS];
+	unsigned int size = *size_used;
+	ssize_t copied, use;
+	size_t offset;
+
+	while (length > 0) {
+		i = 0;
+		maxpages = to_max_pages - num_elem;
+		if (maxpages == 0) {
+			rc = -EFAULT;
+			goto out;
+		}
+		copied = iov_iter_get_pages(from, pages,
+					    length,
+					    maxpages, &offset);
+		if (copied <= 0) {
+			rc = -EFAULT;
+			goto out;
+		}
+
+		iov_iter_advance(from, copied);
+
+		length -= copied;
+		size += copied;
+		while (copied) {
+			use = min_t(int, copied, PAGE_SIZE - offset);
+
+			sg_set_page(&to[num_elem],
+				    pages[i], use, offset);
+			sg_unmark_end(&to[num_elem]);
+			/* We do not uncharge memory from this API */
+
+			offset = 0;
+			copied -= use;
+
+			i++;
+			num_elem++;
+		}
+	}
+	/* Mark the end in the last sg entry if newly added */
+	if (num_elem > *pages_used)
+		sg_mark_end(&to[num_elem - 1]);
+out:
+	if (rc)
+		iov_iter_revert(from, size - *size_used);
+	*size_used = size;
+	*pages_used = num_elem;
+
+	return rc;
+}
+
+/* This function decrypts the input skb into either out_iov or in out_sg
+ * or in skb buffers itself. The input parameter 'zc' indicates if
+ * zero-copy mode needs to be tried or not. With zero-copy mode, either
+ * out_iov or out_sg must be non-NULL. In case both out_iov and out_sg are
+ * NULL, then the decryption happens inside skb buffers itself, i.e.
+ * zero-copy gets disabled and 'zc' is updated.
+ */
+
+static int decrypt_internal(struct sock *sk, struct sk_buff *skb,
+			    struct iov_iter *out_iov,
+			    struct scatterlist *out_sg,
+			    int *chunk, bool *zc, bool async)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct strp_msg *rxm = strp_msg(skb);
+	int n_sgin, n_sgout, nsg, mem_size, aead_size, err, pages = 0;
+	struct aead_request *aead_req;
+	struct sk_buff *unused;
+	u8 *aad, *iv, *mem = NULL;
+	struct scatterlist *sgin = NULL;
+	struct scatterlist *sgout = NULL;
+	const int data_len = rxm->full_len - prot->overhead_size +
+			     prot->tail_size;
+	int iv_offset = 0;
+
+	if (*zc && (out_iov || out_sg)) {
+		if (out_iov)
+			n_sgout = iov_iter_npages(out_iov, INT_MAX) + 1;
+		else
+			n_sgout = sg_nents(out_sg);
+		n_sgin = skb_nsg(skb, rxm->offset + prot->prepend_size,
+				 rxm->full_len - prot->prepend_size);
+	} else {
+		n_sgout = 0;
+		*zc = false;
+		n_sgin = skb_cow_data(skb, 0, &unused);
+	}
+
+	if (n_sgin < 1)
+		return -EBADMSG;
+
+	/* Increment to accommodate AAD */
+	n_sgin = n_sgin + 1;
+
+	nsg = n_sgin + n_sgout;
+
+	aead_size = sizeof(*aead_req) + crypto_aead_reqsize(ctx->aead_recv);
+	mem_size = aead_size + (nsg * sizeof(struct scatterlist));
+	mem_size = mem_size + prot->aad_size;
+	mem_size = mem_size + crypto_aead_ivsize(ctx->aead_recv);
+
+	/* Allocate a single block of memory which contains
+	 * aead_req || sgin[] || sgout[] || aad || iv.
+	 * This order achieves correct alignment for aead_req, sgin, sgout.
+	 */
+	mem = kmalloc(mem_size, sk->sk_allocation);
+	if (!mem)
+		return -ENOMEM;
+
+	/* Segment the allocated memory */
+	aead_req = (struct aead_request *)mem;
+	sgin = (struct scatterlist *)(mem + aead_size);
+	sgout = sgin + n_sgin;
+	aad = (u8 *)(sgout + n_sgout);
+	iv = aad + prot->aad_size;
+
+	/* For CCM based ciphers, first byte of nonce+iv is always '2' */
+	if (prot->cipher_type == TLS_CIPHER_AES_CCM_128) {
+		iv[0] = 2;
+		iv_offset = 1;
+	}
+
+	/* Prepare IV */
+	err = skb_copy_bits(skb, rxm->offset + TLS_HEADER_SIZE,
+			    iv + iv_offset + prot->salt_size,
+			    prot->iv_size);
+	if (err < 0) {
+		kfree(mem);
+		return err;
+	}
+	if (prot->version == TLS_1_3_VERSION)
+		memcpy(iv + iv_offset, tls_ctx->rx.iv,
+		       prot->iv_size + prot->salt_size);
+	else
+		memcpy(iv + iv_offset, tls_ctx->rx.iv, prot->salt_size);
+
+	xor_iv_with_seq(prot->version, iv + iv_offset, tls_ctx->rx.rec_seq);
+
+	/* Prepare AAD */
+	tls_make_aad(aad, rxm->full_len - prot->overhead_size +
+		     prot->tail_size,
+		     tls_ctx->rx.rec_seq, prot->rec_seq_size,
+		     ctx->control, prot->version);
+
+	/* Prepare sgin */
+	sg_init_table(sgin, n_sgin);
+	sg_set_buf(&sgin[0], aad, prot->aad_size);
+	err = skb_to_sgvec(skb, &sgin[1],
+			   rxm->offset + prot->prepend_size,
+			   rxm->full_len - prot->prepend_size);
+	if (err < 0) {
+		kfree(mem);
+		return err;
+	}
+
+	if (n_sgout) {
+		if (out_iov) {
+			sg_init_table(sgout, n_sgout);
+			sg_set_buf(&sgout[0], aad, prot->aad_size);
+
+			*chunk = 0;
+			err = tls_setup_from_iter(sk, out_iov, data_len,
+						  &pages, chunk, &sgout[1],
+						  (n_sgout - 1));
+			if (err < 0)
+				goto fallback_to_reg_recv;
+		} else if (out_sg) {
+			memcpy(sgout, out_sg, n_sgout * sizeof(*sgout));
+		} else {
+			goto fallback_to_reg_recv;
+		}
+	} else {
+fallback_to_reg_recv:
+		sgout = sgin;
+		pages = 0;
+		*chunk = data_len;
+		*zc = false;
+	}
+
+	/* Prepare and submit AEAD request */
+	err = tls_do_decryption(sk, skb, sgin, sgout, iv,
+				data_len, aead_req, async);
+	if (err == -EINPROGRESS)
+		return err;
+
+	/* Release the pages in case iov was mapped to pages */
+	for (; pages > 0; pages--)
+		put_page(sg_page(&sgout[pages]));
+
+	kfree(mem);
+	return err;
+}
+
+static int decrypt_skb_update(struct sock *sk, struct sk_buff *skb,
+			      struct iov_iter *dest, int *chunk, bool *zc,
+			      bool async)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct strp_msg *rxm = strp_msg(skb);
+	int pad, err = 0;
+
+	if (!ctx->decrypted) {
+		if (tls_ctx->rx_conf == TLS_HW) {
+			err = tls_device_decrypted(sk, tls_ctx, skb, rxm);
+			if (err < 0)
+				return err;
+		}
+
+		/* Still not decrypted after tls_device */
+		if (!ctx->decrypted) {
+			err = decrypt_internal(sk, skb, dest, NULL, chunk, zc,
+					       async);
+			if (err < 0) {
+				if (err == -EINPROGRESS)
+					tls_advance_record_sn(sk, prot,
+							      &tls_ctx->rx);
+				else if (err == -EBADMSG)
+					TLS_INC_STATS(sock_net(sk),
+						      LINUX_MIB_TLSDECRYPTERROR);
+				return err;
+			}
+		} else {
+			*zc = false;
+		}
+
+		pad = padding_length(ctx, prot, skb);
+		if (pad < 0)
+			return pad;
+
+		rxm->full_len -= pad;
+		rxm->offset += prot->prepend_size;
+		rxm->full_len -= prot->overhead_size;
+		tls_advance_record_sn(sk, prot, &tls_ctx->rx);
+		ctx->decrypted = 1;
+		ctx->saved_data_ready(sk);
+	} else {
+		*zc = false;
+	}
+
+	return err;
+}
+
+int decrypt_skb(struct sock *sk, struct sk_buff *skb,
+		struct scatterlist *sgout)
+{
+	bool zc = true;
+	int chunk;
+
+	return decrypt_internal(sk, skb, NULL, sgout, &chunk, &zc, false);
+}
+
+static bool tls_sw_advance_skb(struct sock *sk, struct sk_buff *skb,
+			       unsigned int len)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+
+	if (skb) {
+		struct strp_msg *rxm = strp_msg(skb);
+
+		if (len < rxm->full_len) {
+			rxm->offset += len;
+			rxm->full_len -= len;
+			return false;
+		}
+		consume_skb(skb);
+	}
+
+	/* Finished with message */
+	ctx->recv_pkt = NULL;
+	__strp_unpause(&ctx->strp);
+
+	return true;
+}
+
+/* This function traverses the rx_list in tls receive context to copies the
+ * decrypted records into the buffer provided by caller zero copy is not
+ * true. Further, the records are removed from the rx_list if it is not a peek
+ * case and the record has been consumed completely.
+ */
+static int process_rx_list(struct tls_sw_context_rx *ctx,
+			   struct msghdr *msg,
+			   u8 *control,
+			   bool *cmsg,
+			   size_t skip,
+			   size_t len,
+			   bool zc,
+			   bool is_peek)
+{
+	struct sk_buff *skb = skb_peek(&ctx->rx_list);
+	u8 ctrl = *control;
+	u8 msgc = *cmsg;
+	struct tls_msg *tlm;
+	ssize_t copied = 0;
+
+	/* Set the record type in 'control' if caller didn't pass it */
+	if (!ctrl && skb) {
+		tlm = tls_msg(skb);
+		ctrl = tlm->control;
+	}
+
+	while (skip && skb) {
+		struct strp_msg *rxm = strp_msg(skb);
+		tlm = tls_msg(skb);
+
+		/* Cannot process a record of different type */
+		if (ctrl != tlm->control)
+			return 0;
+
+		if (skip < rxm->full_len)
+			break;
+
+		skip = skip - rxm->full_len;
+		skb = skb_peek_next(skb, &ctx->rx_list);
+	}
+
+	while (len && skb) {
+		struct sk_buff *next_skb;
+		struct strp_msg *rxm = strp_msg(skb);
+		int chunk = min_t(unsigned int, rxm->full_len - skip, len);
+
+		tlm = tls_msg(skb);
+
+		/* Cannot process a record of different type */
+		if (ctrl != tlm->control)
+			return 0;
+
+		/* Set record type if not already done. For a non-data record,
+		 * do not proceed if record type could not be copied.
+		 */
+		if (!msgc) {
+			int cerr = put_cmsg(msg, SOL_TLS, TLS_GET_RECORD_TYPE,
+					    sizeof(ctrl), &ctrl);
+			msgc = true;
+			if (ctrl != TLS_RECORD_TYPE_DATA) {
+				if (cerr || msg->msg_flags & MSG_CTRUNC)
+					return -EIO;
+
+				*cmsg = msgc;
+			}
+		}
+
+		if (!zc || (rxm->full_len - skip) > len) {
+			int err = skb_copy_datagram_msg(skb, rxm->offset + skip,
+						    msg, chunk);
+			if (err < 0)
+				return err;
+		}
+
+		len = len - chunk;
+		copied = copied + chunk;
+
+		/* Consume the data from record if it is non-peek case*/
+		if (!is_peek) {
+			rxm->offset = rxm->offset + chunk;
+			rxm->full_len = rxm->full_len - chunk;
+
+			/* Return if there is unconsumed data in the record */
+			if (rxm->full_len - skip)
+				break;
+		}
+
+		/* The remaining skip-bytes must lie in 1st record in rx_list.
+		 * So from the 2nd record, 'skip' should be 0.
+		 */
+		skip = 0;
+
+		if (msg)
+			msg->msg_flags |= MSG_EOR;
+
+		next_skb = skb_peek_next(skb, &ctx->rx_list);
+
+		if (!is_peek) {
+			skb_unlink(skb, &ctx->rx_list);
+			consume_skb(skb);
+		}
+
+		skb = next_skb;
+	}
+
+	*control = ctrl;
+	return copied;
+}
+
+int tls_sw_recvmsg(struct sock *sk,
+		   struct msghdr *msg,
+		   size_t len,
+		   int nonblock,
+		   int flags,
+		   int *addr_len)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct sk_psock *psock;
+	unsigned char control = 0;
+	ssize_t decrypted = 0;
+	struct strp_msg *rxm;
+	struct tls_msg *tlm;
+	struct sk_buff *skb;
+	ssize_t copied = 0;
+	bool cmsg = false;
+	int target, err = 0;
+	long timeo;
+	bool is_kvec = iov_iter_is_kvec(&msg->msg_iter);
+	bool is_peek = flags & MSG_PEEK;
+	bool bpf_strp_enabled;
+	int num_async = 0;
+	int pending;
+
+	flags |= nonblock;
+
+	if (unlikely(flags & MSG_ERRQUEUE))
+		return sock_recv_errqueue(sk, msg, len, SOL_IP, IP_RECVERR);
+
+	psock = sk_psock_get(sk);
+	lock_sock(sk);
+	bpf_strp_enabled = sk_psock_strp_enabled(psock);
+
+	/* Process pending decrypted records. It must be non-zero-copy */
+	err = process_rx_list(ctx, msg, &control, &cmsg, 0, len, false,
+			      is_peek);
+	if (err < 0) {
+		tls_err_abort(sk, err);
+		goto end;
+	} else {
+		copied = err;
+	}
+
+	if (len <= copied)
+		goto recv_end;
+
+	target = sock_rcvlowat(sk, flags & MSG_WAITALL, len);
+	len = len - copied;
+	timeo = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
+
+	while (len && (decrypted + copied < target || ctx->recv_pkt)) {
+		bool retain_skb = false;
+		bool zc = false;
+		int to_decrypt;
+		int chunk = 0;
+		bool async_capable;
+		bool async = false;
+
+		skb = tls_wait_data(sk, psock, flags & MSG_DONTWAIT, timeo, &err);
+		if (!skb) {
+			if (psock) {
+				int ret = __tcp_bpf_recvmsg(sk, psock,
+							    msg, len, flags);
+
+				if (ret > 0) {
+					decrypted += ret;
+					len -= ret;
+					continue;
+				}
+			}
+			goto recv_end;
+		} else {
+			tlm = tls_msg(skb);
+			if (prot->version == TLS_1_3_VERSION)
+				tlm->control = 0;
+			else
+				tlm->control = ctx->control;
+		}
+
+		rxm = strp_msg(skb);
+
+		to_decrypt = rxm->full_len - prot->overhead_size;
+
+		if (to_decrypt <= len && !is_kvec && !is_peek &&
+		    ctx->control == TLS_RECORD_TYPE_DATA &&
+		    prot->version != TLS_1_3_VERSION &&
+		    !bpf_strp_enabled)
+			zc = true;
+
+		/* Do not use async mode if record is non-data */
+		if (ctx->control == TLS_RECORD_TYPE_DATA && !bpf_strp_enabled)
+			async_capable = ctx->async_capable;
+		else
+			async_capable = false;
+
+		err = decrypt_skb_update(sk, skb, &msg->msg_iter,
+					 &chunk, &zc, async_capable);
+		if (err < 0 && err != -EINPROGRESS) {
+			tls_err_abort(sk, -EBADMSG);
+			goto recv_end;
+		}
+
+		if (err == -EINPROGRESS) {
+			async = true;
+			num_async++;
+		} else if (prot->version == TLS_1_3_VERSION) {
+			tlm->control = ctx->control;
+		}
+
+		/* If the type of records being processed is not known yet,
+		 * set it to record type just dequeued. If it is already known,
+		 * but does not match the record type just dequeued, go to end.
+		 * We always get record type here since for tls1.2, record type
+		 * is known just after record is dequeued from stream parser.
+		 * For tls1.3, we disable async.
+		 */
+
+		if (!control)
+			control = tlm->control;
+		else if (control != tlm->control)
+			goto recv_end;
+
+		if (!cmsg) {
+			int cerr;
+
+			cerr = put_cmsg(msg, SOL_TLS, TLS_GET_RECORD_TYPE,
+					sizeof(control), &control);
+			cmsg = true;
+			if (control != TLS_RECORD_TYPE_DATA) {
+				if (cerr || msg->msg_flags & MSG_CTRUNC) {
+					err = -EIO;
+					goto recv_end;
+				}
+			}
+		}
+
+		if (async)
+			goto pick_next_record;
+
+		if (!zc) {
+			if (bpf_strp_enabled) {
+				err = sk_psock_tls_strp_read(psock, skb);
+				if (err != __SK_PASS) {
+					rxm->offset = rxm->offset + rxm->full_len;
+					rxm->full_len = 0;
+					if (err == __SK_DROP)
+						consume_skb(skb);
+					ctx->recv_pkt = NULL;
+					__strp_unpause(&ctx->strp);
+					continue;
+				}
+			}
+
+			if (rxm->full_len > len) {
+				retain_skb = true;
+				chunk = len;
+			} else {
+				chunk = rxm->full_len;
+			}
+
+			err = skb_copy_datagram_msg(skb, rxm->offset,
+						    msg, chunk);
+			if (err < 0)
+				goto recv_end;
+
+			if (!is_peek) {
+				rxm->offset = rxm->offset + chunk;
+				rxm->full_len = rxm->full_len - chunk;
+			}
+		}
+
+pick_next_record:
+		if (chunk > len)
+			chunk = len;
+
+		decrypted += chunk;
+		len -= chunk;
+
+		/* For async or peek case, queue the current skb */
+		if (async || is_peek || retain_skb) {
+			skb_queue_tail(&ctx->rx_list, skb);
+			skb = NULL;
+		}
+
+		if (tls_sw_advance_skb(sk, skb, chunk)) {
+			/* Return full control message to
+			 * userspace before trying to parse
+			 * another message type
+			 */
+			msg->msg_flags |= MSG_EOR;
+			if (control != TLS_RECORD_TYPE_DATA)
+				goto recv_end;
+		} else {
+			break;
+		}
+	}
+
+recv_end:
+	if (num_async) {
+		/* Wait for all previously submitted records to be decrypted */
+		spin_lock_bh(&ctx->decrypt_compl_lock);
+		ctx->async_notify = true;
+		pending = atomic_read(&ctx->decrypt_pending);
+		spin_unlock_bh(&ctx->decrypt_compl_lock);
+		if (pending) {
+			err = crypto_wait_req(-EINPROGRESS, &ctx->async_wait);
+			if (err) {
+				/* one of async decrypt failed */
+				tls_err_abort(sk, err);
+				copied = 0;
+				decrypted = 0;
+				goto end;
+			}
+		} else {
+			reinit_completion(&ctx->async_wait.completion);
+		}
+
+		/* There can be no concurrent accesses, since we have no
+		 * pending decrypt operations
+		 */
+		WRITE_ONCE(ctx->async_notify, false);
+
+		/* Drain records from the rx_list & copy if required */
+		if (is_peek || is_kvec)
+			err = process_rx_list(ctx, msg, &control, &cmsg, copied,
+					      decrypted, false, is_peek);
+		else
+			err = process_rx_list(ctx, msg, &control, &cmsg, 0,
+					      decrypted, true, is_peek);
+		if (err < 0) {
+			tls_err_abort(sk, err);
+			copied = 0;
+			goto end;
+		}
+	}
+
+	copied += decrypted;
+
+end:
+	release_sock(sk);
+	if (psock)
+		sk_psock_put(sk, psock);
+	return copied ? : err;
+}
+
+ssize_t tls_sw_splice_read(struct socket *sock,  loff_t *ppos,
+			   struct pipe_inode_info *pipe,
+			   size_t len, unsigned int flags)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sock->sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+	struct strp_msg *rxm = NULL;
+	struct sock *sk = sock->sk;
+	struct sk_buff *skb;
+	ssize_t copied = 0;
+	int err = 0;
+	long timeo;
+	int chunk;
+	bool zc = false;
+
+	lock_sock(sk);
+
+	timeo = sock_rcvtimeo(sk, flags & SPLICE_F_NONBLOCK);
+
+	skb = tls_wait_data(sk, NULL, flags & SPLICE_F_NONBLOCK, timeo, &err);
+	if (!skb)
+		goto splice_read_end;
+
+	err = decrypt_skb_update(sk, skb, NULL, &chunk, &zc, false);
+	if (err < 0) {
+		tls_err_abort(sk, -EBADMSG);
+		goto splice_read_end;
+	}
+
+	/* splice does not support reading control messages */
+	if (ctx->control != TLS_RECORD_TYPE_DATA) {
+		err = -EINVAL;
+		goto splice_read_end;
+	}
+
+	rxm = strp_msg(skb);
+
+	chunk = min_t(unsigned int, rxm->full_len, len);
+	copied = skb_splice_bits(skb, sk, rxm->offset, pipe, chunk, flags);
+	if (copied < 0)
+		goto splice_read_end;
+
+	if (likely(!(flags & MSG_PEEK)))
+		tls_sw_advance_skb(sk, skb, copied);
+
+splice_read_end:
+	release_sock(sk);
+	return copied ? : err;
+}
+
+bool tls_sw_stream_read(const struct sock *sk)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+	bool ingress_empty = true;
+	struct sk_psock *psock;
+
+	rcu_read_lock();
+	psock = sk_psock(sk);
+	if (psock)
+		ingress_empty = list_empty(&psock->ingress_msg);
+	rcu_read_unlock();
+
+	return !ingress_empty || ctx->recv_pkt ||
+		!skb_queue_empty(&ctx->rx_list);
+}
+
+static int tls_read_size(struct strparser *strp, struct sk_buff *skb)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(strp->sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	char header[TLS_HEADER_SIZE + MAX_IV_SIZE];
+	struct strp_msg *rxm = strp_msg(skb);
+	size_t cipher_overhead;
+	size_t data_len = 0;
+	int ret;
+
+	/* Verify that we have a full TLS header, or wait for more data */
+	if (rxm->offset + prot->prepend_size > skb->len)
+		return 0;
+
+	/* Sanity-check size of on-stack buffer. */
+	if (WARN_ON(prot->prepend_size > sizeof(header))) {
+		ret = -EINVAL;
+		goto read_failure;
+	}
+
+	/* Linearize header to local buffer */
+	ret = skb_copy_bits(skb, rxm->offset, header, prot->prepend_size);
+
+	if (ret < 0)
+		goto read_failure;
+
+	ctx->control = header[0];
+
+	data_len = ((header[4] & 0xFF) | (header[3] << 8));
+
+	cipher_overhead = prot->tag_size;
+	if (prot->version != TLS_1_3_VERSION)
+		cipher_overhead += prot->iv_size;
+
+	if (data_len > TLS_MAX_PAYLOAD_SIZE + cipher_overhead +
+	    prot->tail_size) {
+		ret = -EMSGSIZE;
+		goto read_failure;
+	}
+	if (data_len < cipher_overhead) {
+		ret = -EBADMSG;
+		goto read_failure;
+	}
+
+	/* Note that both TLS1.3 and TLS1.2 use TLS_1_2 version here */
+	if (header[1] != TLS_1_2_VERSION_MINOR ||
+	    header[2] != TLS_1_2_VERSION_MAJOR) {
+		ret = -EINVAL;
+		goto read_failure;
+	}
+
+	tls_device_rx_resync_new_rec(strp->sk, data_len + TLS_HEADER_SIZE,
+				     TCP_SKB_CB(skb)->seq + rxm->offset);
+	return data_len + TLS_HEADER_SIZE;
+
+read_failure:
+	tls_err_abort(strp->sk, ret);
+
+	return ret;
+}
+
+static void tls_queue(struct strparser *strp, struct sk_buff *skb)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(strp->sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+
+	ctx->decrypted = 0;
+
+	ctx->recv_pkt = skb;
+	strp_pause(strp);
+
+	ctx->saved_data_ready(strp->sk);
+}
+
+static void tls_data_ready(struct sock *sk)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+	struct sk_psock *psock;
+
+	strp_data_ready(&ctx->strp);
+
+	psock = sk_psock_get(sk);
+	if (psock) {
+		if (!list_empty(&psock->ingress_msg))
+			ctx->saved_data_ready(sk);
+		sk_psock_put(sk, psock);
+	}
+}
+
+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);
+	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_tx_records(sk, -1);
+
+	/* Free up un-sent records in tx_list. First, free
+	 * the partially sent record if any at head of tx_list.
+	 */
+	if (tls_ctx->partially_sent_record) {
+		tls_free_partial_record(sk, tls_ctx);
+		rec = list_first_entry(&ctx->tx_list,
+				       struct tls_rec, list);
+		list_del(&rec->list);
+		sk_msg_free(sk, &rec->msg_plaintext);
+		kfree(rec);
+	}
+
+	list_for_each_entry_safe(rec, tmp, &ctx->tx_list, list) {
+		list_del(&rec->list);
+		sk_msg_free(sk, &rec->msg_encrypted);
+		sk_msg_free(sk, &rec->msg_plaintext);
+		kfree(rec);
+	}
+
+	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);
+}
+
+void tls_sw_release_resources_rx(struct sock *sk)
+{
+	struct tls_context *tls_ctx = tls_get_ctx(sk);
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+
+	kfree(tls_ctx->rx.rec_seq);
+	kfree(tls_ctx->rx.iv);
+
+	if (ctx->aead_recv) {
+		kfree_skb(ctx->recv_pkt);
+		ctx->recv_pkt = NULL;
+		skb_queue_purge(&ctx->rx_list);
+		crypto_free_aead(ctx->aead_recv);
+		strp_stop(&ctx->strp);
+		/* 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_strparser_done(struct tls_context *tls_ctx)
+{
+	struct tls_sw_context_rx *ctx = tls_sw_ctx_rx(tls_ctx);
+
+	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)
+{
+	struct delayed_work *delayed_work = to_delayed_work(work);
+	struct tx_work *tx_work = container_of(delayed_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;
+
+	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;
+
+	if (mutex_trylock(&tls_ctx->tx_lock)) {
+		lock_sock(sk);
+		tls_tx_records(sk, -1);
+		release_sock(sk);
+		mutex_unlock(&tls_ctx->tx_lock);
+	} else if (!test_and_set_bit(BIT_TX_SCHEDULED, &ctx->tx_bitmask)) {
+		/* Someone is holding the tx_lock, they will likely run Tx
+		 * and cancel the work on their way out of the lock section.
+		 * Schedule a long delay just in case.
+		 */
+		schedule_delayed_work(&ctx->tx_work.work, msecs_to_jiffies(10));
+	}
+}
+
+void tls_sw_write_space(struct sock *sk, struct tls_context *ctx)
+{
+	struct tls_sw_context_tx *tx_ctx = tls_sw_ctx_tx(ctx);
+
+	/* Schedule the transmission if tx list is ready */
+	if (is_tx_ready(tx_ctx) &&
+	    !test_and_set_bit(BIT_TX_SCHEDULED, &tx_ctx->tx_bitmask))
+		schedule_delayed_work(&tx_ctx->tx_work.work, 0);
+}
+
+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);
+	struct tls_prot_info *prot = &tls_ctx->prot_info;
+	struct tls_crypto_info *crypto_info;
+	struct tls12_crypto_info_aes_gcm_128 *gcm_128_info;
+	struct tls12_crypto_info_aes_gcm_256 *gcm_256_info;
+	struct tls12_crypto_info_aes_ccm_128 *ccm_128_info;
+	struct tls_sw_context_tx *sw_ctx_tx = NULL;
+	struct tls_sw_context_rx *sw_ctx_rx = NULL;
+	struct cipher_context *cctx;
+	struct crypto_aead **aead;
+	struct strp_callbacks cb;
+	u16 nonce_size, tag_size, iv_size, rec_seq_size, salt_size;
+	struct crypto_tfm *tfm;
+	char *iv, *rec_seq, *key, *salt, *cipher_name;
+	size_t keysize;
+	int rc = 0;
+
+	if (!ctx) {
+		rc = -EINVAL;
+		goto out;
+	}
+
+	if (tx) {
+		if (!ctx->priv_ctx_tx) {
+			sw_ctx_tx = kzalloc(sizeof(*sw_ctx_tx), GFP_KERNEL);
+			if (!sw_ctx_tx) {
+				rc = -ENOMEM;
+				goto out;
+			}
+			ctx->priv_ctx_tx = sw_ctx_tx;
+		} else {
+			sw_ctx_tx =
+				(struct tls_sw_context_tx *)ctx->priv_ctx_tx;
+		}
+	} else {
+		if (!ctx->priv_ctx_rx) {
+			sw_ctx_rx = kzalloc(sizeof(*sw_ctx_rx), GFP_KERNEL);
+			if (!sw_ctx_rx) {
+				rc = -ENOMEM;
+				goto out;
+			}
+			ctx->priv_ctx_rx = sw_ctx_rx;
+		} else {
+			sw_ctx_rx =
+				(struct tls_sw_context_rx *)ctx->priv_ctx_rx;
+		}
+	}
+
+	if (tx) {
+		crypto_init_wait(&sw_ctx_tx->async_wait);
+		spin_lock_init(&sw_ctx_tx->encrypt_compl_lock);
+		crypto_info = &ctx->crypto_send.info;
+		cctx = &ctx->tx;
+		aead = &sw_ctx_tx->aead_send;
+		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;
+	} else {
+		crypto_init_wait(&sw_ctx_rx->async_wait);
+		spin_lock_init(&sw_ctx_rx->decrypt_compl_lock);
+		crypto_info = &ctx->crypto_recv.info;
+		cctx = &ctx->rx;
+		skb_queue_head_init(&sw_ctx_rx->rx_list);
+		aead = &sw_ctx_rx->aead_recv;
+	}
+
+	switch (crypto_info->cipher_type) {
+	case TLS_CIPHER_AES_GCM_128: {
+		nonce_size = TLS_CIPHER_AES_GCM_128_IV_SIZE;
+		tag_size = TLS_CIPHER_AES_GCM_128_TAG_SIZE;
+		iv_size = TLS_CIPHER_AES_GCM_128_IV_SIZE;
+		iv = ((struct tls12_crypto_info_aes_gcm_128 *)crypto_info)->iv;
+		rec_seq_size = TLS_CIPHER_AES_GCM_128_REC_SEQ_SIZE;
+		rec_seq =
+		 ((struct tls12_crypto_info_aes_gcm_128 *)crypto_info)->rec_seq;
+		gcm_128_info =
+			(struct tls12_crypto_info_aes_gcm_128 *)crypto_info;
+		keysize = TLS_CIPHER_AES_GCM_128_KEY_SIZE;
+		key = gcm_128_info->key;
+		salt = gcm_128_info->salt;
+		salt_size = TLS_CIPHER_AES_GCM_128_SALT_SIZE;
+		cipher_name = "gcm(aes)";
+		break;
+	}
+	case TLS_CIPHER_AES_GCM_256: {
+		nonce_size = TLS_CIPHER_AES_GCM_256_IV_SIZE;
+		tag_size = TLS_CIPHER_AES_GCM_256_TAG_SIZE;
+		iv_size = TLS_CIPHER_AES_GCM_256_IV_SIZE;
+		iv = ((struct tls12_crypto_info_aes_gcm_256 *)crypto_info)->iv;
+		rec_seq_size = TLS_CIPHER_AES_GCM_256_REC_SEQ_SIZE;
+		rec_seq =
+		 ((struct tls12_crypto_info_aes_gcm_256 *)crypto_info)->rec_seq;
+		gcm_256_info =
+			(struct tls12_crypto_info_aes_gcm_256 *)crypto_info;
+		keysize = TLS_CIPHER_AES_GCM_256_KEY_SIZE;
+		key = gcm_256_info->key;
+		salt = gcm_256_info->salt;
+		salt_size = TLS_CIPHER_AES_GCM_256_SALT_SIZE;
+		cipher_name = "gcm(aes)";
+		break;
+	}
+	case TLS_CIPHER_AES_CCM_128: {
+		nonce_size = TLS_CIPHER_AES_CCM_128_IV_SIZE;
+		tag_size = TLS_CIPHER_AES_CCM_128_TAG_SIZE;
+		iv_size = TLS_CIPHER_AES_CCM_128_IV_SIZE;
+		iv = ((struct tls12_crypto_info_aes_ccm_128 *)crypto_info)->iv;
+		rec_seq_size = TLS_CIPHER_AES_CCM_128_REC_SEQ_SIZE;
+		rec_seq =
+		((struct tls12_crypto_info_aes_ccm_128 *)crypto_info)->rec_seq;
+		ccm_128_info =
+		(struct tls12_crypto_info_aes_ccm_128 *)crypto_info;
+		keysize = TLS_CIPHER_AES_CCM_128_KEY_SIZE;
+		key = ccm_128_info->key;
+		salt = ccm_128_info->salt;
+		salt_size = TLS_CIPHER_AES_CCM_128_SALT_SIZE;
+		cipher_name = "ccm(aes)";
+		break;
+	}
+	default:
+		rc = -EINVAL;
+		goto free_priv;
+	}
+
+	/* Sanity-check the sizes for stack allocations. */
+	if (iv_size > MAX_IV_SIZE || nonce_size > MAX_IV_SIZE ||
+	    rec_seq_size > TLS_MAX_REC_SEQ_SIZE) {
+		rc = -EINVAL;
+		goto free_priv;
+	}
+
+	if (crypto_info->version == TLS_1_3_VERSION) {
+		nonce_size = 0;
+		prot->aad_size = TLS_HEADER_SIZE;
+		prot->tail_size = 1;
+	} else {
+		prot->aad_size = TLS_AAD_SPACE_SIZE;
+		prot->tail_size = 0;
+	}
+
+	prot->version = crypto_info->version;
+	prot->cipher_type = crypto_info->cipher_type;
+	prot->prepend_size = TLS_HEADER_SIZE + nonce_size;
+	prot->tag_size = tag_size;
+	prot->overhead_size = prot->prepend_size +
+			      prot->tag_size + prot->tail_size;
+	prot->iv_size = iv_size;
+	prot->salt_size = salt_size;
+	cctx->iv = kmalloc(iv_size + salt_size, GFP_KERNEL);
+	if (!cctx->iv) {
+		rc = -ENOMEM;
+		goto free_priv;
+	}
+	/* Note: 128 & 256 bit salt are the same size */
+	prot->rec_seq_size = rec_seq_size;
+	memcpy(cctx->iv, salt, salt_size);
+	memcpy(cctx->iv + salt_size, iv, iv_size);
+	cctx->rec_seq = kmemdup(rec_seq, rec_seq_size, GFP_KERNEL);
+	if (!cctx->rec_seq) {
+		rc = -ENOMEM;
+		goto free_iv;
+	}
+
+	if (!*aead) {
+		*aead = crypto_alloc_aead(cipher_name, 0, 0);
+		if (IS_ERR(*aead)) {
+			rc = PTR_ERR(*aead);
+			*aead = NULL;
+			goto free_rec_seq;
+		}
+	}
+
+	ctx->push_pending_record = tls_sw_push_pending_record;
+
+	rc = crypto_aead_setkey(*aead, key, keysize);
+
+	if (rc)
+		goto free_aead;
+
+	rc = crypto_aead_setauthsize(*aead, prot->tag_size);
+	if (rc)
+		goto free_aead;
+
+	if (sw_ctx_rx) {
+		tfm = crypto_aead_tfm(sw_ctx_rx->aead_recv);
+
+		if (crypto_info->version == TLS_1_3_VERSION)
+			sw_ctx_rx->async_capable = 0;
+		else
+			sw_ctx_rx->async_capable =
+				!!(tfm->__crt_alg->cra_flags &
+				   CRYPTO_ALG_ASYNC);
+
+		/* Set up strparser */
+		memset(&cb, 0, sizeof(cb));
+		cb.rcv_msg = tls_queue;
+		cb.parse_msg = tls_read_size;
+
+		strp_init(&sw_ctx_rx->strp, sk, &cb);
+	}
+
+	goto out;
+
+free_aead:
+	crypto_free_aead(*aead);
+	*aead = NULL;
+free_rec_seq:
+	kfree(cctx->rec_seq);
+	cctx->rec_seq = NULL;
+free_iv:
+	kfree(cctx->iv);
+	cctx->iv = NULL;
+free_priv:
+	if (tx) {
+		kfree(ctx->priv_ctx_tx);
+		ctx->priv_ctx_tx = NULL;
+	} else {
+		kfree(ctx->priv_ctx_rx);
+		ctx->priv_ctx_rx = NULL;
+	}
+out:
+	return rc;
+}
diff --git a/net/tls/tls_toe.c b/net/tls/tls_toe.c
new file mode 100644
index 000000000..7e1330f19
--- /dev/null
+++ b/net/tls/tls_toe.c
@@ -0,0 +1,139 @@
+/*
+ * Copyright (c) 2016-2017, Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2016-2017, Dave Watson <davejwatson@fb.com>. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/list.h>
+#include <linux/rcupdate.h>
+#include <linux/spinlock.h>
+#include <net/inet_connection_sock.h>
+#include <net/tls.h>
+#include <net/tls_toe.h>
+
+static LIST_HEAD(device_list);
+static DEFINE_SPINLOCK(device_spinlock);
+
+static void tls_toe_sk_destruct(struct sock *sk)
+{
+	struct inet_connection_sock *icsk = inet_csk(sk);
+	struct tls_context *ctx = tls_get_ctx(sk);
+
+	ctx->sk_destruct(sk);
+	/* Free ctx */
+	rcu_assign_pointer(icsk->icsk_ulp_data, NULL);
+	tls_ctx_free(sk, ctx);
+}
+
+int tls_toe_bypass(struct sock *sk)
+{
+	struct tls_toe_device *dev;
+	struct tls_context *ctx;
+	int rc = 0;
+
+	spin_lock_bh(&device_spinlock);
+	list_for_each_entry(dev, &device_list, dev_list) {
+		if (dev->feature && dev->feature(dev)) {
+			ctx = tls_ctx_create(sk);
+			if (!ctx)
+				goto out;
+
+			ctx->sk_destruct = sk->sk_destruct;
+			sk->sk_destruct = tls_toe_sk_destruct;
+			ctx->rx_conf = TLS_HW_RECORD;
+			ctx->tx_conf = TLS_HW_RECORD;
+			update_sk_prot(sk, ctx);
+			rc = 1;
+			break;
+		}
+	}
+out:
+	spin_unlock_bh(&device_spinlock);
+	return rc;
+}
+
+void tls_toe_unhash(struct sock *sk)
+{
+	struct tls_context *ctx = tls_get_ctx(sk);
+	struct tls_toe_device *dev;
+
+	spin_lock_bh(&device_spinlock);
+	list_for_each_entry(dev, &device_list, dev_list) {
+		if (dev->unhash) {
+			kref_get(&dev->kref);
+			spin_unlock_bh(&device_spinlock);
+			dev->unhash(dev, sk);
+			kref_put(&dev->kref, dev->release);
+			spin_lock_bh(&device_spinlock);
+		}
+	}
+	spin_unlock_bh(&device_spinlock);
+	ctx->sk_proto->unhash(sk);
+}
+
+int tls_toe_hash(struct sock *sk)
+{
+	struct tls_context *ctx = tls_get_ctx(sk);
+	struct tls_toe_device *dev;
+	int err;
+
+	err = ctx->sk_proto->hash(sk);
+	spin_lock_bh(&device_spinlock);
+	list_for_each_entry(dev, &device_list, dev_list) {
+		if (dev->hash) {
+			kref_get(&dev->kref);
+			spin_unlock_bh(&device_spinlock);
+			err |= dev->hash(dev, sk);
+			kref_put(&dev->kref, dev->release);
+			spin_lock_bh(&device_spinlock);
+		}
+	}
+	spin_unlock_bh(&device_spinlock);
+
+	if (err)
+		tls_toe_unhash(sk);
+	return err;
+}
+
+void tls_toe_register_device(struct tls_toe_device *device)
+{
+	spin_lock_bh(&device_spinlock);
+	list_add_tail(&device->dev_list, &device_list);
+	spin_unlock_bh(&device_spinlock);
+}
+EXPORT_SYMBOL(tls_toe_register_device);
+
+void tls_toe_unregister_device(struct tls_toe_device *device)
+{
+	spin_lock_bh(&device_spinlock);
+	list_del(&device->dev_list);
+	spin_unlock_bh(&device_spinlock);
+}
+EXPORT_SYMBOL(tls_toe_unregister_device);
diff --git a/net/tls/trace.c b/net/tls/trace.c
new file mode 100644
index 000000000..e374913cf
--- /dev/null
+++ b/net/tls/trace.c
@@ -0,0 +1,10 @@
+// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+/* Copyright (C) 2019 Netronome Systems, Inc. */
+
+#include <linux/module.h>
+
+#ifndef __CHECKER__
+#define CREATE_TRACE_POINTS
+#include "trace.h"
+
+#endif
diff --git a/net/tls/trace.h b/net/tls/trace.h
new file mode 100644
index 000000000..9ba5f600e
--- /dev/null
+++ b/net/tls/trace.h
@@ -0,0 +1,202 @@
+/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
+/* Copyright (C) 2019 Netronome Systems, Inc. */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM tls
+
+#if !defined(_TLS_TRACE_H_) || defined(TRACE_HEADER_MULTI_READ)
+#define _TLS_TRACE_H_
+
+#include <asm/unaligned.h>
+#include <linux/tracepoint.h>
+
+struct sock;
+
+TRACE_EVENT(tls_device_offload_set,
+
+	TP_PROTO(struct sock *sk, int dir, u32 tcp_seq, u8 *rec_no, int ret),
+
+	TP_ARGS(sk, dir, tcp_seq, rec_no, ret),
+
+	TP_STRUCT__entry(
+		__field(	struct sock *,	sk		)
+		__field(	u64,		rec_no		)
+		__field(	int,		dir		)
+		__field(	u32,		tcp_seq		)
+		__field(	int,		ret		)
+	),
+
+	TP_fast_assign(
+		__entry->sk = sk;
+		__entry->rec_no = get_unaligned_be64(rec_no);
+		__entry->dir = dir;
+		__entry->tcp_seq = tcp_seq;
+		__entry->ret = ret;
+	),
+
+	TP_printk(
+		"sk=%p direction=%d tcp_seq=%u rec_no=%llu ret=%d",
+		__entry->sk, __entry->dir, __entry->tcp_seq, __entry->rec_no,
+		__entry->ret
+	)
+);
+
+TRACE_EVENT(tls_device_decrypted,
+
+	TP_PROTO(struct sock *sk, u32 tcp_seq, u8 *rec_no, u32 rec_len,
+		 bool encrypted, bool decrypted),
+
+	TP_ARGS(sk, tcp_seq, rec_no, rec_len, encrypted, decrypted),
+
+	TP_STRUCT__entry(
+		__field(	struct sock *,	sk		)
+		__field(	u64,		rec_no		)
+		__field(	u32,		tcp_seq		)
+		__field(	u32,		rec_len		)
+		__field(	bool,		encrypted	)
+		__field(	bool,		decrypted	)
+	),
+
+	TP_fast_assign(
+		__entry->sk = sk;
+		__entry->rec_no = get_unaligned_be64(rec_no);
+		__entry->tcp_seq = tcp_seq;
+		__entry->rec_len = rec_len;
+		__entry->encrypted = encrypted;
+		__entry->decrypted = decrypted;
+	),
+
+	TP_printk(
+		"sk=%p tcp_seq=%u rec_no=%llu len=%u encrypted=%d decrypted=%d",
+		__entry->sk, __entry->tcp_seq,
+		__entry->rec_no, __entry->rec_len,
+		__entry->encrypted, __entry->decrypted
+	)
+);
+
+TRACE_EVENT(tls_device_rx_resync_send,
+
+	TP_PROTO(struct sock *sk, u32 tcp_seq, u8 *rec_no, int sync_type),
+
+	TP_ARGS(sk, tcp_seq, rec_no, sync_type),
+
+	TP_STRUCT__entry(
+		__field(	struct sock *,	sk		)
+		__field(	u64,		rec_no		)
+		__field(	u32,		tcp_seq		)
+		__field(	int,		sync_type	)
+	),
+
+	TP_fast_assign(
+		__entry->sk = sk;
+		__entry->rec_no = get_unaligned_be64(rec_no);
+		__entry->tcp_seq = tcp_seq;
+		__entry->sync_type = sync_type;
+	),
+
+	TP_printk(
+		"sk=%p tcp_seq=%u rec_no=%llu sync_type=%d",
+		__entry->sk, __entry->tcp_seq, __entry->rec_no,
+		__entry->sync_type
+	)
+);
+
+TRACE_EVENT(tls_device_rx_resync_nh_schedule,
+
+	TP_PROTO(struct sock *sk),
+
+	TP_ARGS(sk),
+
+	TP_STRUCT__entry(
+		__field(	struct sock *,	sk		)
+	),
+
+	TP_fast_assign(
+		__entry->sk = sk;
+	),
+
+	TP_printk(
+		"sk=%p", __entry->sk
+	)
+);
+
+TRACE_EVENT(tls_device_rx_resync_nh_delay,
+
+	TP_PROTO(struct sock *sk, u32 sock_data, u32 rec_len),
+
+	TP_ARGS(sk, sock_data, rec_len),
+
+	TP_STRUCT__entry(
+		__field(	struct sock *,	sk		)
+		__field(	u32,		sock_data	)
+		__field(	u32,		rec_len		)
+	),
+
+	TP_fast_assign(
+		__entry->sk = sk;
+		__entry->sock_data = sock_data;
+		__entry->rec_len = rec_len;
+	),
+
+	TP_printk(
+		"sk=%p sock_data=%u rec_len=%u",
+		__entry->sk, __entry->sock_data, __entry->rec_len
+	)
+);
+
+TRACE_EVENT(tls_device_tx_resync_req,
+
+	TP_PROTO(struct sock *sk, u32 tcp_seq, u32 exp_tcp_seq),
+
+	TP_ARGS(sk, tcp_seq, exp_tcp_seq),
+
+	TP_STRUCT__entry(
+		__field(	struct sock *,	sk		)
+		__field(	u32,		tcp_seq		)
+		__field(	u32,		exp_tcp_seq	)
+	),
+
+	TP_fast_assign(
+		__entry->sk = sk;
+		__entry->tcp_seq = tcp_seq;
+		__entry->exp_tcp_seq = exp_tcp_seq;
+	),
+
+	TP_printk(
+		"sk=%p tcp_seq=%u exp_tcp_seq=%u",
+		__entry->sk, __entry->tcp_seq, __entry->exp_tcp_seq
+	)
+);
+
+TRACE_EVENT(tls_device_tx_resync_send,
+
+	TP_PROTO(struct sock *sk, u32 tcp_seq, u8 *rec_no),
+
+	TP_ARGS(sk, tcp_seq, rec_no),
+
+	TP_STRUCT__entry(
+		__field(	struct sock *,	sk		)
+		__field(	u64,		rec_no		)
+		__field(	u32,		tcp_seq		)
+	),
+
+	TP_fast_assign(
+		__entry->sk = sk;
+		__entry->rec_no = get_unaligned_be64(rec_no);
+		__entry->tcp_seq = tcp_seq;
+	),
+
+	TP_printk(
+		"sk=%p tcp_seq=%u rec_no=%llu",
+		__entry->sk, __entry->tcp_seq, __entry->rec_no
+	)
+);
+
+#endif /* _TLS_TRACE_H_ */
+
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH .
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE trace
+
+#include <trace/define_trace.h>