[net-next,v2,2/3] nfp: add framework to support ipsec offloading

Commit Message

Simon Horman Sept. 27, 2022, 10:27 a.m. UTC

From: Huanhuan Wang <[e-mail]>

A new metadata type and config structure are introduced to
interact with firmware to support ipsec offloading. This
feature relies on specific firmware that supports ipsec
encrypt/decrypt by advertising related capability bit.

The xfrm callbacks which interact with upper layer are
implemented in the following patch.

Based on initial work of Norm Bagley <[e-mail]>.

Signed-off-by: Huanhuan Wang <[e-mail]>
Signed-off-by: Simon Horman <[e-mail]>
---
 drivers/net/ethernet/netronome/Kconfig        |  11 +
 drivers/net/ethernet/netronome/nfp/Makefile   |   6 +
 .../ethernet/netronome/nfp/crypto/crypto.h    |  35 +++
 .../net/ethernet/netronome/nfp/crypto/ipsec.c | 216 ++++++++++++++++++
 drivers/net/ethernet/netronome/nfp/nfd3/dp.c  |  43 +++-
 .../net/ethernet/netronome/nfp/nfd3/ipsec.c   |  19 ++
 .../net/ethernet/netronome/nfp/nfd3/nfd3.h    |   8 +
 drivers/net/ethernet/netronome/nfp/nfp_net.h  |   9 +
 .../ethernet/netronome/nfp/nfp_net_common.c   |  10 +-
 .../net/ethernet/netronome/nfp/nfp_net_ctrl.h |   4 +
 10 files changed, 354 insertions(+), 7 deletions(-)
 create mode 100644 drivers/net/ethernet/netronome/nfp/crypto/ipsec.c
 create mode 100644 drivers/net/ethernet/netronome/nfp/nfd3/ipsec.c

Comments

Leon Romanovsky Sept. 29, 2022, 8:10 a.m. UTC | #1

On Tue, Sep 27, 2022 at 12:27:06PM +0200, Simon Horman wrote:
> From: Huanhuan Wang <[e-mail]>
> 
> A new metadata type and config structure are introduced to
> interact with firmware to support ipsec offloading. This
> feature relies on specific firmware that supports ipsec
> encrypt/decrypt by advertising related capability bit.
> 
> The xfrm callbacks which interact with upper layer are
> implemented in the following patch.
> 
> Based on initial work of Norm Bagley <[e-mail]>.
> 
> Signed-off-by: Huanhuan Wang <[e-mail]>
> Signed-off-by: Simon Horman <[e-mail]>
> ---
>  drivers/net/ethernet/netronome/Kconfig        |  11 +
>  drivers/net/ethernet/netronome/nfp/Makefile   |   6 +
>  .../ethernet/netronome/nfp/crypto/crypto.h    |  35 +++
>  .../net/ethernet/netronome/nfp/crypto/ipsec.c | 216 ++++++++++++++++++
>  drivers/net/ethernet/netronome/nfp/nfd3/dp.c  |  43 +++-
>  .../net/ethernet/netronome/nfp/nfd3/ipsec.c   |  19 ++
>  .../net/ethernet/netronome/nfp/nfd3/nfd3.h    |   8 +
>  drivers/net/ethernet/netronome/nfp/nfp_net.h  |   9 +
>  .../ethernet/netronome/nfp/nfp_net_common.c   |  10 +-
>  .../net/ethernet/netronome/nfp/nfp_net_ctrl.h |   4 +
>  10 files changed, 354 insertions(+), 7 deletions(-)
>  create mode 100644 drivers/net/ethernet/netronome/nfp/crypto/ipsec.c
>  create mode 100644 drivers/net/ethernet/netronome/nfp/nfd3/ipsec.c
> 
> diff --git a/drivers/net/ethernet/netronome/Kconfig b/drivers/net/ethernet/netronome/Kconfig
> index 8844d1ac053a..7f669d39e471 100644
> --- a/drivers/net/ethernet/netronome/Kconfig
> +++ b/drivers/net/ethernet/netronome/Kconfig
> @@ -54,6 +54,17 @@ config NFP_APP_ABM_NIC
>  	  functionality.
>  	  Code will be built into the nfp.ko driver.
>  
> +config NFP_NET_IPSEC
> +	bool "NFP Ipsec offload support"

You are adding "IPsec crypto offload", please be specific in all places.

> +	depends on NFP
> +	depends on XFRM_OFFLOAD
> +	default y
> +	help
> +	  Enable driver support Ipsec offload on NFP NIC. Say Y, if
> +	  you are planning to make use of Ipsec offload.
> +	  NOTE that Ipsec offload on NFP Nic requires specific FW to
> +	  work.
> +
>  config NFP_DEBUG
>  	bool "Debug support for Netronome(R) NFP4000/NFP6000 NIC drivers"
>  	depends on NFP
> diff --git a/drivers/net/ethernet/netronome/nfp/Makefile b/drivers/net/ethernet/netronome/nfp/Makefile
> index 9c0861d03634..3d33b2838e0d 100644
> --- a/drivers/net/ethernet/netronome/nfp/Makefile
> +++ b/drivers/net/ethernet/netronome/nfp/Makefile
> @@ -80,4 +80,10 @@ nfp-objs += \
>  	    abm/main.o
>  endif
>  
> +ifeq ($(CONFIG_NFP_NET_IPSEC),y)

Just curious, why ifeq and not nfp-$(...) format?

> +nfp-objs += \
> +	    crypto/ipsec.o \
> +	    nfd3/ipsec.o
> +endif
> +
>  nfp-$(CONFIG_NFP_DEBUG) += nfp_net_debugfs.o
> diff --git a/drivers/net/ethernet/netronome/nfp/crypto/crypto.h b/drivers/net/ethernet/netronome/nfp/crypto/crypto.h
> index bffe58bb2f27..a27a378e3ebe 100644
> --- a/drivers/net/ethernet/netronome/nfp/crypto/crypto.h
> +++ b/drivers/net/ethernet/netronome/nfp/crypto/crypto.h
> @@ -39,4 +39,39 @@ nfp_net_tls_rx_resync_req(struct net_device *netdev,
>  }
>  #endif
>  
> +/* Ipsec related structures and functions */
> +struct nfp_ipsec_offload {
> +	u32 seq_hi;
> +	u32 seq_low;
> +	u32 handle;
> +};
> +
> +#ifndef CONFIG_NFP_NET_IPSEC
> +static inline int nfp_net_ipsec_init(struct nfp_net *nn)
> +{
> +	return 0;
> +}
> +
> +static inline void nfp_net_ipsec_clean(struct nfp_net *nn)
> +{
> +}
> +
> +static inline bool nfp_net_ipsec_tx_prep(struct nfp_net_dp *dp, struct sk_buff *skb,
> +					 struct nfp_ipsec_offload *offload_info)
> +{
> +	return false;
> +}
> +
> +static inline int nfp_net_ipsec_rx(struct nfp_meta_parsed *meta, struct sk_buff *skb)
> +{
> +	return 0;
> +}
> +#else
> +int nfp_net_ipsec_init(struct nfp_net *nn);
> +void nfp_net_ipsec_clean(struct nfp_net *nn);
> +bool nfp_net_ipsec_tx_prep(struct nfp_net_dp *dp, struct sk_buff *skb,
> +			   struct nfp_ipsec_offload *offload_info);
> +int nfp_net_ipsec_rx(struct nfp_meta_parsed *meta, struct sk_buff *skb);
> +#endif
> +
>  #endif
> diff --git a/drivers/net/ethernet/netronome/nfp/crypto/ipsec.c b/drivers/net/ethernet/netronome/nfp/crypto/ipsec.c
> new file mode 100644
> index 000000000000..658fcba8e733
> --- /dev/null
> +++ b/drivers/net/ethernet/netronome/nfp/crypto/ipsec.c
> @@ -0,0 +1,216 @@
> +// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> +/* Copyright (C) 2018 Netronome Systems, Inc */
> +/* Copyright (C) 2021 Corigine, Inc */
> +
> +#include <linux/module.h>
> +#include <linux/kernel.h>
> +#include <linux/init.h>
> +#include <linux/netdevice.h>
> +#include <asm/unaligned.h>
> +#include <linux/ktime.h>
> +#include <net/xfrm.h>
> +
> +#include "../nfp_net_ctrl.h"
> +#include "../nfp_net.h"
> +#include "crypto.h"
> +
> +#define NFP_NET_IPSEC_MAX_SA_CNT  (16 * 1024) /* Firmware support a maximum of 16K sa offload */
> +#define OFFLOAD_HANDLE_ERROR      0xffffffff

You don't set this define in this patch, so why can't 0 be non-valid?

> +
> +/* IPSEC_CFG_MSSG_ADD_SA */
> +struct nfp_ipsec_cfg_add_sa {
> +	u32 ciph_key[8];		  /* Cipher Key */
> +	union {
> +		u32 auth_key[16];	  /* Authentication Key */
> +		struct nfp_ipsec_aesgcm { /* AES-GCM-ESP fields */
> +			u32 salt;	  /* Initialized with sa */
> +			u32 iv[2];	  /* Firmware use only */
> +			u32 cntr;
> +			u32 zeros[4];	  /* Init to 0 with sa */
> +			u32 len_a[2];	  /* Firmware use only */
> +			u32 len_c[2];
> +			u32 spare0[4];
> +		} aesgcm_fields;
> +	};
> +	struct sa_ctrl_word {
> +		uint32_t hash   :4;	  /* From nfp_ipsec_sa_hash_type */
> +		uint32_t cimode :4;	  /* From nfp_ipsec_sa_cipher_mode */
> +		uint32_t cipher :4;	  /* From nfp_ipsec_sa_cipher */
> +		uint32_t mode   :2;	  /* From nfp_ipsec_sa_mode */
> +		uint32_t proto  :2;	  /* From nfp_ipsec_sa_prot */
> +		uint32_t dir :1;	  /* Sa direction */
> +		uint32_t ena_arw:1;	  /* Anti-Replay Window */
> +		uint32_t ext_seq:1;	  /* 64-bit Sequence Num */
> +		uint32_t ext_arw:1;	  /* 64b Anti-Replay Window */
> +		uint32_t spare2 :9;	  /* Must be set to 0 */
> +		uint32_t encap_dsbl:1;	  /* Encap/decap disable */
> +		uint32_t gen_seq:1;	  /* Firmware Generate Seq */
> +		uint32_t spare8 :1;	  /* Must be set to 0 */
> +	} ctrl_word;
> +	u32 spi;			  /* SPI Value */
> +	uint32_t pmtu_limit :16;	  /* PMTU Limit */
> +	uint32_t spare3     :1;
> +	uint32_t frag_check :1;		  /* Stateful fragment checking flag */
> +	uint32_t bypass_DSCP:1;		  /* Bypass DSCP Flag */
> +	uint32_t df_ctrl    :2;		  /* DF Control bits */
> +	uint32_t ipv6       :1;		  /* Outbound IPv6 addr format */
> +	uint32_t udp_enable :1;		  /* Add/Remove UDP header for NAT */
> +	uint32_t tfc_enable :1;		  /* Traffic Flow Confidentiality */
> +	uint32_t spare4	 :8;
> +	u32 soft_lifetime_byte_count;
> +	u32 hard_lifetime_byte_count;
> +	u32 src_ip[4];			  /* Src IP addr */
> +	u32 dst_ip[4];			  /* Dst IP addr */
> +	uint32_t natt_dst_port :16;	  /* NAT-T UDP Header dst port */
> +	uint32_t natt_src_port :16;	  /* NAT-T UDP Header src port */
> +	u32 soft_lifetime_time_limit;
> +	u32 hard_lifetime_time_limit;
> +	u32 sa_creation_time_lo_32;	  /* Ucode fills this in */
> +	u32 sa_creation_time_hi_32;	  /* Ucode fills this in */
> +	uint32_t reserved0   :16;
> +	uint32_t tfc_padding :16;	  /* Traffic Flow Confidential Pad */
> +};
> +
> +struct nfp_net_ipsec_sa_data {
> +	struct nfp_ipsec_cfg_add_sa nfp_sa;
> +	struct xfrm_state *x;
> +	int invalidated;
> +};
> +
> +struct nfp_net_ipsec_data {
> +	struct nfp_net_ipsec_sa_data sa_entries[NFP_NET_IPSEC_MAX_SA_CNT];
> +	unsigned int sa_free_stack[NFP_NET_IPSEC_MAX_SA_CNT];
> +	unsigned int sa_free_cnt;

I don't see in this patch what are you doing with this free_stack array,
but whole nfp_net_ipsec_data is more than 32Kb of arrays.

> +	struct mutex lock;	/* Protects nfp_net_ipsec_data struct */
> +};
> +
> +static int nfp_net_xfrm_add_state(struct xfrm_state *x)
> +{
> +	return -EOPNOTSUPP;
> +}
> +
> +static void nfp_net_xfrm_del_state(struct xfrm_state *x)
> +{
> +}
> +
> +static void nfp_net_xfrm_free_state(struct xfrm_state *x)
> +{
> +}
> +
> +static bool nfp_net_ipsec_offload_ok(struct sk_buff *skb, struct xfrm_state *x)
> +{
> +	return false;
> +}
> +
> +static const struct xfrmdev_ops nfp_net_ipsec_xfrmdev_ops = {
> +	.xdo_dev_state_add = nfp_net_xfrm_add_state,
> +	.xdo_dev_state_delete = nfp_net_xfrm_del_state,
> +	.xdo_dev_state_free = nfp_net_xfrm_free_state,
> +	.xdo_dev_offload_ok = nfp_net_ipsec_offload_ok,
> +};
> +
> +int nfp_net_ipsec_init(struct nfp_net *nn)
> +{
> +	if (nn->cap_w1 & NFP_NET_CFG_CTRL_IPSEC) {

Success oriented flow, please

If (...)
  return 0;
....

> +		struct nfp_net_ipsec_data *ipd;
> +		int i;
> +
> +		nn->dp.netdev->xfrmdev_ops = &nfp_net_ipsec_xfrmdev_ops;

You set ops before nn->ipsec_data which requires from you always check
the validity of nn->ipsec_data in xfrm calls.

Please set it when this init can't fail and remove checks in your xfrm
callbacks, because XFRM will call for already initialized nfp ipsec
core.

> +
> +		ipd = kzalloc(sizeof(*ipd), GFP_KERNEL);
> +		if (!ipd)
> +			return -ENOMEM;
> +
> +		for (i = 0; i < NFP_NET_IPSEC_MAX_SA_CNT; i++)
> +			ipd->sa_free_stack[i] = NFP_NET_IPSEC_MAX_SA_CNT - i - 1;
> +
> +		ipd->sa_free_cnt = NFP_NET_IPSEC_MAX_SA_CNT;
> +		mutex_init(&ipd->lock);
> +		nn->ipsec_data = ipd;
> +	}
> +
> +	return 0;
> +}
> +
> +void nfp_net_ipsec_clean(struct nfp_net *nn)
> +{
> +	if (!nn->ipsec_data)
> +		return;
> +
> +	mutex_destroy(&nn->ipsec_data->lock);
> +	kfree(nn->ipsec_data);
> +	nn->ipsec_data = NULL;

Can you call again to nfp_net_ipsec_clean() after this line?
If not, please remove NULL assignment. It hides bugs.

> +}
> +
> +bool nfp_net_ipsec_tx_prep(struct nfp_net_dp *dp, struct sk_buff *skb,
> +			   struct nfp_ipsec_offload *offload_info)
> +{
> +	struct xfrm_offload *xo = xfrm_offload(skb);
> +	struct xfrm_state *x;
> +
> +	if (!xo)
> +		return false;

How is it possible in offload path?
Why do all drivers check sec_path length and not xo?

> +
> +	x = xfrm_input_state(skb);
> +	if (!x)
> +		return false;
> +
> +	if (x->xso.offload_handle == OFFLOAD_HANDLE_ERROR) {
> +		nn_dp_warn(dp, "Invalid xfrm offload handle\n");
> +		return false;
> +	}
> +
> +	offload_info->seq_hi = xo->seq.hi;
> +	offload_info->seq_low = xo->seq.low;
> +	offload_info->handle = x->xso.offload_handle;
> +
> +	return true;
> +}
> +
> +int nfp_net_ipsec_rx(struct nfp_meta_parsed *meta, struct sk_buff *skb)
> +{
> +	struct nfp_net_ipsec_sa_data *sa_data;
> +	struct net_device *netdev = skb->dev;
> +	struct nfp_net_ipsec_data *ipd;
> +	struct xfrm_offload *xo;
> +	struct nfp_net_dp *dp;
> +	struct xfrm_state *x;
> +	struct sec_path *sp;
> +	struct nfp_net *nn;
> +	int saidx;
> +
> +	nn = netdev_priv(netdev);
> +	ipd = nn->ipsec_data;
> +	dp = &nn->dp;
> +
> +	if (meta->ipsec_saidx == 0)
> +		return 0; /* No offload took place */
> +
> +	saidx = meta->ipsec_saidx - 1;
> +	if (saidx > NFP_NET_IPSEC_MAX_SA_CNT || saidx < 0) {
> +		nn_dp_warn(dp, "Invalid SAIDX from NIC %d\n", saidx);

No prints in data path that can be triggered from the network, please.

> +		return -EINVAL;
> +	}
> +
> +	sa_data = &ipd->sa_entries[saidx];
> +	if (!sa_data->x) {
> +		nn_dp_warn(dp, "Unused SAIDX from NIC %d\n", saidx);
> +		return -EINVAL;
> +	}
> +
> +	x = sa_data->x;
> +	xfrm_state_hold(x);

You should call to xfrm_state_hold() after secpath_set(), because in
failure of latter you will leak reference to state.

> +	sp = secpath_set(skb);
> +	if (unlikely(!sp)) {
> +		nn_dp_warn(dp, "Failed to alloc secpath for RX offload\n");
> +		return -ENOMEM;
> +	}
> +
> +	sp->xvec[sp->len++] = x;
> +	sp->olen++;
> +	xo = xfrm_offload(skb);
> +	xo->flags = CRYPTO_DONE;
> +	xo->status = CRYPTO_SUCCESS;
> +
> +	return 0;
> +}
> diff --git a/drivers/net/ethernet/netronome/nfp/nfd3/dp.c b/drivers/net/ethernet/netronome/nfp/nfd3/dp.c
> index 448c1c1afaee..db53e0392923 100644
> --- a/drivers/net/ethernet/netronome/nfp/nfd3/dp.c
> +++ b/drivers/net/ethernet/netronome/nfp/nfd3/dp.c
> @@ -167,14 +167,18 @@ nfp_nfd3_tx_csum(struct nfp_net_dp *dp, struct nfp_net_r_vector *r_vec,
>  	u64_stats_update_end(&r_vec->tx_sync);
>  }
>  
> -static int nfp_nfd3_prep_tx_meta(struct nfp_net_dp *dp, struct sk_buff *skb, u64 tls_handle)
> +static int nfp_nfd3_prep_tx_meta(struct nfp_net_dp *dp, struct sk_buff *skb,
> +				 u64 tls_handle, bool *ipsec)
>  {
>  	struct metadata_dst *md_dst = skb_metadata_dst(skb);
> +	struct nfp_ipsec_offload offload_info;
>  	unsigned char *data;
>  	bool vlan_insert;
>  	u32 meta_id = 0;
>  	int md_bytes;
>  
> +	*ipsec = nfp_net_ipsec_tx_prep(dp, skb, &offload_info);
> +
>  	if (unlikely(md_dst || tls_handle)) {
>  		if (unlikely(md_dst && md_dst->type != METADATA_HW_PORT_MUX))
>  			md_dst = NULL;
> @@ -182,13 +186,14 @@ static int nfp_nfd3_prep_tx_meta(struct nfp_net_dp *dp, struct sk_buff *skb, u64
>  
>  	vlan_insert = skb_vlan_tag_present(skb) && (dp->ctrl & NFP_NET_CFG_CTRL_TXVLAN_V2);
>  
> -	if (!(md_dst || tls_handle || vlan_insert))
> +	if (!(md_dst || tls_handle || vlan_insert || *ipsec))
>  		return 0;
>  
>  	md_bytes = sizeof(meta_id) +
>  		   !!md_dst * NFP_NET_META_PORTID_SIZE +
>  		   !!tls_handle * NFP_NET_META_CONN_HANDLE_SIZE +
> -		   vlan_insert * NFP_NET_META_VLAN_SIZE;
> +		   vlan_insert * NFP_NET_META_VLAN_SIZE +
> +		   *ipsec * NFP_NET_META_IPSEC_FIELD_SIZE; /* Ipsec has 12-bytes metadata */
>  
>  	if (unlikely(skb_cow_head(skb, md_bytes)))
>  		return -ENOMEM;
> @@ -218,6 +223,19 @@ static int nfp_nfd3_prep_tx_meta(struct nfp_net_dp *dp, struct sk_buff *skb, u64
>  		meta_id <<= NFP_NET_META_FIELD_SIZE;
>  		meta_id |= NFP_NET_META_VLAN;
>  	}
> +	if (*ipsec) {
> +		/* The ipsec has three consecutive 4-bit ipsec Metadate types
> +		 * so ipsec has three 4-bytes of Metadata
> +		 */
> +		data -= NFP_NET_META_IPSEC_SIZE;
> +		put_unaligned_be32(offload_info.seq_hi, data);
> +		data -= NFP_NET_META_IPSEC_SIZE;
> +		put_unaligned_be32(offload_info.seq_low, data);
> +		data -= NFP_NET_META_IPSEC_SIZE;
> +		put_unaligned_be32(offload_info.handle - 1, data);
> +		meta_id <<= NFP_NET_META_IPSEC_FIELD_SIZE;
> +		meta_id |= NFP_NET_META_IPSEC << 8 | NFP_NET_META_IPSEC << 4 | NFP_NET_META_IPSEC;
> +	}
>  
>  	data -= sizeof(meta_id);
>  	put_unaligned_be32(meta_id, data);
> @@ -246,6 +264,7 @@ netdev_tx_t nfp_nfd3_tx(struct sk_buff *skb, struct net_device *netdev)
>  	dma_addr_t dma_addr;
>  	unsigned int fsize;
>  	u64 tls_handle = 0;
> +	bool ipsec = false;
>  	u16 qidx;
>  
>  	dp = &nn->dp;
> @@ -273,7 +292,7 @@ netdev_tx_t nfp_nfd3_tx(struct sk_buff *skb, struct net_device *netdev)
>  		return NETDEV_TX_OK;
>  	}
>  
> -	md_bytes = nfp_nfd3_prep_tx_meta(dp, skb, tls_handle);
> +	md_bytes = nfp_nfd3_prep_tx_meta(dp, skb, tls_handle, &ipsec);
>  	if (unlikely(md_bytes < 0))
>  		goto err_flush;
>  
> @@ -312,6 +331,8 @@ netdev_tx_t nfp_nfd3_tx(struct sk_buff *skb, struct net_device *netdev)
>  		txd->vlan = cpu_to_le16(skb_vlan_tag_get(skb));
>  	}
>  
> +	if (ipsec)
> +		nfp_nfd3_ipsec_tx(txd, skb);
>  	/* Gather DMA */
>  	if (nr_frags > 0) {
>  		__le64 second_half;
> @@ -764,6 +785,12 @@ nfp_nfd3_parse_meta(struct net_device *netdev, struct nfp_meta_parsed *meta,
>  				return false;
>  			data += sizeof(struct nfp_net_tls_resync_req);
>  			break;
> +#ifdef CONFIG_NFP_NET_IPSEC
> +		case NFP_NET_META_IPSEC:
> +			meta->ipsec_saidx = get_unaligned_be32(data) + 1;
> +			data += 4;
> +			break;
> +#endif
>  		default:
>  			return true;
>  		}
> @@ -876,12 +903,11 @@ static int nfp_nfd3_rx(struct nfp_net_rx_ring *rx_ring, int budget)
>  	struct nfp_net_dp *dp = &r_vec->nfp_net->dp;
>  	struct nfp_net_tx_ring *tx_ring;
>  	struct bpf_prog *xdp_prog;
> +	int idx, pkts_polled = 0;
>  	bool xdp_tx_cmpl = false;
>  	unsigned int true_bufsz;
>  	struct sk_buff *skb;
> -	int pkts_polled = 0;
>  	struct xdp_buff xdp;
> -	int idx;
>  
>  	xdp_prog = READ_ONCE(dp->xdp_prog);
>  	true_bufsz = xdp_prog ? PAGE_SIZE : dp->fl_bufsz;
> @@ -1081,6 +1107,11 @@ static int nfp_nfd3_rx(struct nfp_net_rx_ring *rx_ring, int budget)
>  			continue;
>  		}
>  
> +		if (unlikely(nfp_net_ipsec_rx(&meta, skb))) {
> +			nfp_nfd3_rx_drop(dp, r_vec, rx_ring, NULL, skb);

Silent drop without any user visible counters, ok.

> +			continue;
> +		}
> +
>  		if (meta_len_xdp)
>  			skb_metadata_set(skb, meta_len_xdp);
>  
> diff --git a/drivers/net/ethernet/netronome/nfp/nfd3/ipsec.c b/drivers/net/ethernet/netronome/nfp/nfd3/ipsec.c
> new file mode 100644
> index 000000000000..f0d74d6b49d0
> --- /dev/null
> +++ b/drivers/net/ethernet/netronome/nfp/nfd3/ipsec.c
> @@ -0,0 +1,19 @@
> +// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> +/* Copyright (C) 2018 Netronome Systems, Inc */
> +/* Copyright (C) 2021 Corigine, Inc */
> +
> +#include <net/xfrm.h>
> +
> +#include "../nfp_net.h"
> +#include "nfd3.h"
> +
> +void nfp_nfd3_ipsec_tx(struct nfp_nfd3_tx_desc *txd, struct sk_buff *skb)
> +{
> +	struct xfrm_state *x;
> +
> +	x = xfrm_input_state(skb);
> +	if (x->xso.dev && (x->xso.dev->features & NETIF_F_HW_ESP_TX_CSUM)) {
> +		txd->flags |= NFD3_DESC_TX_CSUM | NFD3_DESC_TX_IP4_CSUM |
> +			      NFD3_DESC_TX_TCP_CSUM | NFD3_DESC_TX_UDP_CSUM;
> +	}
> +}
> diff --git a/drivers/net/ethernet/netronome/nfp/nfd3/nfd3.h b/drivers/net/ethernet/netronome/nfp/nfd3/nfd3.h
> index 7a0df9e6c3c4..9c1c10dcbaee 100644
> --- a/drivers/net/ethernet/netronome/nfp/nfd3/nfd3.h
> +++ b/drivers/net/ethernet/netronome/nfp/nfd3/nfd3.h
> @@ -103,4 +103,12 @@ void nfp_nfd3_rx_ring_fill_freelist(struct nfp_net_dp *dp,
>  void nfp_nfd3_xsk_tx_free(struct nfp_nfd3_tx_buf *txbuf);
>  int nfp_nfd3_xsk_poll(struct napi_struct *napi, int budget);
>  
> +#ifndef CONFIG_NFP_NET_IPSEC
> +static inline void nfp_nfd3_ipsec_tx(struct nfp_nfd3_tx_desc *txd, struct sk_buff *skb)
> +{
> +}
> +#else
> +void nfp_nfd3_ipsec_tx(struct nfp_nfd3_tx_desc *txd, struct sk_buff *skb);
> +#endif
> +
>  #endif
> diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net.h b/drivers/net/ethernet/netronome/nfp/nfp_net.h
> index 0c3e7e2f856d..768539f12214 100644
> --- a/drivers/net/ethernet/netronome/nfp/nfp_net.h
> +++ b/drivers/net/ethernet/netronome/nfp/nfp_net.h
> @@ -263,6 +263,10 @@ struct nfp_meta_parsed {
>  		u8 tpid;
>  		u16 tci;
>  	} vlan;
> +
> +#ifdef CONFIG_NFP_NET_IPSEC
> +	u32 ipsec_saidx;
> +#endif
>  };
>  
>  struct nfp_net_rx_hash {
> @@ -584,6 +588,7 @@ struct nfp_net_dp {
>   * @qcp_cfg:            Pointer to QCP queue used for configuration notification
>   * @tx_bar:             Pointer to mapped TX queues
>   * @rx_bar:             Pointer to mapped FL/RX queues
> + * @ipsec_data:         Ipsec Sa data
>   * @tlv_caps:		Parsed TLV capabilities
>   * @ktls_tx_conn_cnt:	Number of offloaded kTLS TX connections
>   * @ktls_rx_conn_cnt:	Number of offloaded kTLS RX connections
> @@ -672,6 +677,10 @@ struct nfp_net {
>  	u8 __iomem *tx_bar;
>  	u8 __iomem *rx_bar;
>  
> +#ifdef CONFIG_NFP_NET_IPSEC
> +	struct nfp_net_ipsec_data *ipsec_data;
> +#endif
> +
>  	struct nfp_net_tlv_caps tlv_caps;
>  
>  	unsigned int ktls_tx_conn_cnt;
> diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
> index 7e4424d626a6..040c0c2aad80 100644
> --- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
> +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
> @@ -2565,9 +2565,13 @@ int nfp_net_init(struct nfp_net *nn)
>  		if (err)
>  			return err;
>  
> -		err = nfp_net_tls_init(nn);
> +		err = nfp_net_ipsec_init(nn);
>  		if (err)
>  			goto err_clean_mbox;
> +
> +		err = nfp_net_tls_init(nn);
> +		if (err)
> +			goto err_clean_ipsec;
>  	}
>  
>  	nfp_net_vecs_init(nn);
> @@ -2576,6 +2580,9 @@ int nfp_net_init(struct nfp_net *nn)
>  		return 0;
>  	return register_netdev(nn->dp.netdev);
>  
> +err_clean_ipsec:
> +	nfp_net_ipsec_clean(nn);
> +
>  err_clean_mbox:
>  	nfp_ccm_mbox_clean(nn);
>  	return err;
> @@ -2591,6 +2598,7 @@ void nfp_net_clean(struct nfp_net *nn)
>  		return;
>  
>  	unregister_netdev(nn->dp.netdev);
> +	nfp_net_ipsec_clean(nn);
>  	nfp_ccm_mbox_clean(nn);
>  	nfp_net_reconfig_wait_posted(nn);
>  }
> diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h b/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h
> index 80346c1c266b..fff05496152d 100644
> --- a/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h
> +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h
> @@ -45,6 +45,7 @@
>  #define NFP_NET_META_CSUM		6 /* checksum complete type */
>  #define NFP_NET_META_CONN_HANDLE	7
>  #define NFP_NET_META_RESYNC_INFO	8 /* RX resync info request */
> +#define NFP_NET_META_IPSEC		9 /* IPSec SA index for tx and rx */
>  
>  #define NFP_META_PORT_ID_CTRL		~0U
>  
> @@ -52,6 +53,8 @@
>  #define NFP_NET_META_VLAN_SIZE			4
>  #define NFP_NET_META_PORTID_SIZE		4
>  #define NFP_NET_META_CONN_HANDLE_SIZE		8
> +#define NFP_NET_META_IPSEC_SIZE			4
> +#define NFP_NET_META_IPSEC_FIELD_SIZE		12
>  /* Hash type pre-pended when a RSS hash was computed */
>  #define NFP_NET_RSS_NONE		0
>  #define NFP_NET_RSS_IPV4		1
> @@ -260,6 +263,7 @@
>   */
>  #define NFP_NET_CFG_CTRL_WORD1		0x0098
>  #define   NFP_NET_CFG_CTRL_PKT_TYPE	  (0x1 << 0) /* Pkttype offload */
> +#define   NFP_NET_CFG_CTRL_IPSEC	  (0x1 << 1) /* IPSec offload */
>  
>  #define NFP_NET_CFG_CAP_WORD1		0x00a4
>  
> -- 
> 2.30.2
>

Leon Romanovsky Oct. 11, 2022, 8:41 a.m. UTC | #2

On Mon, Oct 10, 2022 at 03:05:12PM +0800, Yinjun Zhang wrote:
> Thanks for your comments and sorry for the late reply.
> 
> On Thu, Sep 29, 2022 at 11:10:12AM +0300, Leon Romanovsky wrote:
> > On Tue, Sep 27, 2022 at 12:27:06PM +0200, Simon Horman wrote:
> > > +struct nfp_net_ipsec_data {
> > > +	struct nfp_net_ipsec_sa_data sa_entries[NFP_NET_IPSEC_MAX_SA_CNT];
> > > +	unsigned int sa_free_stack[NFP_NET_IPSEC_MAX_SA_CNT];
> > > +	unsigned int sa_free_cnt;
> > 
> > I don't see in this patch what are you doing with this free_stack array,
> > but whole nfp_net_ipsec_data is more than 32Kb of arrays.
> >
> 
> `sa_free_stack` is used to maintain the used/available sa entries, which
> is initialized in `nfp_net_ipsec_init`.
> Yes, it's indeed a big array, and we're going to use pointer instead of array
> here.

Why do you want to use array and not Xarray?

> 
> > > +bool nfp_net_ipsec_tx_prep(struct nfp_net_dp *dp, struct sk_buff *skb,
> > > +			   struct nfp_ipsec_offload *offload_info)
> > > +{
> > > +	struct xfrm_offload *xo = xfrm_offload(skb);
> > > +	struct xfrm_state *x;
> > > +
> > > +	if (!xo)
> > > +		return false;
> > 
> > How is it possible in offload path?
> > Why do all drivers check sec_path length and not xo?
> > 
> 
> `tx_prep` is called in the tx datapath, we use `xo` to check if the
> packet needs offload-encrypto or not.

You didn't answer on any of my questions above.

How is it possible in offload path?
Why do all drivers check sec_path length and not xo?

> 
> > > +int nfp_net_ipsec_rx(struct nfp_meta_parsed *meta, struct sk_buff *skb)
> > > +{
> > > +	struct nfp_net_ipsec_sa_data *sa_data;
> > > +	struct net_device *netdev = skb->dev;
> > > +	struct nfp_net_ipsec_data *ipd;
> > > +	struct xfrm_offload *xo;
> > > +	struct nfp_net_dp *dp;
> > > +	struct xfrm_state *x;
> > > +	struct sec_path *sp;
> > > +	struct nfp_net *nn;
> > > +	int saidx;
> > > +
> > > +	nn = netdev_priv(netdev);
> > > +	ipd = nn->ipsec_data;
> > > +	dp = &nn->dp;
> > > +
> > > +	if (meta->ipsec_saidx == 0)
> > > +		return 0; /* No offload took place */
> > > +
> > > +	saidx = meta->ipsec_saidx - 1;
> > > +	if (saidx > NFP_NET_IPSEC_MAX_SA_CNT || saidx < 0) {
> > > +		nn_dp_warn(dp, "Invalid SAIDX from NIC %d\n", saidx);
> > 
> > No prints in data path that can be triggered from the network, please.
> > 
> 
> It's a ratelimit print, and it means severe error happens, probably
> unrecoverable, when running into this path.

The main part of the sentence is "... can be triggered from the network ..."

Thanks

Yinjun Zhang Oct. 13, 2022, 8:21 a.m. UTC | #3

On Tue, Oct 11, 2022 at 11:41:47AM +0300, Leon Romanovsky wrote:
> > 
> > `sa_free_stack` is used to maintain the used/available sa entries, which
> > is initialized in `nfp_net_ipsec_init`.
> > Yes, it's indeed a big array, and we're going to use pointer instead of array
> > here.
> 
> Why do you want to use array and not Xarray?
> 

We'll try that. Thanks.

> > 
> > > > +bool nfp_net_ipsec_tx_prep(struct nfp_net_dp *dp, struct sk_buff *skb,
> > > > +			   struct nfp_ipsec_offload *offload_info)
> > > > +{
> > > > +	struct xfrm_offload *xo = xfrm_offload(skb);
> > > > +	struct xfrm_state *x;
> > > > +
> > > > +	if (!xo)
> > > > +		return false;
> > > 
> > > How is it possible in offload path?
> > > Why do all drivers check sec_path length and not xo?
> > > 
> > 
> > `tx_prep` is called in the tx datapath, we use `xo` to check if the
> > packet needs offload-encrypto or not.
> 
> You didn't answer on any of my questions above.
> 
> How is it possible in offload path?
> Why do all drivers check sec_path length and not xo?
> 

It's not a offload-only path, but a normal tx data path. Only if xo is
not NULL, we're going to do crypto-offload. Not every transimitted
packet needs crypto, right? 
We're going to move this check to its parent function to avoid
unnecessary jump when it's not a to-crypto packet.

> > 
> > > > +	saidx = meta->ipsec_saidx - 1;
> > > > +	if (saidx > NFP_NET_IPSEC_MAX_SA_CNT || saidx < 0) {
> > > > +		nn_dp_warn(dp, "Invalid SAIDX from NIC %d\n", saidx);
> > > 
> > > No prints in data path that can be triggered from the network, please.
> > > 
> > 
> > It's a ratelimit print, and it means severe error happens, probably
> > unrecoverable, when running into this path.
> 
> The main part of the sentence is "... can be triggered from the network ..."

OK, we'll remove this print, and maybe introduce some error counters instead in
following patch series.

> 
> Thanks

Patch

diff --git a/drivers/net/ethernet/netronome/Kconfig b/drivers/net/ethernet/netronome/Kconfig
index 8844d1ac053a..7f669d39e471 100644
--- a/drivers/net/ethernet/netronome/Kconfig
+++ b/drivers/net/ethernet/netronome/Kconfig
@@ -54,6 +54,17 @@  config NFP_APP_ABM_NIC
 	  functionality.
 	  Code will be built into the nfp.ko driver.
 
+config NFP_NET_IPSEC
+	bool "NFP Ipsec offload support"
+	depends on NFP
+	depends on XFRM_OFFLOAD
+	default y
+	help
+	  Enable driver support Ipsec offload on NFP NIC. Say Y, if
+	  you are planning to make use of Ipsec offload.
+	  NOTE that Ipsec offload on NFP Nic requires specific FW to
+	  work.
+
 config NFP_DEBUG
 	bool "Debug support for Netronome(R) NFP4000/NFP6000 NIC drivers"
 	depends on NFP
diff --git a/drivers/net/ethernet/netronome/nfp/Makefile b/drivers/net/ethernet/netronome/nfp/Makefile
index 9c0861d03634..3d33b2838e0d 100644
--- a/drivers/net/ethernet/netronome/nfp/Makefile
+++ b/drivers/net/ethernet/netronome/nfp/Makefile
@@ -80,4 +80,10 @@  nfp-objs += \
 	    abm/main.o
 endif
 
+ifeq ($(CONFIG_NFP_NET_IPSEC),y)
+nfp-objs += \
+	    crypto/ipsec.o \
+	    nfd3/ipsec.o
+endif
+
 nfp-$(CONFIG_NFP_DEBUG) += nfp_net_debugfs.o
diff --git a/drivers/net/ethernet/netronome/nfp/crypto/crypto.h b/drivers/net/ethernet/netronome/nfp/crypto/crypto.h
index bffe58bb2f27..a27a378e3ebe 100644
--- a/drivers/net/ethernet/netronome/nfp/crypto/crypto.h
+++ b/drivers/net/ethernet/netronome/nfp/crypto/crypto.h
@@ -39,4 +39,39 @@  nfp_net_tls_rx_resync_req(struct net_device *netdev,
 }
 #endif
 
+/* Ipsec related structures and functions */
+struct nfp_ipsec_offload {
+	u32 seq_hi;
+	u32 seq_low;
+	u32 handle;
+};
+
+#ifndef CONFIG_NFP_NET_IPSEC
+static inline int nfp_net_ipsec_init(struct nfp_net *nn)
+{
+	return 0;
+}
+
+static inline void nfp_net_ipsec_clean(struct nfp_net *nn)
+{
+}
+
+static inline bool nfp_net_ipsec_tx_prep(struct nfp_net_dp *dp, struct sk_buff *skb,
+					 struct nfp_ipsec_offload *offload_info)
+{
+	return false;
+}
+
+static inline int nfp_net_ipsec_rx(struct nfp_meta_parsed *meta, struct sk_buff *skb)
+{
+	return 0;
+}
+#else
+int nfp_net_ipsec_init(struct nfp_net *nn);
+void nfp_net_ipsec_clean(struct nfp_net *nn);
+bool nfp_net_ipsec_tx_prep(struct nfp_net_dp *dp, struct sk_buff *skb,
+			   struct nfp_ipsec_offload *offload_info);
+int nfp_net_ipsec_rx(struct nfp_meta_parsed *meta, struct sk_buff *skb);
+#endif
+
 #endif
diff --git a/drivers/net/ethernet/netronome/nfp/crypto/ipsec.c b/drivers/net/ethernet/netronome/nfp/crypto/ipsec.c
new file mode 100644
index 000000000000..658fcba8e733
--- /dev/null
+++ b/drivers/net/ethernet/netronome/nfp/crypto/ipsec.c
@@ -0,0 +1,216 @@ 
+// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+/* Copyright (C) 2018 Netronome Systems, Inc */
+/* Copyright (C) 2021 Corigine, Inc */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/netdevice.h>
+#include <asm/unaligned.h>
+#include <linux/ktime.h>
+#include <net/xfrm.h>
+
+#include "../nfp_net_ctrl.h"
+#include "../nfp_net.h"
+#include "crypto.h"
+
+#define NFP_NET_IPSEC_MAX_SA_CNT  (16 * 1024) /* Firmware support a maximum of 16K sa offload */
+#define OFFLOAD_HANDLE_ERROR      0xffffffff
+
+/* IPSEC_CFG_MSSG_ADD_SA */
+struct nfp_ipsec_cfg_add_sa {
+	u32 ciph_key[8];		  /* Cipher Key */
+	union {
+		u32 auth_key[16];	  /* Authentication Key */
+		struct nfp_ipsec_aesgcm { /* AES-GCM-ESP fields */
+			u32 salt;	  /* Initialized with sa */
+			u32 iv[2];	  /* Firmware use only */
+			u32 cntr;
+			u32 zeros[4];	  /* Init to 0 with sa */
+			u32 len_a[2];	  /* Firmware use only */
+			u32 len_c[2];
+			u32 spare0[4];
+		} aesgcm_fields;
+	};
+	struct sa_ctrl_word {
+		uint32_t hash   :4;	  /* From nfp_ipsec_sa_hash_type */
+		uint32_t cimode :4;	  /* From nfp_ipsec_sa_cipher_mode */
+		uint32_t cipher :4;	  /* From nfp_ipsec_sa_cipher */
+		uint32_t mode   :2;	  /* From nfp_ipsec_sa_mode */
+		uint32_t proto  :2;	  /* From nfp_ipsec_sa_prot */
+		uint32_t dir :1;	  /* Sa direction */
+		uint32_t ena_arw:1;	  /* Anti-Replay Window */
+		uint32_t ext_seq:1;	  /* 64-bit Sequence Num */
+		uint32_t ext_arw:1;	  /* 64b Anti-Replay Window */
+		uint32_t spare2 :9;	  /* Must be set to 0 */
+		uint32_t encap_dsbl:1;	  /* Encap/decap disable */
+		uint32_t gen_seq:1;	  /* Firmware Generate Seq */
+		uint32_t spare8 :1;	  /* Must be set to 0 */
+	} ctrl_word;
+	u32 spi;			  /* SPI Value */
+	uint32_t pmtu_limit :16;	  /* PMTU Limit */
+	uint32_t spare3     :1;
+	uint32_t frag_check :1;		  /* Stateful fragment checking flag */
+	uint32_t bypass_DSCP:1;		  /* Bypass DSCP Flag */
+	uint32_t df_ctrl    :2;		  /* DF Control bits */
+	uint32_t ipv6       :1;		  /* Outbound IPv6 addr format */
+	uint32_t udp_enable :1;		  /* Add/Remove UDP header for NAT */
+	uint32_t tfc_enable :1;		  /* Traffic Flow Confidentiality */
+	uint32_t spare4	 :8;
+	u32 soft_lifetime_byte_count;
+	u32 hard_lifetime_byte_count;
+	u32 src_ip[4];			  /* Src IP addr */
+	u32 dst_ip[4];			  /* Dst IP addr */
+	uint32_t natt_dst_port :16;	  /* NAT-T UDP Header dst port */
+	uint32_t natt_src_port :16;	  /* NAT-T UDP Header src port */
+	u32 soft_lifetime_time_limit;
+	u32 hard_lifetime_time_limit;
+	u32 sa_creation_time_lo_32;	  /* Ucode fills this in */
+	u32 sa_creation_time_hi_32;	  /* Ucode fills this in */
+	uint32_t reserved0   :16;
+	uint32_t tfc_padding :16;	  /* Traffic Flow Confidential Pad */
+};
+
+struct nfp_net_ipsec_sa_data {
+	struct nfp_ipsec_cfg_add_sa nfp_sa;
+	struct xfrm_state *x;
+	int invalidated;
+};
+
+struct nfp_net_ipsec_data {
+	struct nfp_net_ipsec_sa_data sa_entries[NFP_NET_IPSEC_MAX_SA_CNT];
+	unsigned int sa_free_stack[NFP_NET_IPSEC_MAX_SA_CNT];
+	unsigned int sa_free_cnt;
+	struct mutex lock;	/* Protects nfp_net_ipsec_data struct */
+};
+
+static int nfp_net_xfrm_add_state(struct xfrm_state *x)
+{
+	return -EOPNOTSUPP;
+}
+
+static void nfp_net_xfrm_del_state(struct xfrm_state *x)
+{
+}
+
+static void nfp_net_xfrm_free_state(struct xfrm_state *x)
+{
+}
+
+static bool nfp_net_ipsec_offload_ok(struct sk_buff *skb, struct xfrm_state *x)
+{
+	return false;
+}
+
+static const struct xfrmdev_ops nfp_net_ipsec_xfrmdev_ops = {
+	.xdo_dev_state_add = nfp_net_xfrm_add_state,
+	.xdo_dev_state_delete = nfp_net_xfrm_del_state,
+	.xdo_dev_state_free = nfp_net_xfrm_free_state,
+	.xdo_dev_offload_ok = nfp_net_ipsec_offload_ok,
+};
+
+int nfp_net_ipsec_init(struct nfp_net *nn)
+{
+	if (nn->cap_w1 & NFP_NET_CFG_CTRL_IPSEC) {
+		struct nfp_net_ipsec_data *ipd;
+		int i;
+
+		nn->dp.netdev->xfrmdev_ops = &nfp_net_ipsec_xfrmdev_ops;
+
+		ipd = kzalloc(sizeof(*ipd), GFP_KERNEL);
+		if (!ipd)
+			return -ENOMEM;
+
+		for (i = 0; i < NFP_NET_IPSEC_MAX_SA_CNT; i++)
+			ipd->sa_free_stack[i] = NFP_NET_IPSEC_MAX_SA_CNT - i - 1;
+
+		ipd->sa_free_cnt = NFP_NET_IPSEC_MAX_SA_CNT;
+		mutex_init(&ipd->lock);
+		nn->ipsec_data = ipd;
+	}
+
+	return 0;
+}
+
+void nfp_net_ipsec_clean(struct nfp_net *nn)
+{
+	if (!nn->ipsec_data)
+		return;
+
+	mutex_destroy(&nn->ipsec_data->lock);
+	kfree(nn->ipsec_data);
+	nn->ipsec_data = NULL;
+}
+
+bool nfp_net_ipsec_tx_prep(struct nfp_net_dp *dp, struct sk_buff *skb,
+			   struct nfp_ipsec_offload *offload_info)
+{
+	struct xfrm_offload *xo = xfrm_offload(skb);
+	struct xfrm_state *x;
+
+	if (!xo)
+		return false;
+
+	x = xfrm_input_state(skb);
+	if (!x)
+		return false;
+
+	if (x->xso.offload_handle == OFFLOAD_HANDLE_ERROR) {
+		nn_dp_warn(dp, "Invalid xfrm offload handle\n");
+		return false;
+	}
+
+	offload_info->seq_hi = xo->seq.hi;
+	offload_info->seq_low = xo->seq.low;
+	offload_info->handle = x->xso.offload_handle;
+
+	return true;
+}
+
+int nfp_net_ipsec_rx(struct nfp_meta_parsed *meta, struct sk_buff *skb)
+{
+	struct nfp_net_ipsec_sa_data *sa_data;
+	struct net_device *netdev = skb->dev;
+	struct nfp_net_ipsec_data *ipd;
+	struct xfrm_offload *xo;
+	struct nfp_net_dp *dp;
+	struct xfrm_state *x;
+	struct sec_path *sp;
+	struct nfp_net *nn;
+	int saidx;
+
+	nn = netdev_priv(netdev);
+	ipd = nn->ipsec_data;
+	dp = &nn->dp;
+
+	if (meta->ipsec_saidx == 0)
+		return 0; /* No offload took place */
+
+	saidx = meta->ipsec_saidx - 1;
+	if (saidx > NFP_NET_IPSEC_MAX_SA_CNT || saidx < 0) {
+		nn_dp_warn(dp, "Invalid SAIDX from NIC %d\n", saidx);
+		return -EINVAL;
+	}
+
+	sa_data = &ipd->sa_entries[saidx];
+	if (!sa_data->x) {
+		nn_dp_warn(dp, "Unused SAIDX from NIC %d\n", saidx);
+		return -EINVAL;
+	}
+
+	x = sa_data->x;
+	xfrm_state_hold(x);
+	sp = secpath_set(skb);
+	if (unlikely(!sp)) {
+		nn_dp_warn(dp, "Failed to alloc secpath for RX offload\n");
+		return -ENOMEM;
+	}
+
+	sp->xvec[sp->len++] = x;
+	sp->olen++;
+	xo = xfrm_offload(skb);
+	xo->flags = CRYPTO_DONE;
+	xo->status = CRYPTO_SUCCESS;
+
+	return 0;
+}
diff --git a/drivers/net/ethernet/netronome/nfp/nfd3/dp.c b/drivers/net/ethernet/netronome/nfp/nfd3/dp.c
index 448c1c1afaee..db53e0392923 100644
--- a/drivers/net/ethernet/netronome/nfp/nfd3/dp.c
+++ b/drivers/net/ethernet/netronome/nfp/nfd3/dp.c
@@ -167,14 +167,18 @@  nfp_nfd3_tx_csum(struct nfp_net_dp *dp, struct nfp_net_r_vector *r_vec,
 	u64_stats_update_end(&r_vec->tx_sync);
 }
 
-static int nfp_nfd3_prep_tx_meta(struct nfp_net_dp *dp, struct sk_buff *skb, u64 tls_handle)
+static int nfp_nfd3_prep_tx_meta(struct nfp_net_dp *dp, struct sk_buff *skb,
+				 u64 tls_handle, bool *ipsec)
 {
 	struct metadata_dst *md_dst = skb_metadata_dst(skb);
+	struct nfp_ipsec_offload offload_info;
 	unsigned char *data;
 	bool vlan_insert;
 	u32 meta_id = 0;
 	int md_bytes;
 
+	*ipsec = nfp_net_ipsec_tx_prep(dp, skb, &offload_info);
+
 	if (unlikely(md_dst || tls_handle)) {
 		if (unlikely(md_dst && md_dst->type != METADATA_HW_PORT_MUX))
 			md_dst = NULL;
@@ -182,13 +186,14 @@  static int nfp_nfd3_prep_tx_meta(struct nfp_net_dp *dp, struct sk_buff *skb, u64
 
 	vlan_insert = skb_vlan_tag_present(skb) && (dp->ctrl & NFP_NET_CFG_CTRL_TXVLAN_V2);
 
-	if (!(md_dst || tls_handle || vlan_insert))
+	if (!(md_dst || tls_handle || vlan_insert || *ipsec))
 		return 0;
 
 	md_bytes = sizeof(meta_id) +
 		   !!md_dst * NFP_NET_META_PORTID_SIZE +
 		   !!tls_handle * NFP_NET_META_CONN_HANDLE_SIZE +
-		   vlan_insert * NFP_NET_META_VLAN_SIZE;
+		   vlan_insert * NFP_NET_META_VLAN_SIZE +
+		   *ipsec * NFP_NET_META_IPSEC_FIELD_SIZE; /* Ipsec has 12-bytes metadata */
 
 	if (unlikely(skb_cow_head(skb, md_bytes)))
 		return -ENOMEM;
@@ -218,6 +223,19 @@  static int nfp_nfd3_prep_tx_meta(struct nfp_net_dp *dp, struct sk_buff *skb, u64
 		meta_id <<= NFP_NET_META_FIELD_SIZE;
 		meta_id |= NFP_NET_META_VLAN;
 	}
+	if (*ipsec) {
+		/* The ipsec has three consecutive 4-bit ipsec Metadate types
+		 * so ipsec has three 4-bytes of Metadata
+		 */
+		data -= NFP_NET_META_IPSEC_SIZE;
+		put_unaligned_be32(offload_info.seq_hi, data);
+		data -= NFP_NET_META_IPSEC_SIZE;
+		put_unaligned_be32(offload_info.seq_low, data);
+		data -= NFP_NET_META_IPSEC_SIZE;
+		put_unaligned_be32(offload_info.handle - 1, data);
+		meta_id <<= NFP_NET_META_IPSEC_FIELD_SIZE;
+		meta_id |= NFP_NET_META_IPSEC << 8 | NFP_NET_META_IPSEC << 4 | NFP_NET_META_IPSEC;
+	}
 
 	data -= sizeof(meta_id);
 	put_unaligned_be32(meta_id, data);
@@ -246,6 +264,7 @@  netdev_tx_t nfp_nfd3_tx(struct sk_buff *skb, struct net_device *netdev)
 	dma_addr_t dma_addr;
 	unsigned int fsize;
 	u64 tls_handle = 0;
+	bool ipsec = false;
 	u16 qidx;
 
 	dp = &nn->dp;
@@ -273,7 +292,7 @@  netdev_tx_t nfp_nfd3_tx(struct sk_buff *skb, struct net_device *netdev)
 		return NETDEV_TX_OK;
 	}
 
-	md_bytes = nfp_nfd3_prep_tx_meta(dp, skb, tls_handle);
+	md_bytes = nfp_nfd3_prep_tx_meta(dp, skb, tls_handle, &ipsec);
 	if (unlikely(md_bytes < 0))
 		goto err_flush;
 
@@ -312,6 +331,8 @@  netdev_tx_t nfp_nfd3_tx(struct sk_buff *skb, struct net_device *netdev)
 		txd->vlan = cpu_to_le16(skb_vlan_tag_get(skb));
 	}
 
+	if (ipsec)
+		nfp_nfd3_ipsec_tx(txd, skb);
 	/* Gather DMA */
 	if (nr_frags > 0) {
 		__le64 second_half;
@@ -764,6 +785,12 @@  nfp_nfd3_parse_meta(struct net_device *netdev, struct nfp_meta_parsed *meta,
 				return false;
 			data += sizeof(struct nfp_net_tls_resync_req);
 			break;
+#ifdef CONFIG_NFP_NET_IPSEC
+		case NFP_NET_META_IPSEC:
+			meta->ipsec_saidx = get_unaligned_be32(data) + 1;
+			data += 4;
+			break;
+#endif
 		default:
 			return true;
 		}
@@ -876,12 +903,11 @@  static int nfp_nfd3_rx(struct nfp_net_rx_ring *rx_ring, int budget)
 	struct nfp_net_dp *dp = &r_vec->nfp_net->dp;
 	struct nfp_net_tx_ring *tx_ring;
 	struct bpf_prog *xdp_prog;
+	int idx, pkts_polled = 0;
 	bool xdp_tx_cmpl = false;
 	unsigned int true_bufsz;
 	struct sk_buff *skb;
-	int pkts_polled = 0;
 	struct xdp_buff xdp;
-	int idx;
 
 	xdp_prog = READ_ONCE(dp->xdp_prog);
 	true_bufsz = xdp_prog ? PAGE_SIZE : dp->fl_bufsz;
@@ -1081,6 +1107,11 @@  static int nfp_nfd3_rx(struct nfp_net_rx_ring *rx_ring, int budget)
 			continue;
 		}
 
+		if (unlikely(nfp_net_ipsec_rx(&meta, skb))) {
+			nfp_nfd3_rx_drop(dp, r_vec, rx_ring, NULL, skb);
+			continue;
+		}
+
 		if (meta_len_xdp)
 			skb_metadata_set(skb, meta_len_xdp);
 
diff --git a/drivers/net/ethernet/netronome/nfp/nfd3/ipsec.c b/drivers/net/ethernet/netronome/nfp/nfd3/ipsec.c
new file mode 100644
index 000000000000..f0d74d6b49d0
--- /dev/null
+++ b/drivers/net/ethernet/netronome/nfp/nfd3/ipsec.c
@@ -0,0 +1,19 @@ 
+// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+/* Copyright (C) 2018 Netronome Systems, Inc */
+/* Copyright (C) 2021 Corigine, Inc */
+
+#include <net/xfrm.h>
+
+#include "../nfp_net.h"
+#include "nfd3.h"
+
+void nfp_nfd3_ipsec_tx(struct nfp_nfd3_tx_desc *txd, struct sk_buff *skb)
+{
+	struct xfrm_state *x;
+
+	x = xfrm_input_state(skb);
+	if (x->xso.dev && (x->xso.dev->features & NETIF_F_HW_ESP_TX_CSUM)) {
+		txd->flags |= NFD3_DESC_TX_CSUM | NFD3_DESC_TX_IP4_CSUM |
+			      NFD3_DESC_TX_TCP_CSUM | NFD3_DESC_TX_UDP_CSUM;
+	}
+}
diff --git a/drivers/net/ethernet/netronome/nfp/nfd3/nfd3.h b/drivers/net/ethernet/netronome/nfp/nfd3/nfd3.h
index 7a0df9e6c3c4..9c1c10dcbaee 100644
--- a/drivers/net/ethernet/netronome/nfp/nfd3/nfd3.h
+++ b/drivers/net/ethernet/netronome/nfp/nfd3/nfd3.h
@@ -103,4 +103,12 @@  void nfp_nfd3_rx_ring_fill_freelist(struct nfp_net_dp *dp,
 void nfp_nfd3_xsk_tx_free(struct nfp_nfd3_tx_buf *txbuf);
 int nfp_nfd3_xsk_poll(struct napi_struct *napi, int budget);
 
+#ifndef CONFIG_NFP_NET_IPSEC
+static inline void nfp_nfd3_ipsec_tx(struct nfp_nfd3_tx_desc *txd, struct sk_buff *skb)
+{
+}
+#else
+void nfp_nfd3_ipsec_tx(struct nfp_nfd3_tx_desc *txd, struct sk_buff *skb);
+#endif
+
 #endif
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net.h b/drivers/net/ethernet/netronome/nfp/nfp_net.h
index 0c3e7e2f856d..768539f12214 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net.h
@@ -263,6 +263,10 @@  struct nfp_meta_parsed {
 		u8 tpid;
 		u16 tci;
 	} vlan;
+
+#ifdef CONFIG_NFP_NET_IPSEC
+	u32 ipsec_saidx;
+#endif
 };
 
 struct nfp_net_rx_hash {
@@ -584,6 +588,7 @@  struct nfp_net_dp {
  * @qcp_cfg:            Pointer to QCP queue used for configuration notification
  * @tx_bar:             Pointer to mapped TX queues
  * @rx_bar:             Pointer to mapped FL/RX queues
+ * @ipsec_data:         Ipsec Sa data
  * @tlv_caps:		Parsed TLV capabilities
  * @ktls_tx_conn_cnt:	Number of offloaded kTLS TX connections
  * @ktls_rx_conn_cnt:	Number of offloaded kTLS RX connections
@@ -672,6 +677,10 @@  struct nfp_net {
 	u8 __iomem *tx_bar;
 	u8 __iomem *rx_bar;
 
+#ifdef CONFIG_NFP_NET_IPSEC
+	struct nfp_net_ipsec_data *ipsec_data;
+#endif
+
 	struct nfp_net_tlv_caps tlv_caps;
 
 	unsigned int ktls_tx_conn_cnt;
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index 7e4424d626a6..040c0c2aad80 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -2565,9 +2565,13 @@  int nfp_net_init(struct nfp_net *nn)
 		if (err)
 			return err;
 
-		err = nfp_net_tls_init(nn);
+		err = nfp_net_ipsec_init(nn);
 		if (err)
 			goto err_clean_mbox;
+
+		err = nfp_net_tls_init(nn);
+		if (err)
+			goto err_clean_ipsec;
 	}
 
 	nfp_net_vecs_init(nn);
@@ -2576,6 +2580,9 @@  int nfp_net_init(struct nfp_net *nn)
 		return 0;
 	return register_netdev(nn->dp.netdev);
 
+err_clean_ipsec:
+	nfp_net_ipsec_clean(nn);
+
 err_clean_mbox:
 	nfp_ccm_mbox_clean(nn);
 	return err;
@@ -2591,6 +2598,7 @@  void nfp_net_clean(struct nfp_net *nn)
 		return;
 
 	unregister_netdev(nn->dp.netdev);
+	nfp_net_ipsec_clean(nn);
 	nfp_ccm_mbox_clean(nn);
 	nfp_net_reconfig_wait_posted(nn);
 }
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h b/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h
index 80346c1c266b..fff05496152d 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h
@@ -45,6 +45,7 @@ 
 #define NFP_NET_META_CSUM		6 /* checksum complete type */
 #define NFP_NET_META_CONN_HANDLE	7
 #define NFP_NET_META_RESYNC_INFO	8 /* RX resync info request */
+#define NFP_NET_META_IPSEC		9 /* IPSec SA index for tx and rx */
 
 #define NFP_META_PORT_ID_CTRL		~0U
 
@@ -52,6 +53,8 @@ 
 #define NFP_NET_META_VLAN_SIZE			4
 #define NFP_NET_META_PORTID_SIZE		4
 #define NFP_NET_META_CONN_HANDLE_SIZE		8
+#define NFP_NET_META_IPSEC_SIZE			4
+#define NFP_NET_META_IPSEC_FIELD_SIZE		12
 /* Hash type pre-pended when a RSS hash was computed */
 #define NFP_NET_RSS_NONE		0
 #define NFP_NET_RSS_IPV4		1
@@ -260,6 +263,7 @@ 
  */
 #define NFP_NET_CFG_CTRL_WORD1		0x0098
 #define   NFP_NET_CFG_CTRL_PKT_TYPE	  (0x1 << 0) /* Pkttype offload */
+#define   NFP_NET_CFG_CTRL_IPSEC	  (0x1 << 1) /* IPSec offload */
 
 #define NFP_NET_CFG_CAP_WORD1		0x00a4