1.. SPDX-License-Identifier: GPL-2.0
2.. _xfrm_device:
3
4===============================================
5XFRM device - offloading the IPsec computations
6===============================================
7
8Shannon Nelson <shannon.nelson@oracle.com>
9Leon Romanovsky <leonro@nvidia.com>
10
11
12Overview
13========
14
15IPsec is a useful feature for securing network traffic, but the
16computational cost is high: a 10Gbps link can easily be brought down
17to under 1Gbps, depending on the traffic and link configuration.
18Luckily, there are NICs that offer a hardware based IPsec offload which
19can radically increase throughput and decrease CPU utilization.  The XFRM
20Device interface allows NIC drivers to offer to the stack access to the
21hardware offload.
22
23Right now, there are two types of hardware offload that kernel supports.
24 * IPsec crypto offload:
25   * NIC performs encrypt/decrypt
26   * Kernel does everything else
27 * IPsec packet offload:
28   * NIC performs encrypt/decrypt
29   * NIC does encapsulation
30   * Kernel and NIC have SA and policy in-sync
31   * NIC handles the SA and policies states
32   * The Kernel talks to the keymanager
33
34Userland access to the offload is typically through a system such as
35libreswan or KAME/raccoon, but the iproute2 'ip xfrm' command set can
36be handy when experimenting.  An example command might look something
37like this for crypto offload:
38
39  ip x s add proto esp dst 14.0.0.70 src 14.0.0.52 spi 0x07 mode transport \
40     reqid 0x07 replay-window 32 \
41     aead 'rfc4106(gcm(aes))' 0x44434241343332312423222114131211f4f3f2f1 128 \
42     sel src 14.0.0.52/24 dst 14.0.0.70/24 proto tcp \
43     offload dev eth4 dir in
44
45and for packet offload
46
47  ip x s add proto esp dst 14.0.0.70 src 14.0.0.52 spi 0x07 mode transport \
48     reqid 0x07 replay-window 32 \
49     aead 'rfc4106(gcm(aes))' 0x44434241343332312423222114131211f4f3f2f1 128 \
50     sel src 14.0.0.52/24 dst 14.0.0.70/24 proto tcp \
51     offload packet dev eth4 dir in
52
53  ip x p add src 14.0.0.70 dst 14.0.0.52 offload packet dev eth4 dir in
54  tmpl src 14.0.0.70 dst 14.0.0.52 proto esp reqid 10000 mode transport
55
56Yes, that's ugly, but that's what shell scripts and/or libreswan are for.
57
58
59
60Callbacks to implement
61======================
62
63::
64
65  /* from include/linux/netdevice.h */
66  struct xfrmdev_ops {
67        /* Crypto and Packet offload callbacks */
68	int	(*xdo_dev_state_add) (struct xfrm_state *x, struct netlink_ext_ack *extack);
69	void	(*xdo_dev_state_delete) (struct xfrm_state *x);
70	void	(*xdo_dev_state_free) (struct xfrm_state *x);
71	bool	(*xdo_dev_offload_ok) (struct sk_buff *skb,
72				       struct xfrm_state *x);
73	void    (*xdo_dev_state_advance_esn) (struct xfrm_state *x);
74
75        /* Solely packet offload callbacks */
76	void    (*xdo_dev_state_update_curlft) (struct xfrm_state *x);
77	int	(*xdo_dev_policy_add) (struct xfrm_policy *x, struct netlink_ext_ack *extack);
78	void	(*xdo_dev_policy_delete) (struct xfrm_policy *x);
79	void	(*xdo_dev_policy_free) (struct xfrm_policy *x);
80  };
81
82The NIC driver offering ipsec offload will need to implement callbacks
83relevant to supported offload to make the offload available to the network
84stack's XFRM subsystem. Additionally, the feature bits NETIF_F_HW_ESP and
85NETIF_F_HW_ESP_TX_CSUM will signal the availability of the offload.
86
87
88
89Flow
90====
91
92At probe time and before the call to register_netdev(), the driver should
93set up local data structures and XFRM callbacks, and set the feature bits.
94The XFRM code's listener will finish the setup on NETDEV_REGISTER.
95
96::
97
98		adapter->netdev->xfrmdev_ops = &ixgbe_xfrmdev_ops;
99		adapter->netdev->features |= NETIF_F_HW_ESP;
100		adapter->netdev->hw_enc_features |= NETIF_F_HW_ESP;
101
102When new SAs are set up with a request for "offload" feature, the
103driver's xdo_dev_state_add() will be given the new SA to be offloaded
104and an indication of whether it is for Rx or Tx.  The driver should
105
106	- verify the algorithm is supported for offloads
107	- store the SA information (key, salt, target-ip, protocol, etc)
108	- enable the HW offload of the SA
109	- return status value:
110
111		===========   ===================================
112		0             success
113		-EOPNETSUPP   offload not supported, try SW IPsec,
114                              not applicable for packet offload mode
115		other         fail the request
116		===========   ===================================
117
118The driver can also set an offload_handle in the SA, an opaque void pointer
119that can be used to convey context into the fast-path offload requests::
120
121		xs->xso.offload_handle = context;
122
123
124When the network stack is preparing an IPsec packet for an SA that has
125been setup for offload, it first calls into xdo_dev_offload_ok() with
126the skb and the intended offload state to ask the driver if the offload
127will serviceable.  This can check the packet information to be sure the
128offload can be supported (e.g. IPv4 or IPv6, no IPv4 options, etc) and
129return true of false to signify its support.
130
131Crypto offload mode:
132When ready to send, the driver needs to inspect the Tx packet for the
133offload information, including the opaque context, and set up the packet
134send accordingly::
135
136		xs = xfrm_input_state(skb);
137		context = xs->xso.offload_handle;
138		set up HW for send
139
140The stack has already inserted the appropriate IPsec headers in the
141packet data, the offload just needs to do the encryption and fix up the
142header values.
143
144
145When a packet is received and the HW has indicated that it offloaded a
146decryption, the driver needs to add a reference to the decoded SA into
147the packet's skb.  At this point the data should be decrypted but the
148IPsec headers are still in the packet data; they are removed later up
149the stack in xfrm_input().
150
151	find and hold the SA that was used to the Rx skb::
152
153		get spi, protocol, and destination IP from packet headers
154		xs = find xs from (spi, protocol, dest_IP)
155		xfrm_state_hold(xs);
156
157	store the state information into the skb::
158
159		sp = secpath_set(skb);
160		if (!sp) return;
161		sp->xvec[sp->len++] = xs;
162		sp->olen++;
163
164	indicate the success and/or error status of the offload::
165
166		xo = xfrm_offload(skb);
167		xo->flags = CRYPTO_DONE;
168		xo->status = crypto_status;
169
170	hand the packet to napi_gro_receive() as usual
171
172In ESN mode, xdo_dev_state_advance_esn() is called from xfrm_replay_advance_esn().
173Driver will check packet seq number and update HW ESN state machine if needed.
174
175Packet offload mode:
176HW adds and deletes XFRM headers. So in RX path, XFRM stack is bypassed if HW
177reported success. In TX path, the packet lefts kernel without extra header
178and not encrypted, the HW is responsible to perform it.
179
180When the SA is removed by the user, the driver's xdo_dev_state_delete()
181and xdo_dev_policy_delete() are asked to disable the offload.  Later,
182xdo_dev_state_free() and xdo_dev_policy_free() are called from a garbage
183collection routine after all reference counts to the state and policy
184have been removed and any remaining resources can be cleared for the
185offload state.  How these are used by the driver will depend on specific
186hardware needs.
187
188As a netdev is set to DOWN the XFRM stack's netdev listener will call
189xdo_dev_state_delete(), xdo_dev_policy_delete(), xdo_dev_state_free() and
190xdo_dev_policy_free() on any remaining offloaded states.
191
192Outcome of HW handling packets, the XFRM core can't count hard, soft limits.
193The HW/driver are responsible to perform it and provide accurate data when
194xdo_dev_state_update_curlft() is called. In case of one of these limits
195occuried, the driver needs to call to xfrm_state_check_expire() to make sure
196that XFRM performs rekeying sequence.
197