Merge branch 'net-tls-small-general-improvements'

Jakub Kicinski says:

====================
net/tls: small general improvements

This series cleans up and improves the tls code, mostly the offload
parts.

First a slight performance optimization - avoiding unnecessary re-
-encryption of records in patch 1.  Next patch 2 makes the code
more resilient by checking for errors in skb_copy_bits().  Next
commit removes a warning which can be triggered in normal operation,
(especially for devices explicitly making use of the fallback path).
Next two paths change the condition checking around the call to
tls_device_decrypted() to make it easier to extend.  Remaining
commits are centered around reorganizing struct tls_context for
better cache utilization.
====================

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

8 files changed, 75 insertions, 65 deletions

diff --git a/Documentation/networking/tls-offload.rst b/Documentation/networking/tls-offload.rst
index cb85af559dff..eb7c9b81ccf5 100644
--- a/Documentation/networking/tls-offload.rst
+++ b/Documentation/networking/tls-offload.rst
@@ -379,7 +379,6 @@ by the driver:
    but did not arrive in the expected order
  * ``tx_tls_drop_no_sync_data`` - number of TX packets dropped because
    they arrived out of order and associated record could not be found
-   (see also :ref:`pre_tls_data`)
 
 Notable corner cases, exceptions and additional requirements
 ============================================================
@@ -462,21 +461,3 @@ Redirects leak clear text
 
 In the RX direction, if segment has already been decrypted by the device
 and it gets redirected or mirrored - clear text will be transmitted out.
-
-.. _pre_tls_data:
-
-Transmission of pre-TLS data
-----------------------------
-
-User can enqueue some already encrypted and framed records before enabling
-``ktls`` on the socket. Those records have to get sent as they are. This is
-perfectly easy to handle in the software case - such data will be waiting
-in the TCP layer, TLS ULP won't see it. In the offloaded case when pre-queued
-segment reaches transmission point it appears to be out of order (before the
-expected TCP sequence number) and the stack does not have a record information
-associated.
-
-All segments without record information cannot, however, be assumed to be
-pre-queued data, because a race condition exists between TCP stack queuing
-a retransmission, the driver seeing the retransmission and TCP ACK arriving
-for the retransmitted data.
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 2ee5e63195c0..98ff5ac98caa 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -1063,6 +1063,7 @@ struct sk_buff *alloc_skb_with_frags(unsigned long header_len,
 				     int max_page_order,
 				     int *errcode,
 				     gfp_t gfp_mask);
+struct sk_buff *alloc_skb_for_msg(struct sk_buff *first);
 
 /* Layout of fast clones : [skb1][skb2][fclone_ref] */
 struct sk_buff_fclones {
diff --git a/include/net/tls.h b/include/net/tls.h
index 39ea62f0c1f6..0a0072636009 100644
--- a/include/net/tls.h
+++ b/include/net/tls.h
@@ -236,34 +236,32 @@ struct tls_prot_info {
 };
 
 struct tls_context {
+	/* read-only cache line */
 	struct tls_prot_info prot_info;
 
-	union tls_crypto_context crypto_send;
-	union tls_crypto_context crypto_recv;
+	u8 tx_conf:3;
+	u8 rx_conf:3;
 
-	struct list_head list;
-	struct net_device *netdev;
-	refcount_t refcount;
+	int (*push_pending_record)(struct sock *sk, int flags);
+	void (*sk_write_space)(struct sock *sk);
 
 	void *priv_ctx_tx;
 	void *priv_ctx_rx;
 
-	u8 tx_conf:3;
-	u8 rx_conf:3;
+	struct net_device *netdev;
 
+	/* rw cache line */
 	struct cipher_context tx;
 	struct cipher_context rx;
 
 	struct scatterlist *partially_sent_record;
 	u16 partially_sent_offset;
 
-	unsigned long flags;
 	bool in_tcp_sendpages;
 	bool pending_open_record_frags;
+	unsigned long flags;
 
-	int (*push_pending_record)(struct sock *sk, int flags);
-
-	void (*sk_write_space)(struct sock *sk);
+	/* cache cold stuff */
 	void (*sk_destruct)(struct sock *sk);
 	void (*sk_proto_close)(struct sock *sk, long timeout);
 
@@ -275,6 +273,12 @@ struct tls_context {
 			   int __user *optlen);
 	int  (*hash)(struct sock *sk);
 	void (*unhash)(struct sock *sk);
+
+	union tls_crypto_context crypto_send;
+	union tls_crypto_context crypto_recv;
+
+	struct list_head list;
+	refcount_t refcount;
 };
 
 enum tls_offload_ctx_dir {
@@ -442,19 +446,15 @@ static inline struct tls_context *tls_get_ctx(const struct sock *sk)
 }
 
 static inline void tls_advance_record_sn(struct sock *sk,
-					 struct cipher_context *ctx,
-					 int version)
+					 struct tls_prot_info *prot,
+					 struct cipher_context *ctx)
 {
-	struct tls_context *tls_ctx = tls_get_ctx(sk);
-	struct tls_prot_info *prot = &tls_ctx->prot_info;
-
 	if (tls_bigint_increment(ctx->rec_seq, prot->rec_seq_size))
 		tls_err_abort(sk, EBADMSG);
 
-	if (version != TLS_1_3_VERSION) {
+	if (prot->version != TLS_1_3_VERSION)
 		tls_bigint_increment(ctx->iv + TLS_CIPHER_AES_GCM_128_SALT_SIZE,
 				     prot->iv_size);
-	}
 }
 
 static inline void tls_fill_prepend(struct tls_context *ctx,
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 4a712a00243a..b50a5e3ac4e4 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -914,6 +914,31 @@ static struct sk_buff *__skb_clone(struct sk_buff *n, struct sk_buff *skb)
 }
 
 /**
+ * alloc_skb_for_msg() - allocate sk_buff to wrap frag list forming a msg
+ * @first: first sk_buff of the msg
+ */
+struct sk_buff *alloc_skb_for_msg(struct sk_buff *first)
+{
+	struct sk_buff *n;
+
+	n = alloc_skb(0, GFP_ATOMIC);
+	if (!n)
+		return NULL;
+
+	n->len = first->len;
+	n->data_len = first->len;
+	n->truesize = first->truesize;
+
+	skb_shinfo(n)->frag_list = first;
+
+	__copy_skb_header(n, first);
+	n->destructor = NULL;
+
+	return n;
+}
+EXPORT_SYMBOL_GPL(alloc_skb_for_msg);
+
+/**
  *	skb_morph	-	morph one skb into another
  *	@dst: the skb to receive the contents
  *	@src: the skb to supply the contents
diff --git a/net/strparser/strparser.c b/net/strparser/strparser.c
index e137698e8aef..3fe541b746b0 100644
--- a/net/strparser/strparser.c
+++ b/net/strparser/strparser.c
@@ -160,18 +160,14 @@ static int __strp_recv(read_descriptor_t *desc, struct sk_buff *orig_skb,
 					return 0;
 				}
 
-				skb = alloc_skb(0, GFP_ATOMIC);
+				skb = alloc_skb_for_msg(head);
 				if (!skb) {
 					STRP_STATS_INCR(strp->stats.mem_fail);
 					desc->error = -ENOMEM;
 					return 0;
 				}
-				skb->len = head->len;
-				skb->data_len = head->len;
-				skb->truesize = head->truesize;
-				*_strp_msg(skb) = *_strp_msg(head);
+
 				strp->skb_nextp = &head->next;
-				skb_shinfo(skb)->frag_list = head;
 				strp->skb_head = skb;
 				head = skb;
 			} else {
diff --git a/net/tls/tls_device.c b/net/tls/tls_device.c
index b95c408fd771..51e556e79371 100644
--- a/net/tls/tls_device.c
+++ b/net/tls/tls_device.c
@@ -252,7 +252,7 @@ static int tls_push_record(struct sock *sk,
 			 skb_frag_address(frag),
 			 record->len - prot->prepend_size,
 			 record_type,
-			 ctx->crypto_send.info.version);
+			 prot->version);
 
 	/* HW doesn't care about the data in the tag, because it fills it. */
 	dummy_tag_frag.page = skb_frag_page(frag);
@@ -264,7 +264,7 @@ static int tls_push_record(struct sock *sk,
 	list_add_tail(&record->list, &offload_ctx->records_list);
 	spin_unlock_irq(&offload_ctx->lock);
 	offload_ctx->open_record = NULL;
-	tls_advance_record_sn(sk, &ctx->tx, ctx->crypto_send.info.version);
+	tls_advance_record_sn(sk, prot, &ctx->tx);
 
 	for (i = 0; i < record->num_frags; i++) {
 		frag = &record->frags[i];
@@ -603,8 +603,10 @@ static int tls_device_reencrypt(struct sock *sk, struct sk_buff *skb)
 	sg_set_buf(&sg[0], buf,
 		   rxm->full_len + TLS_HEADER_SIZE +
 		   TLS_CIPHER_AES_GCM_128_IV_SIZE);
-	skb_copy_bits(skb, offset, buf,
-		      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);
@@ -618,8 +620,11 @@ static int tls_device_reencrypt(struct sock *sk, struct sk_buff *skb)
 	if (skb_pagelen(skb) > offset) {
 		copy = min_t(int, skb_pagelen(skb) - offset, data_len);
 
-		if (skb->decrypted)
-			skb_store_bits(skb, offset, buf, copy);
+		if (skb->decrypted) {
+			err = skb_store_bits(skb, offset, buf, copy);
+			if (err)
+				goto free_buf;
+		}
 
 		offset += copy;
 		buf += copy;
@@ -642,8 +647,11 @@ static int tls_device_reencrypt(struct sock *sk, struct sk_buff *skb)
 		copy = min_t(int, skb_iter->len - frag_pos,
 			     data_len + rxm->offset - offset);
 
-		if (skb_iter->decrypted)
-			skb_store_bits(skb_iter, frag_pos, buf, copy);
+		if (skb_iter->decrypted) {
+			err = skb_store_bits(skb_iter, frag_pos, buf, copy);
+			if (err)
+				goto free_buf;
+		}
 
 		offset += copy;
 		buf += copy;
@@ -664,10 +672,6 @@ int tls_device_decrypted(struct sock *sk, struct sk_buff *skb)
 	int is_encrypted = !is_decrypted;
 	struct sk_buff *skb_iter;
 
-	/* Skip if it is already decrypted */
-	if (ctx->sw.decrypted)
-		return 0;
-
 	/* Check if all the data is decrypted already */
 	skb_walk_frags(skb, skb_iter) {
 		is_decrypted &= skb_iter->decrypted;
diff --git a/net/tls/tls_device_fallback.c b/net/tls/tls_device_fallback.c
index c3a5fe624b4e..5a087e1981c3 100644
--- a/net/tls/tls_device_fallback.c
+++ b/net/tls/tls_device_fallback.c
@@ -240,7 +240,6 @@ static int fill_sg_in(struct scatterlist *sg_in,
 	record = tls_get_record(ctx, tcp_seq, rcd_sn);
 	if (!record) {
 		spin_unlock_irqrestore(&ctx->lock, flags);
-		WARN(1, "Record not found for seq %u\n", tcp_seq);
 		return -EINVAL;
 	}
 
@@ -409,7 +408,10 @@ put_sg:
 		put_page(sg_page(&sg_in[--resync_sgs]));
 	kfree(sg_in);
 free_orig:
-	kfree_skb(skb);
+	if (nskb)
+		consume_skb(skb);
+	else
+		kfree_skb(skb);
 	return nskb;
 }
 
diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c
index 960494f437ac..bef71e54fad0 100644
--- a/net/tls/tls_sw.c
+++ b/net/tls/tls_sw.c
@@ -534,7 +534,7 @@ static int tls_do_encryption(struct sock *sk,
 
 	/* Unhook the record from context if encryption is not failure */
 	ctx->open_rec = NULL;
-	tls_advance_record_sn(sk, &tls_ctx->tx, prot->version);
+	tls_advance_record_sn(sk, prot, &tls_ctx->tx);
 	return rc;
 }
 
@@ -1486,15 +1486,16 @@ static int decrypt_skb_update(struct sock *sk, struct sk_buff *skb,
 	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;
-	int version = prot->version;
 	struct strp_msg *rxm = strp_msg(skb);
 	int pad, err = 0;
 
 	if (!ctx->decrypted) {
 #ifdef CONFIG_TLS_DEVICE
-		err = tls_device_decrypted(sk, skb);
-		if (err < 0)
-			return err;
+		if (tls_ctx->rx_conf == TLS_HW) {
+			err = tls_device_decrypted(sk, skb);
+			if (err < 0)
+				return err;
+		}
 #endif
 		/* Still not decrypted after tls_device */
 		if (!ctx->decrypted) {
@@ -1502,8 +1503,8 @@ static int decrypt_skb_update(struct sock *sk, struct sk_buff *skb,
 					       async);
 			if (err < 0) {
 				if (err == -EINPROGRESS)
-					tls_advance_record_sn(sk, &tls_ctx->rx,
-							      version);
+					tls_advance_record_sn(sk, prot,
+							      &tls_ctx->rx);
 
 				return err;
 			}
@@ -1518,7 +1519,7 @@ static int decrypt_skb_update(struct sock *sk, struct sk_buff *skb,
 		rxm->full_len -= pad;
 		rxm->offset += prot->prepend_size;
 		rxm->full_len -= prot->overhead_size;
-		tls_advance_record_sn(sk, &tls_ctx->rx, version);
+		tls_advance_record_sn(sk, prot, &tls_ctx->rx);
 		ctx->decrypted = true;
 		ctx->saved_data_ready(sk);
 	} else {