9 files changed, 345 insertions, 285 deletions
diff --git a/net/xfrm/xfrm_algo.c b/net/xfrm/xfrm_algo.c
index e6da7e8495c9..749011e031c0 100644
--- a/net/xfrm/xfrm_algo.c
+++ b/net/xfrm/xfrm_algo.c
@@ -5,13 +5,13 @@
  * Copyright (c) 2002 James Morris <jmorris@intercode.com.au>
  */
 
+#include <crypto/acompress.h>
 #include <crypto/aead.h>
 #include <crypto/hash.h>
 #include <crypto/skcipher.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/pfkeyv2.h>
-#include <linux/crypto.h>
 #include <linux/scatterlist.h>
 #include <net/xfrm.h>
 #if IS_ENABLED(CONFIG_INET_ESP) || IS_ENABLED(CONFIG_INET6_ESP)
@@ -669,7 +669,7 @@ static const struct xfrm_algo_list xfrm_ealg_list = {
 };
 
 static const struct xfrm_algo_list xfrm_calg_list = {
-	.find = crypto_has_comp,
+	.find = crypto_has_acomp,
 	.algs = calg_list,
 	.entries = ARRAY_SIZE(calg_list),
 };
@@ -828,8 +828,7 @@ void xfrm_probe_algs(void)
 	}
 
 	for (i = 0; i < calg_entries(); i++) {
-		status = crypto_has_comp(calg_list[i].name, 0,
-					 CRYPTO_ALG_ASYNC);
+		status = crypto_has_acomp(calg_list[i].name, 0, 0);
 		if (calg_list[i].available != status)
 			calg_list[i].available = status;
 	}
diff --git a/net/xfrm/xfrm_device.c b/net/xfrm/xfrm_device.c
index d1fa94e52cea..d62f76161d83 100644
--- a/net/xfrm/xfrm_device.c
+++ b/net/xfrm/xfrm_device.c
@@ -244,11 +244,6 @@ int xfrm_dev_state_add(struct net *net, struct xfrm_state *x,
 	xfrm_address_t *daddr;
 	bool is_packet_offload;
 
-	if (!x->type_offload) {
-		NL_SET_ERR_MSG(extack, "Type doesn't support offload");
-		return -EINVAL;
-	}
-
 	if (xuo->flags &
 	    ~(XFRM_OFFLOAD_IPV6 | XFRM_OFFLOAD_INBOUND | XFRM_OFFLOAD_PACKET)) {
 		NL_SET_ERR_MSG(extack, "Unrecognized flags in offload request");
@@ -310,6 +305,13 @@ int xfrm_dev_state_add(struct net *net, struct xfrm_state *x,
 		return -EINVAL;
 	}
 
+	xfrm_set_type_offload(x);
+	if (!x->type_offload) {
+		NL_SET_ERR_MSG(extack, "Type doesn't support offload");
+		dev_put(dev);
+		return -EINVAL;
+	}
+
 	xso->dev = dev;
 	netdev_tracker_alloc(dev, &xso->dev_tracker, GFP_ATOMIC);
 	xso->real_dev = dev;
@@ -332,6 +334,7 @@ int xfrm_dev_state_add(struct net *net, struct xfrm_state *x,
 		netdev_put(dev, &xso->dev_tracker);
 		xso->type = XFRM_DEV_OFFLOAD_UNSPECIFIED;
 
+		xfrm_unset_type_offload(x);
 		/* User explicitly requested packet offload mode and configured
 		 * policy in addition to the XFRM state. So be civil to users,
 		 * and return an error instead of taking fallback path.
@@ -415,14 +418,12 @@ bool xfrm_dev_offload_ok(struct sk_buff *skb, struct xfrm_state *x)
 	struct dst_entry *dst = skb_dst(skb);
 	struct xfrm_dst *xdst = (struct xfrm_dst *)dst;
 	struct net_device *dev = x->xso.dev;
+	bool check_tunnel_size;
 
-	if (!x->type_offload ||
-	    (x->xso.type == XFRM_DEV_OFFLOAD_UNSPECIFIED && x->encap))
+	if (x->xso.type == XFRM_DEV_OFFLOAD_UNSPECIFIED)
 		return false;
 
-	if (x->xso.type == XFRM_DEV_OFFLOAD_PACKET ||
-	    ((!dev || (dev == xfrm_dst_path(dst)->dev)) &&
-	     !xdst->child->xfrm)) {
+	if ((dev == xfrm_dst_path(dst)->dev) && !xdst->child->xfrm) {
 		mtu = xfrm_state_mtu(x, xdst->child_mtu_cached);
 		if (skb->len <= mtu)
 			goto ok;
@@ -434,8 +435,29 @@ bool xfrm_dev_offload_ok(struct sk_buff *skb, struct xfrm_state *x)
 	return false;
 
 ok:
-	if (dev && dev->xfrmdev_ops && dev->xfrmdev_ops->xdo_dev_offload_ok)
-		return x->xso.dev->xfrmdev_ops->xdo_dev_offload_ok(skb, x);
+	check_tunnel_size = x->xso.type == XFRM_DEV_OFFLOAD_PACKET &&
+			    x->props.mode == XFRM_MODE_TUNNEL;
+	switch (x->props.family) {
+	case AF_INET:
+		/* Check for IPv4 options */
+		if (ip_hdr(skb)->ihl != 5)
+			return false;
+		if (check_tunnel_size && xfrm4_tunnel_check_size(skb))
+			return false;
+		break;
+	case AF_INET6:
+		/* Check for IPv6 extensions */
+		if (ipv6_ext_hdr(ipv6_hdr(skb)->nexthdr))
+			return false;
+		if (check_tunnel_size && xfrm6_tunnel_check_size(skb))
+			return false;
+		break;
+	default:
+		break;
+	}
+
+	if (dev->xfrmdev_ops->xdo_dev_offload_ok)
+		return dev->xfrmdev_ops->xdo_dev_offload_ok(skb, x);
 
 	return true;
 }
diff --git a/net/xfrm/xfrm_interface_core.c b/net/xfrm/xfrm_interface_core.c
index c397eb99d867..622445f041d3 100644
--- a/net/xfrm/xfrm_interface_core.c
+++ b/net/xfrm/xfrm_interface_core.c
@@ -242,10 +242,9 @@ static void xfrmi_dev_free(struct net_device *dev)
 	gro_cells_destroy(&xi->gro_cells);
 }
 
-static int xfrmi_create(struct net_device *dev)
+static int xfrmi_create(struct net *net, struct net_device *dev)
 {
 	struct xfrm_if *xi = netdev_priv(dev);
-	struct net *net = dev_net(dev);
 	struct xfrmi_net *xfrmn = net_generic(net, xfrmi_net_id);
 	int err;
 
@@ -814,15 +813,17 @@ static void xfrmi_netlink_parms(struct nlattr *data[],
 		parms->collect_md = true;
 }
 
-static int xfrmi_newlink(struct net *src_net, struct net_device *dev,
-			struct nlattr *tb[], struct nlattr *data[],
-			struct netlink_ext_ack *extack)
+static int xfrmi_newlink(struct net_device *dev,
+			 struct rtnl_newlink_params *params,
+			 struct netlink_ext_ack *extack)
 {
-	struct net *net = dev_net(dev);
+	struct nlattr **data = params->data;
 	struct xfrm_if_parms p = {};
 	struct xfrm_if *xi;
+	struct net *net;
 	int err;
 
+	net = params->link_net ? : dev_net(dev);
 	xfrmi_netlink_parms(data, &p);
 	if (p.collect_md) {
 		struct xfrmi_net *xfrmn = net_generic(net, xfrmi_net_id);
@@ -851,7 +852,7 @@ static int xfrmi_newlink(struct net *src_net, struct net_device *dev,
 	xi->net = net;
 	xi->dev = dev;
 
-	err = xfrmi_create(dev);
+	err = xfrmi_create(net, dev);
 	return err;
 }
 
diff --git a/net/xfrm/xfrm_ipcomp.c b/net/xfrm/xfrm_ipcomp.c
index 9c0fa0e1786a..0c1420534394 100644
--- a/net/xfrm/xfrm_ipcomp.c
+++ b/net/xfrm/xfrm_ipcomp.c
@@ -3,7 +3,7 @@
  * IP Payload Compression Protocol (IPComp) - RFC3173.
  *
  * Copyright (c) 2003 James Morris <jmorris@intercode.com.au>
- * Copyright (c) 2003-2008 Herbert Xu <herbert@gondor.apana.org.au>
+ * Copyright (c) 2003-2025 Herbert Xu <herbert@gondor.apana.org.au>
  *
  * Todo:
  *   - Tunable compression parameters.
@@ -11,303 +11,302 @@
  *   - Adaptive compression.
  */
 
-#include <linux/crypto.h>
+#include <crypto/acompress.h>
 #include <linux/err.h>
-#include <linux/list.h>
 #include <linux/module.h>
-#include <linux/mutex.h>
-#include <linux/percpu.h>
+#include <linux/skbuff_ref.h>
 #include <linux/slab.h>
-#include <linux/smp.h>
-#include <linux/vmalloc.h>
-#include <net/ip.h>
 #include <net/ipcomp.h>
 #include <net/xfrm.h>
 
-struct ipcomp_tfms {
-	struct list_head list;
-	struct crypto_comp * __percpu *tfms;
-	int users;
+#define IPCOMP_SCRATCH_SIZE 65400
+
+struct ipcomp_skb_cb {
+	struct xfrm_skb_cb xfrm;
+	struct acomp_req *req;
 };
 
-static DEFINE_MUTEX(ipcomp_resource_mutex);
-static void * __percpu *ipcomp_scratches;
-static int ipcomp_scratch_users;
-static LIST_HEAD(ipcomp_tfms_list);
+struct ipcomp_data {
+	u16 threshold;
+	struct crypto_acomp *tfm;
+};
 
-static int ipcomp_decompress(struct xfrm_state *x, struct sk_buff *skb)
+struct ipcomp_req_extra {
+	struct xfrm_state *x;
+	struct scatterlist sg[];
+};
+
+static inline struct ipcomp_skb_cb *ipcomp_cb(struct sk_buff *skb)
 {
-	struct ipcomp_data *ipcd = x->data;
-	const int plen = skb->len;
-	int dlen = IPCOMP_SCRATCH_SIZE;
-	const u8 *start = skb->data;
-	u8 *scratch = *this_cpu_ptr(ipcomp_scratches);
-	struct crypto_comp *tfm = *this_cpu_ptr(ipcd->tfms);
-	int err = crypto_comp_decompress(tfm, start, plen, scratch, &dlen);
-	int len;
+	struct ipcomp_skb_cb *cb = (void *)skb->cb;
 
-	if (err)
-		return err;
+	BUILD_BUG_ON(sizeof(*cb) > sizeof(skb->cb));
+	return cb;
+}
 
-	if (dlen < (plen + sizeof(struct ip_comp_hdr)))
-		return -EINVAL;
+static int ipcomp_post_acomp(struct sk_buff *skb, int err, int hlen)
+{
+	struct acomp_req *req = ipcomp_cb(skb)->req;
+	struct ipcomp_req_extra *extra;
+	const int plen = skb->data_len;
+	struct scatterlist *dsg;
+	int len, dlen;
 
-	len = dlen - plen;
-	if (len > skb_tailroom(skb))
-		len = skb_tailroom(skb);
+	if (unlikely(err))
+		goto out_free_req;
 
-	__skb_put(skb, len);
+	extra = acomp_request_extra(req);
+	dsg = extra->sg;
+	dlen = req->dlen;
 
-	len += plen;
-	skb_copy_to_linear_data(skb, scratch, len);
+	pskb_trim_unique(skb, 0);
+	__skb_put(skb, hlen);
 
-	while ((scratch += len, dlen -= len) > 0) {
+	/* Only update truesize on input. */
+	if (!hlen)
+		skb->truesize += dlen - plen;
+	skb->data_len = dlen;
+	skb->len += dlen;
+
+	do {
 		skb_frag_t *frag;
 		struct page *page;
 
-		if (WARN_ON(skb_shinfo(skb)->nr_frags >= MAX_SKB_FRAGS))
-			return -EMSGSIZE;
-
 		frag = skb_shinfo(skb)->frags + skb_shinfo(skb)->nr_frags;
-		page = alloc_page(GFP_ATOMIC);
-
-		if (!page)
-			return -ENOMEM;
+		page = sg_page(dsg);
+		dsg = sg_next(dsg);
 
 		len = PAGE_SIZE;
 		if (dlen < len)
 			len = dlen;
 
 		skb_frag_fill_page_desc(frag, page, 0, len);
-		memcpy(skb_frag_address(frag), scratch, len);
-
-		skb->truesize += len;
-		skb->data_len += len;
-		skb->len += len;
 
 		skb_shinfo(skb)->nr_frags++;
-	}
+	} while ((dlen -= len));
 
-	return 0;
+	for (; dsg; dsg = sg_next(dsg))
+		__free_page(sg_page(dsg));
+
+out_free_req:
+	acomp_request_free(req);
+	return err;
 }
 
-int ipcomp_input(struct xfrm_state *x, struct sk_buff *skb)
+static int ipcomp_input_done2(struct sk_buff *skb, int err)
 {
-	int nexthdr;
-	int err = -ENOMEM;
-	struct ip_comp_hdr *ipch;
-
-	if (skb_linearize_cow(skb))
-		goto out;
-
-	skb->ip_summed = CHECKSUM_NONE;
+	struct ip_comp_hdr *ipch = ip_comp_hdr(skb);
+	const int plen = skb->len;
 
-	/* Remove ipcomp header and decompress original payload */
-	ipch = (void *)skb->data;
-	nexthdr = ipch->nexthdr;
+	skb_reset_transport_header(skb);
 
-	skb->transport_header = skb->network_header + sizeof(*ipch);
-	__skb_pull(skb, sizeof(*ipch));
-	err = ipcomp_decompress(x, skb);
-	if (err)
-		goto out;
+	return ipcomp_post_acomp(skb, err, 0) ?:
+	       skb->len < (plen + sizeof(ip_comp_hdr)) ? -EINVAL :
+	       ipch->nexthdr;
+}
 
-	err = nexthdr;
+static void ipcomp_input_done(void *data, int err)
+{
+	struct sk_buff *skb = data;
 
-out:
-	return err;
+	xfrm_input_resume(skb, ipcomp_input_done2(skb, err));
 }
-EXPORT_SYMBOL_GPL(ipcomp_input);
 
-static int ipcomp_compress(struct xfrm_state *x, struct sk_buff *skb)
+static struct acomp_req *ipcomp_setup_req(struct xfrm_state *x,
+					  struct sk_buff *skb, int minhead,
+					  int dlen)
 {
+	const int dnfrags = min(MAX_SKB_FRAGS, 16);
 	struct ipcomp_data *ipcd = x->data;
+	struct ipcomp_req_extra *extra;
+	struct scatterlist *sg, *dsg;
 	const int plen = skb->len;
-	int dlen = IPCOMP_SCRATCH_SIZE;
-	u8 *start = skb->data;
-	struct crypto_comp *tfm;
-	u8 *scratch;
+	struct crypto_acomp *tfm;
+	struct acomp_req *req;
+	int nfrags;
+	int total;
 	int err;
+	int i;
 
-	local_bh_disable();
-	scratch = *this_cpu_ptr(ipcomp_scratches);
-	tfm = *this_cpu_ptr(ipcd->tfms);
-	err = crypto_comp_compress(tfm, start, plen, scratch, &dlen);
-	if (err)
-		goto out;
-
-	if ((dlen + sizeof(struct ip_comp_hdr)) >= plen) {
-		err = -EMSGSIZE;
-		goto out;
-	}
+	ipcomp_cb(skb)->req = NULL;
 
-	memcpy(start + sizeof(struct ip_comp_hdr), scratch, dlen);
-	local_bh_enable();
+	do {
+		struct sk_buff *trailer;
 
-	pskb_trim(skb, dlen + sizeof(struct ip_comp_hdr));
-	return 0;
+		if (skb->len > PAGE_SIZE) {
+			if (skb_linearize_cow(skb))
+				return ERR_PTR(-ENOMEM);
+			nfrags = 1;
+			break;
+		}
 
-out:
-	local_bh_enable();
-	return err;
-}
+		if (!skb_cloned(skb) && skb_headlen(skb) >= minhead) {
+			if (!skb_is_nonlinear(skb)) {
+				nfrags = 1;
+				break;
+			} else if (!skb_has_frag_list(skb)) {
+				nfrags = skb_shinfo(skb)->nr_frags;
+				nfrags++;
+				break;
+			}
+		}
 
-int ipcomp_output(struct xfrm_state *x, struct sk_buff *skb)
-{
-	int err;
-	struct ip_comp_hdr *ipch;
-	struct ipcomp_data *ipcd = x->data;
+		nfrags = skb_cow_data(skb, skb_headlen(skb) < minhead ?
+					   minhead - skb_headlen(skb) : 0,
+				      &trailer);
+		if (nfrags < 0)
+			return ERR_PTR(nfrags);
+	} while (0);
+
+	tfm = ipcd->tfm;
+	req = acomp_request_alloc_extra(
+		tfm, sizeof(*extra) + sizeof(*sg) * (nfrags + dnfrags),
+		GFP_ATOMIC);
+	ipcomp_cb(skb)->req = req;
+	if (!req)
+		return ERR_PTR(-ENOMEM);
+
+	extra = acomp_request_extra(req);
+	extra->x = x;
+
+	dsg = extra->sg;
+	sg = dsg + dnfrags;
+	sg_init_table(sg, nfrags);
+	err = skb_to_sgvec(skb, sg, 0, plen);
+	if (unlikely(err < 0))
+		return ERR_PTR(err);
+
+	sg_init_table(dsg, dnfrags);
+	total = 0;
+	for (i = 0; i < dnfrags && total < dlen; i++) {
+		struct page *page;
 
-	if (skb->len < ipcd->threshold) {
-		/* Don't bother compressing */
-		goto out_ok;
+		page = alloc_page(GFP_ATOMIC);
+		if (!page)
+			break;
+		sg_set_page(dsg + i, page, PAGE_SIZE, 0);
+		total += PAGE_SIZE;
 	}
+	if (!i)
+		return ERR_PTR(-ENOMEM);
+	sg_mark_end(dsg + i - 1);
+	dlen = min(dlen, total);
 
-	if (skb_linearize_cow(skb))
-		goto out_ok;
-
-	err = ipcomp_compress(x, skb);
-
-	if (err) {
-		goto out_ok;
-	}
+	acomp_request_set_params(req, sg, dsg, plen, dlen);
 
-	/* Install ipcomp header, convert into ipcomp datagram. */
-	ipch = ip_comp_hdr(skb);
-	ipch->nexthdr = *skb_mac_header(skb);
-	ipch->flags = 0;
-	ipch->cpi = htons((u16 )ntohl(x->id.spi));
-	*skb_mac_header(skb) = IPPROTO_COMP;
-out_ok:
-	skb_push(skb, -skb_network_offset(skb));
-	return 0;
+	return req;
 }
-EXPORT_SYMBOL_GPL(ipcomp_output);
 
-static void ipcomp_free_scratches(void)
+static int ipcomp_decompress(struct xfrm_state *x, struct sk_buff *skb)
 {
-	int i;
-	void * __percpu *scratches;
-
-	if (--ipcomp_scratch_users)
-		return;
+	struct acomp_req *req;
+	int err;
 
-	scratches = ipcomp_scratches;
-	if (!scratches)
-		return;
+	req = ipcomp_setup_req(x, skb, 0, IPCOMP_SCRATCH_SIZE);
+	err = PTR_ERR(req);
+	if (IS_ERR(req))
+		goto out;
 
-	for_each_possible_cpu(i)
-		vfree(*per_cpu_ptr(scratches, i));
+	acomp_request_set_callback(req, 0, ipcomp_input_done, skb);
+	err = crypto_acomp_decompress(req);
+	if (err == -EINPROGRESS)
+		return err;
 
-	free_percpu(scratches);
-	ipcomp_scratches = NULL;
+out:
+	return ipcomp_input_done2(skb, err);
 }
 
-static void * __percpu *ipcomp_alloc_scratches(void)
+int ipcomp_input(struct xfrm_state *x, struct sk_buff *skb)
 {
-	void * __percpu *scratches;
-	int i;
-
-	if (ipcomp_scratch_users++)
-		return ipcomp_scratches;
-
-	scratches = alloc_percpu(void *);
-	if (!scratches)
-		return NULL;
+	struct ip_comp_hdr *ipch __maybe_unused;
 
-	ipcomp_scratches = scratches;
+	if (!pskb_may_pull(skb, sizeof(*ipch)))
+		return -EINVAL;
 
-	for_each_possible_cpu(i) {
-		void *scratch;
+	skb->ip_summed = CHECKSUM_NONE;
 
-		scratch = vmalloc_node(IPCOMP_SCRATCH_SIZE, cpu_to_node(i));
-		if (!scratch)
-			return NULL;
-		*per_cpu_ptr(scratches, i) = scratch;
-	}
+	/* Remove ipcomp header and decompress original payload */
+	__skb_pull(skb, sizeof(*ipch));
 
-	return scratches;
+	return ipcomp_decompress(x, skb);
 }
+EXPORT_SYMBOL_GPL(ipcomp_input);
 
-static void ipcomp_free_tfms(struct crypto_comp * __percpu *tfms)
+static int ipcomp_output_push(struct sk_buff *skb)
 {
-	struct ipcomp_tfms *pos;
-	int cpu;
-
-	list_for_each_entry(pos, &ipcomp_tfms_list, list) {
-		if (pos->tfms == tfms)
-			break;
-	}
-
-	WARN_ON(list_entry_is_head(pos, &ipcomp_tfms_list, list));
-
-	if (--pos->users)
-		return;
+	skb_push(skb, -skb_network_offset(skb));
+	return 0;
+}
 
-	list_del(&pos->list);
-	kfree(pos);
+static int ipcomp_output_done2(struct xfrm_state *x, struct sk_buff *skb,
+			       int err)
+{
+	struct ip_comp_hdr *ipch;
 
-	if (!tfms)
-		return;
+	err = ipcomp_post_acomp(skb, err, sizeof(*ipch));
+	if (err)
+		goto out_ok;
 
-	for_each_possible_cpu(cpu) {
-		struct crypto_comp *tfm = *per_cpu_ptr(tfms, cpu);
-		crypto_free_comp(tfm);
-	}
-	free_percpu(tfms);
+	/* Install ipcomp header, convert into ipcomp datagram. */
+	ipch = ip_comp_hdr(skb);
+	ipch->nexthdr = *skb_mac_header(skb);
+	ipch->flags = 0;
+	ipch->cpi = htons((u16 )ntohl(x->id.spi));
+	*skb_mac_header(skb) = IPPROTO_COMP;
+out_ok:
+	return ipcomp_output_push(skb);
 }
 
-static struct crypto_comp * __percpu *ipcomp_alloc_tfms(const char *alg_name)
+static void ipcomp_output_done(void *data, int err)
 {
-	struct ipcomp_tfms *pos;
-	struct crypto_comp * __percpu *tfms;
-	int cpu;
+	struct ipcomp_req_extra *extra;
+	struct sk_buff *skb = data;
+	struct acomp_req *req;
 
+	req = ipcomp_cb(skb)->req;
+	extra = acomp_request_extra(req);
 
-	list_for_each_entry(pos, &ipcomp_tfms_list, list) {
-		struct crypto_comp *tfm;
+	xfrm_output_resume(skb_to_full_sk(skb), skb,
+			   ipcomp_output_done2(extra->x, skb, err));
+}
 
-		/* This can be any valid CPU ID so we don't need locking. */
-		tfm = this_cpu_read(*pos->tfms);
+static int ipcomp_compress(struct xfrm_state *x, struct sk_buff *skb)
+{
+	struct ip_comp_hdr *ipch __maybe_unused;
+	struct acomp_req *req;
+	int err;
 
-		if (!strcmp(crypto_comp_name(tfm), alg_name)) {
-			pos->users++;
-			return pos->tfms;
-		}
-	}
+	req = ipcomp_setup_req(x, skb, sizeof(*ipch),
+			       skb->len - sizeof(*ipch));
+	err = PTR_ERR(req);
+	if (IS_ERR(req))
+		goto out;
 
-	pos = kmalloc(sizeof(*pos), GFP_KERNEL);
-	if (!pos)
-		return NULL;
+	acomp_request_set_callback(req, 0, ipcomp_output_done, skb);
+	err = crypto_acomp_compress(req);
+	if (err == -EINPROGRESS)
+		return err;
 
-	pos->users = 1;
-	INIT_LIST_HEAD(&pos->list);
-	list_add(&pos->list, &ipcomp_tfms_list);
+out:
+	return ipcomp_output_done2(x, skb, err);
+}
 
-	pos->tfms = tfms = alloc_percpu(struct crypto_comp *);
-	if (!tfms)
-		goto error;
+int ipcomp_output(struct xfrm_state *x, struct sk_buff *skb)
+{
+	struct ipcomp_data *ipcd = x->data;
 
-	for_each_possible_cpu(cpu) {
-		struct crypto_comp *tfm = crypto_alloc_comp(alg_name, 0,
-							    CRYPTO_ALG_ASYNC);
-		if (IS_ERR(tfm))
-			goto error;
-		*per_cpu_ptr(tfms, cpu) = tfm;
+	if (skb->len < ipcd->threshold) {
+		/* Don't bother compressing */
+		return ipcomp_output_push(skb);
 	}
 
-	return tfms;
-
-error:
-	ipcomp_free_tfms(tfms);
-	return NULL;
+	return ipcomp_compress(x, skb);
 }
+EXPORT_SYMBOL_GPL(ipcomp_output);
 
 static void ipcomp_free_data(struct ipcomp_data *ipcd)
 {
-	if (ipcd->tfms)
-		ipcomp_free_tfms(ipcd->tfms);
-	ipcomp_free_scratches();
+	crypto_free_acomp(ipcd->tfm);
 }
 
 void ipcomp_destroy(struct xfrm_state *x)
@@ -316,9 +315,7 @@ void ipcomp_destroy(struct xfrm_state *x)
 	if (!ipcd)
 		return;
 	xfrm_state_delete_tunnel(x);
-	mutex_lock(&ipcomp_resource_mutex);
 	ipcomp_free_data(ipcd);
-	mutex_unlock(&ipcomp_resource_mutex);
 	kfree(ipcd);
 }
 EXPORT_SYMBOL_GPL(ipcomp_destroy);
@@ -345,14 +342,9 @@ int ipcomp_init_state(struct xfrm_state *x, struct netlink_ext_ack *extack)
 	if (!ipcd)
 		goto out;
 
-	mutex_lock(&ipcomp_resource_mutex);
-	if (!ipcomp_alloc_scratches())
-		goto error;
-
-	ipcd->tfms = ipcomp_alloc_tfms(x->calg->alg_name);
-	if (!ipcd->tfms)
+	ipcd->tfm = crypto_alloc_acomp(x->calg->alg_name, 0, 0);
+	if (IS_ERR(ipcd->tfm))
 		goto error;
-	mutex_unlock(&ipcomp_resource_mutex);
 
 	calg_desc = xfrm_calg_get_byname(x->calg->alg_name, 0);
 	BUG_ON(!calg_desc);
@@ -364,7 +356,6 @@ out:
 
 error:
 	ipcomp_free_data(ipcd);
-	mutex_unlock(&ipcomp_resource_mutex);
 	kfree(ipcd);
 	goto out;
 }
diff --git a/net/xfrm/xfrm_iptfs.c b/net/xfrm/xfrm_iptfs.c
index 755f1eea8bfa..3b6d7284fc70 100644
--- a/net/xfrm/xfrm_iptfs.c
+++ b/net/xfrm/xfrm_iptfs.c
@@ -2625,12 +2625,10 @@ static void __iptfs_init_state(struct xfrm_state *x,
 			       struct xfrm_iptfs_data *xtfs)
 {
 	__skb_queue_head_init(&xtfs->queue);
-	hrtimer_init(&xtfs->iptfs_timer, CLOCK_MONOTONIC, IPTFS_HRTIMER_MODE);
-	xtfs->iptfs_timer.function = iptfs_delay_timer;
+	hrtimer_setup(&xtfs->iptfs_timer, iptfs_delay_timer, CLOCK_MONOTONIC, IPTFS_HRTIMER_MODE);
 
 	spin_lock_init(&xtfs->drop_lock);
-	hrtimer_init(&xtfs->drop_timer, CLOCK_MONOTONIC, IPTFS_HRTIMER_MODE);
-	xtfs->drop_timer.function = iptfs_drop_timer;
+	hrtimer_setup(&xtfs->drop_timer, iptfs_drop_timer, CLOCK_MONOTONIC, IPTFS_HRTIMER_MODE);
 
 	/* Modify type (esp) adjustment values */
 
diff --git a/net/xfrm/xfrm_output.c b/net/xfrm/xfrm_output.c
index f7abd42c077d..9077730ff7d0 100644
--- a/net/xfrm/xfrm_output.c
+++ b/net/xfrm/xfrm_output.c
@@ -612,6 +612,40 @@ out:
 }
 EXPORT_SYMBOL_GPL(xfrm_output_resume);
 
+static int xfrm_dev_direct_output(struct sock *sk, struct xfrm_state *x,
+				  struct sk_buff *skb)
+{
+	struct dst_entry *dst = skb_dst(skb);
+	struct net *net = xs_net(x);
+	int err;
+
+	dst = skb_dst_pop(skb);
+	if (!dst) {
+		XFRM_INC_STATS(net, LINUX_MIB_XFRMOUTERROR);
+		kfree_skb(skb);
+		return -EHOSTUNREACH;
+	}
+	skb_dst_set(skb, dst);
+	nf_reset_ct(skb);
+
+	err = skb_dst(skb)->ops->local_out(net, sk, skb);
+	if (unlikely(err != 1)) {
+		kfree_skb(skb);
+		return err;
+	}
+
+	/* In transport mode, network destination is
+	 * directly reachable, while in tunnel mode,
+	 * inner packet network may not be. In packet
+	 * offload type, HW is responsible for hard
+	 * header packet mangling so directly xmit skb
+	 * to netdevice.
+	 */
+	skb->dev = x->xso.dev;
+	__skb_push(skb, skb->dev->hard_header_len);
+	return dev_queue_xmit(skb);
+}
+
 static int xfrm_output2(struct net *net, struct sock *sk, struct sk_buff *skb)
 {
 	return xfrm_output_resume(sk, skb, 1);
@@ -735,6 +769,13 @@ int xfrm_output(struct sock *sk, struct sk_buff *skb)
 			return -EHOSTUNREACH;
 		}
 
+		/* Exclusive direct xmit for tunnel mode, as
+		 * some filtering or matching rules may apply
+		 * in transport mode.
+		 */
+		if (x->props.mode == XFRM_MODE_TUNNEL)
+			return xfrm_dev_direct_output(sk, x, skb);
+
 		return xfrm_output_resume(sk, skb, 0);
 	}
 
@@ -758,7 +799,7 @@ int xfrm_output(struct sock *sk, struct sk_buff *skb)
 		skb->encapsulation = 1;
 
 		if (skb_is_gso(skb)) {
-			if (skb->inner_protocol)
+			if (skb->inner_protocol && x->props.mode == XFRM_MODE_TUNNEL)
 				return xfrm_output_gso(net, sk, skb);
 
 			skb_shinfo(skb)->gso_type |= SKB_GSO_ESP;
@@ -786,7 +827,7 @@ out:
 }
 EXPORT_SYMBOL_GPL(xfrm_output);
 
-static int xfrm4_tunnel_check_size(struct sk_buff *skb)
+int xfrm4_tunnel_check_size(struct sk_buff *skb)
 {
 	int mtu, ret = 0;
 
@@ -812,6 +853,7 @@ static int xfrm4_tunnel_check_size(struct sk_buff *skb)
 out:
 	return ret;
 }
+EXPORT_SYMBOL_GPL(xfrm4_tunnel_check_size);
 
 static int xfrm4_extract_output(struct xfrm_state *x, struct sk_buff *skb)
 {
@@ -834,7 +876,7 @@ static int xfrm4_extract_output(struct xfrm_state *x, struct sk_buff *skb)
 }
 
 #if IS_ENABLED(CONFIG_IPV6)
-static int xfrm6_tunnel_check_size(struct sk_buff *skb)
+int xfrm6_tunnel_check_size(struct sk_buff *skb)
 {
 	int mtu, ret = 0;
 	struct dst_entry *dst = skb_dst(skb);
@@ -864,6 +906,7 @@ static int xfrm6_tunnel_check_size(struct sk_buff *skb)
 out:
 	return ret;
 }
+EXPORT_SYMBOL_GPL(xfrm6_tunnel_check_size);
 #endif
 
 static int xfrm6_extract_output(struct xfrm_state *x, struct sk_buff *skb)
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index 6551e588fe52..30970d40a454 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -3294,7 +3294,7 @@ no_transform:
 
 ok:
 	xfrm_pols_put(pols, drop_pols);
-	if (dst && dst->xfrm &&
+	if (dst->xfrm &&
 	    (dst->xfrm->props.mode == XFRM_MODE_TUNNEL ||
 	     dst->xfrm->props.mode == XFRM_MODE_IPTFS))
 		dst->flags |= DST_XFRM_TUNNEL;
diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index ad2202fa82f3..d896c3fefb07 100644
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -424,18 +424,18 @@ void xfrm_unregister_type_offload(const struct xfrm_type_offload *type,
 }
 EXPORT_SYMBOL(xfrm_unregister_type_offload);
 
-static const struct xfrm_type_offload *
-xfrm_get_type_offload(u8 proto, unsigned short family, bool try_load)
+void xfrm_set_type_offload(struct xfrm_state *x)
 {
 	const struct xfrm_type_offload *type = NULL;
 	struct xfrm_state_afinfo *afinfo;
+	bool try_load = true;
 
 retry:
-	afinfo = xfrm_state_get_afinfo(family);
+	afinfo = xfrm_state_get_afinfo(x->props.family);
 	if (unlikely(afinfo == NULL))
-		return NULL;
+		goto out;
 
-	switch (proto) {
+	switch (x->id.proto) {
 	case IPPROTO_ESP:
 		type = afinfo->type_offload_esp;
 		break;
@@ -449,18 +449,16 @@ retry:
 	rcu_read_unlock();
 
 	if (!type && try_load) {
-		request_module("xfrm-offload-%d-%d", family, proto);
+		request_module("xfrm-offload-%d-%d", x->props.family,
+			       x->id.proto);
 		try_load = false;
 		goto retry;
 	}
 
-	return type;
-}
-
-static void xfrm_put_type_offload(const struct xfrm_type_offload *type)
-{
-	module_put(type->owner);
+out:
+	x->type_offload = type;
 }
+EXPORT_SYMBOL(xfrm_set_type_offload);
 
 static const struct xfrm_mode xfrm4_mode_map[XFRM_MODE_MAX] = {
 	[XFRM_MODE_BEET] = {
@@ -609,8 +607,6 @@ static void ___xfrm_state_destroy(struct xfrm_state *x)
 	kfree(x->coaddr);
 	kfree(x->replay_esn);
 	kfree(x->preplay_esn);
-	if (x->type_offload)
-		xfrm_put_type_offload(x->type_offload);
 	if (x->type) {
 		x->type->destructor(x);
 		xfrm_put_type(x->type);
@@ -746,8 +742,8 @@ struct xfrm_state *xfrm_state_alloc(struct net *net)
 		INIT_HLIST_NODE(&x->bysrc);
 		INIT_HLIST_NODE(&x->byspi);
 		INIT_HLIST_NODE(&x->byseq);
-		hrtimer_init(&x->mtimer, CLOCK_BOOTTIME, HRTIMER_MODE_ABS_SOFT);
-		x->mtimer.function = xfrm_timer_handler;
+		hrtimer_setup(&x->mtimer, xfrm_timer_handler, CLOCK_BOOTTIME,
+			      HRTIMER_MODE_ABS_SOFT);
 		timer_setup(&x->rtimer, xfrm_replay_timer_handler, 0);
 		x->curlft.add_time = ktime_get_real_seconds();
 		x->lft.soft_byte_limit = XFRM_INF;
@@ -784,6 +780,8 @@ void xfrm_dev_state_free(struct xfrm_state *x)
 	struct xfrm_dev_offload *xso = &x->xso;
 	struct net_device *dev = READ_ONCE(xso->dev);
 
+	xfrm_unset_type_offload(x);
+
 	if (dev && dev->xfrmdev_ops) {
 		spin_lock_bh(&xfrm_state_dev_gc_lock);
 		if (!hlist_unhashed(&x->dev_gclist))
@@ -2315,12 +2313,12 @@ xfrm_state_lookup_byaddr(struct net *net, u32 mark,
 	struct xfrm_hash_state_ptrs state_ptrs;
 	struct xfrm_state *x;
 
-	spin_lock_bh(&net->xfrm.xfrm_state_lock);
+	rcu_read_lock();
 
 	xfrm_hash_ptrs_get(net, &state_ptrs);
 
 	x = __xfrm_state_lookup_byaddr(&state_ptrs, mark, daddr, saddr, proto, family);
-	spin_unlock_bh(&net->xfrm.xfrm_state_lock);
+	rcu_read_unlock();
 	return x;
 }
 EXPORT_SYMBOL(xfrm_state_lookup_byaddr);
@@ -3122,8 +3120,7 @@ u32 xfrm_state_mtu(struct xfrm_state *x, int mtu)
 }
 EXPORT_SYMBOL_GPL(xfrm_state_mtu);
 
-int __xfrm_init_state(struct xfrm_state *x, bool init_replay, bool offload,
-		      struct netlink_ext_ack *extack)
+int __xfrm_init_state(struct xfrm_state *x, struct netlink_ext_ack *extack)
 {
 	const struct xfrm_mode *inner_mode;
 	const struct xfrm_mode *outer_mode;
@@ -3178,8 +3175,6 @@ int __xfrm_init_state(struct xfrm_state *x, bool init_replay, bool offload,
 		goto error;
 	}
 
-	x->type_offload = xfrm_get_type_offload(x->id.proto, family, offload);
-
 	err = x->type->init_state(x, extack);
 	if (err)
 		goto error;
@@ -3192,12 +3187,6 @@ int __xfrm_init_state(struct xfrm_state *x, bool init_replay, bool offload,
 	}
 
 	x->outer_mode = *outer_mode;
-	if (init_replay) {
-		err = xfrm_init_replay(x, extack);
-		if (err)
-			goto error;
-	}
-
 	if (x->nat_keepalive_interval) {
 		if (x->dir != XFRM_SA_DIR_OUT) {
 			NL_SET_ERR_MSG(extack, "NAT keepalive is only supported for outbound SAs");
@@ -3229,11 +3218,16 @@ int xfrm_init_state(struct xfrm_state *x)
 {
 	int err;
 
-	err = __xfrm_init_state(x, true, false, NULL);
-	if (!err)
-		x->km.state = XFRM_STATE_VALID;
+	err = __xfrm_init_state(x, NULL);
+	if (err)
+		return err;
 
-	return err;
+	err = xfrm_init_replay(x, NULL);
+	if (err)
+		return err;
+
+	x->km.state = XFRM_STATE_VALID;
+	return 0;
 }
 
 EXPORT_SYMBOL(xfrm_init_state);
diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c
index 08c6d6f0179f..784a2d124749 100644
--- a/net/xfrm/xfrm_user.c
+++ b/net/xfrm/xfrm_user.c
@@ -178,6 +178,12 @@ static inline int verify_replay(struct xfrm_usersa_info *p,
 				       "Replay seq and seq_hi should be 0 for output SA");
 			return -EINVAL;
 		}
+		if (rs->oseq_hi && !(p->flags & XFRM_STATE_ESN)) {
+			NL_SET_ERR_MSG(
+				extack,
+				"Replay oseq_hi should be 0 in non-ESN mode for output SA");
+			return -EINVAL;
+		}
 		if (rs->bmp_len) {
 			NL_SET_ERR_MSG(extack, "Replay bmp_len should 0 for output SA");
 			return -EINVAL;
@@ -190,6 +196,12 @@ static inline int verify_replay(struct xfrm_usersa_info *p,
 				       "Replay oseq and oseq_hi should be 0 for input SA");
 			return -EINVAL;
 		}
+		if (rs->seq_hi && !(p->flags & XFRM_STATE_ESN)) {
+			NL_SET_ERR_MSG(
+				extack,
+				"Replay seq_hi should be 0 in non-ESN mode for input SA");
+			return -EINVAL;
+		}
 	}
 
 	return 0;
@@ -907,7 +919,7 @@ static struct xfrm_state *xfrm_state_construct(struct net *net,
 			goto error;
 	}
 
-	err = __xfrm_init_state(x, false, attrs[XFRMA_OFFLOAD_DEV], extack);
+	err = __xfrm_init_state(x, extack);
 	if (err)
 		goto error;