diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 31755d496b01..31ae4b74d435 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -733,10 +733,7 @@ struct sk_buff {
 		struct list_head	list;
 	};
 
-	union {
-		struct sock		*sk;
-		int			ip_defrag_offset;
-	};
+	struct sock		*sk;
 
 	union {
 		ktime_t		tstamp;
diff --git a/net/core/sock_destructor.h b/net/core/sock_destructor.h
new file mode 100644
index 000000000000..2f396e6bfba5
--- /dev/null
+++ b/net/core/sock_destructor.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+#ifndef _NET_CORE_SOCK_DESTRUCTOR_H
+#define _NET_CORE_SOCK_DESTRUCTOR_H
+#include <net/tcp.h>
+
+static inline bool is_skb_wmem(const struct sk_buff *skb)
+{
+	return skb->destructor == sock_wfree ||
+	       skb->destructor == __sock_wfree ||
+	       (IS_ENABLED(CONFIG_INET) && skb->destructor == tcp_wfree);
+}
+#endif
diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c
index e0e8a65d561e..12ef3cb26676 100644
--- a/net/ipv4/inet_fragment.c
+++ b/net/ipv4/inet_fragment.c
@@ -24,6 +24,8 @@
 #include <net/ip.h>
 #include <net/ipv6.h>
 
+#include "../core/sock_destructor.h"
+
 /* Use skb->cb to track consecutive/adjacent fragments coming at
  * the end of the queue. Nodes in the rb-tree queue will
  * contain "runs" of one or more adjacent fragments.
@@ -39,6 +41,7 @@ struct ipfrag_skb_cb {
 	};
 	struct sk_buff		*next_frag;
 	int			frag_run_len;
+	int			ip_defrag_offset;
 };
 
 #define FRAG_CB(skb)		((struct ipfrag_skb_cb *)((skb)->cb))
@@ -359,12 +362,12 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb,
 	 */
 	if (!last)
 		fragrun_create(q, skb);  /* First fragment. */
-	else if (last->ip_defrag_offset + last->len < end) {
+	else if (FRAG_CB(last)->ip_defrag_offset + last->len < end) {
 		/* This is the common case: skb goes to the end. */
 		/* Detect and discard overlaps. */
-		if (offset < last->ip_defrag_offset + last->len)
+		if (offset < FRAG_CB(last)->ip_defrag_offset + last->len)
 			return IPFRAG_OVERLAP;
-		if (offset == last->ip_defrag_offset + last->len)
+		if (offset == FRAG_CB(last)->ip_defrag_offset + last->len)
 			fragrun_append_to_last(q, skb);
 		else
 			fragrun_create(q, skb);
@@ -381,13 +384,13 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb,
 
 			parent = *rbn;
 			curr = rb_to_skb(parent);
-			curr_run_end = curr->ip_defrag_offset +
+			curr_run_end = FRAG_CB(curr)->ip_defrag_offset +
 					FRAG_CB(curr)->frag_run_len;
-			if (end <= curr->ip_defrag_offset)
+			if (end <= FRAG_CB(curr)->ip_defrag_offset)
 				rbn = &parent->rb_left;
 			else if (offset >= curr_run_end)
 				rbn = &parent->rb_right;
-			else if (offset >= curr->ip_defrag_offset &&
+			else if (offset >= FRAG_CB(curr)->ip_defrag_offset &&
 				 end <= curr_run_end)
 				return IPFRAG_DUP;
 			else
@@ -401,7 +404,7 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb,
 		rb_insert_color(&skb->rbnode, &q->rb_fragments);
 	}
 
-	skb->ip_defrag_offset = offset;
+	FRAG_CB(skb)->ip_defrag_offset = offset;
 
 	return IPFRAG_OK;
 }
@@ -411,13 +414,28 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb,
 			      struct sk_buff *parent)
 {
 	struct sk_buff *fp, *head = skb_rb_first(&q->rb_fragments);
-	struct sk_buff **nextp;
+	void (*destructor)(struct sk_buff *);
+	unsigned int orig_truesize = 0;
+	struct sk_buff **nextp = NULL;
+	struct sock *sk = skb->sk;
 	int delta;
 
+	if (sk && is_skb_wmem(skb)) {
+		/* TX: skb->sk might have been passed as argument to
+		 * dst->output and must remain valid until tx completes.
+		 *
+		 * Move sk to reassembled skb and fix up wmem accounting.
+		 */
+		orig_truesize = skb->truesize;
+		destructor = skb->destructor;
+	}
+
 	if (head != skb) {
 		fp = skb_clone(skb, GFP_ATOMIC);
-		if (!fp)
-			return NULL;
+		if (!fp) {
+			head = skb;
+			goto out_restore_sk;
+		}
 		FRAG_CB(fp)->next_frag = FRAG_CB(skb)->next_frag;
 		if (RB_EMPTY_NODE(&skb->rbnode))
 			FRAG_CB(parent)->next_frag = fp;
@@ -426,6 +444,12 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb,
 					&q->rb_fragments);
 		if (q->fragments_tail == skb)
 			q->fragments_tail = fp;
+
+		if (orig_truesize) {
+			/* prevent skb_morph from releasing sk */
+			skb->sk = NULL;
+			skb->destructor = NULL;
+		}
 		skb_morph(skb, head);
 		FRAG_CB(skb)->next_frag = FRAG_CB(head)->next_frag;
 		rb_replace_node(&head->rbnode, &skb->rbnode,
@@ -433,13 +457,13 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb,
 		consume_skb(head);
 		head = skb;
 	}
-	WARN_ON(head->ip_defrag_offset != 0);
+	WARN_ON(FRAG_CB(head)->ip_defrag_offset != 0);
 
 	delta = -head->truesize;
 
 	/* Head of list must not be cloned. */
 	if (skb_unclone(head, GFP_ATOMIC))
-		return NULL;
+		goto out_restore_sk;
 
 	delta += head->truesize;
 	if (delta)
@@ -455,7 +479,7 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb,
 
 		clone = alloc_skb(0, GFP_ATOMIC);
 		if (!clone)
-			return NULL;
+			goto out_restore_sk;
 		skb_shinfo(clone)->frag_list = skb_shinfo(head)->frag_list;
 		skb_frag_list_init(head);
 		for (i = 0; i < skb_shinfo(head)->nr_frags; i++)
@@ -472,6 +496,21 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb,
 		nextp = &skb_shinfo(head)->frag_list;
 	}
 
+out_restore_sk:
+	if (orig_truesize) {
+		int ts_delta = head->truesize - orig_truesize;
+
+		/* if this reassembled skb is fragmented later,
+		 * fraglist skbs will get skb->sk assigned from head->sk,
+		 * and each frag skb will be released via sock_wfree.
+		 *
+		 * Update sk_wmem_alloc.
+		 */
+		head->sk = sk;
+		head->destructor = destructor;
+		refcount_add(ts_delta, &sk->sk_wmem_alloc);
+	}
+
 	return nextp;
 }
 EXPORT_SYMBOL(inet_frag_reasm_prepare);
@@ -479,6 +518,8 @@ EXPORT_SYMBOL(inet_frag_reasm_prepare);
 void inet_frag_reasm_finish(struct inet_frag_queue *q, struct sk_buff *head,
 			    void *reasm_data, bool try_coalesce)
 {
+	struct sock *sk = is_skb_wmem(head) ? head->sk : NULL;
+	const unsigned int head_truesize = head->truesize;
 	struct sk_buff **nextp = (struct sk_buff **)reasm_data;
 	struct rb_node *rbn;
 	struct sk_buff *fp;
@@ -541,6 +582,9 @@ void inet_frag_reasm_finish(struct inet_frag_queue *q, struct sk_buff *head,
 	skb_mark_not_on_list(head);
 	head->prev = NULL;
 	head->tstamp = q->stamp;
+
+	if (sk)
+		refcount_add(sum_truesize - head_truesize, &sk->sk_wmem_alloc);
 }
 EXPORT_SYMBOL(inet_frag_reasm_finish);
 
diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c
index fad803d2d711..ec2264adf2a6 100644
--- a/net/ipv4/ip_fragment.c
+++ b/net/ipv4/ip_fragment.c
@@ -377,6 +377,7 @@ static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb)
 	}
 
 	skb_dst_drop(skb);
+	skb_orphan(skb);
 	return -EINPROGRESS;
 
 insert_error:
@@ -479,7 +480,6 @@ int ip_defrag(struct net *net, struct sk_buff *skb, u32 user)
 	struct ipq *qp;
 
 	__IP_INC_STATS(net, IPSTATS_MIB_REASMREQDS);
-	skb_orphan(skb);
 
 	/* Lookup (or create) queue header */
 	qp = ip_find(net, ip_hdr(skb), user, vif);
diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c
index c129ad334eb3..8c2163f95711 100644
--- a/net/ipv6/netfilter/nf_conntrack_reasm.c
+++ b/net/ipv6/netfilter/nf_conntrack_reasm.c
@@ -296,6 +296,7 @@ static int nf_ct_frag6_queue(struct frag_queue *fq, struct sk_buff *skb,
 	}
 
 	skb_dst_drop(skb);
+	skb_orphan(skb);
 	return -EINPROGRESS;
 
 insert_error:
@@ -471,7 +472,6 @@ int nf_ct_frag6_gather(struct net *net, struct sk_buff *skb, u32 user)
 	hdr = ipv6_hdr(skb);
 	fhdr = (struct frag_hdr *)skb_transport_header(skb);
 
-	skb_orphan(skb);
 	fq = fq_find(net, fhdr->identification, user, hdr,
 		     skb->dev ? skb->dev->ifindex : 0);
 	if (fq == NULL) {