Leon Romanovsky says:

============
This is second part with implementation of packet offload.
============
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Enable configuration of IPsec packet offload through XFRM state add
interface together with moving specific to IPsec packet mode limitations
to specific switch-case section.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Extend event logic to update ESN state (esn_msb, esn_overlap)
for an IPsec Offload context.
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Enable object changed event to signal IPsec about hitting
soft and hard limits.
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Implement mlx5 IPsec callback to update current lifetime counters.
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Instead of performing custom hash calculations, rely on FW that returns
unique identifier to every created SA. That identifier is Xarray ready,
which provides better semantic with efficient access.

In addition, store both TX and RX SAs to allow correlation between event
generated by HW when limits are armed and XFRM states.
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Improve readability by providing direct pointer to struct mlx5e_ipsec.
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Software implementation of IPsec skips encryption of packets in TX
path if no matching policy is found. So align HW implementation to
this behavior, by requiring matching reqid for offloaded policy and
SA.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Add the following statistics:
RX successfully IPsec flows:
ipsec_rx_pkts  : Number of packets passed Rx IPsec flow
ipsec_rx_bytes : Number of bytes passed Rx IPsec flow

Rx dropped IPsec policy packets:
ipsec_rx_drop_pkts: Number of packets dropped in Rx datapath due to IPsec drop policy
ipsec_rx_drop_bytes: Number of bytes dropped in Rx datapath due to IPsec drop policy

TX successfully encrypted and encapsulated IPsec packets:
ipsec_tx_pkts  : Number of packets encrypted and encapsulated successfully
ipsec_tx_bytes : Number of bytes encrypted and encapsulated successfully

Tx dropped IPsec policy packets:
ipsec_tx_drop_pkts: Number of packets dropped in Tx datapath due to IPsec drop policy
ipsec_tx_drop_bytes: Number of bytes dropped in Tx datapath due to IPsec drop policy

The above can be seen using:
ethtool -S <ifc> |grep ipsec
Signed-off-by: Raed Salem <raeds@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Flow steering API separates newly created rules based on their
match criteria. Right now, all IPsec tables are created with one
group and suffers from not-optimal FS performance.

Count number of different match criteria for relevant tables, and
set proper value at the table creation.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

In packet offload mode, the HW is responsible to handle ESP headers,
SPI numbers and trailers (ICV) together with different logic for
RX and TX paths.

In order to support packet offload mode, special logic is added
to flow steering rules.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Remove intermediate variable in favor of having similar coding style
for Rx and Tx add rule functions.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Implement mlx5 flow steering logic and mlx5 IPsec code support
XFRM policy offload.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Add empty table to be used for IPsec policy offload.
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Leon Romanovsky says:

============
This series follows previously sent "Extend XFRM core to allow packet
offload configuration" series [1].

It is first part with refactoring to mlx5 allow us natively extend
mlx5 IPsec logic to support both crypto and packet offloads.
============
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Create general function that sets miss group and rule to forward all
not-matched traffic to the next table.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Move miss handles into dedicated struct, so we can reuse it in next
patch when creating IPsec policy flow table.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Reuse existing struct what holds all information about modify
header pointer and rule. This helps to reduce ambiguity from the
name _err_ that doesn't describe the real purpose of that flow
table, rule and function - to copy status result from HW to
the stack.
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Rewrote the IPsec RX add rule path to be less convoluted
and don't rely on pre-initialized variables. The code now has clean
linear flow with clean separation between error and success paths.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

The policy offload logic needs to set flow steering rule that match
on saddr and daddr too, so factor out this code to separate functions,
together with code alignment to netdev coding pattern of relying on
family type.

As part of this change, let's separate more logic from setup_fte_common
to make sure that the function names describe that is done in the
function better than general *common* name.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Even now, to support IPsec crypto, the RX and TX paths use same
logic to create flow tables. In the following patches, we will
add more tables to support IPsec packet offload. So reuse existing
code and rewrite it to support IPsec packet offload from the beginning.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Create initial hardware IPsec packet offload object and connect it
to advanced steering operation (ASO) context and queue, so the data
path can communicate with the stack.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Setup the ASO (Advanced Steering Operation) object that is needed
for IPsec to interact with SW stack about various fast changing
events: replay window, lifetime limits,  e.t.c
Reviewed-by: Raed Salem <raeds@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

mlx5 priv structure is driver main structure that holds high level data.
That information is not needed for IPsec flow steering logic and the
pointer to mlx5e_priv was not supposed to be passed in the first place.

This change "cleans" the logic to rely on internal to IPsec structures
without touching global mlx5e_priv.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Low level mlx5 code needs to use mlx5_core print routines and
not netdev ones, as the failures are relevant to the HW itself
and not to its netdev.

This change allows us to remove access to mlx5 priv structure, which
holds high level driver data that isn't needed for mlx5 IPsec code.
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Remove AF family obfuscation by creating symmetric structs for RX and
TX IPsec flow steering chains. This simplifies to us low level IPsec
FS creation logic without need to dig into multiple levels of structs.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Instead of performing redefinition of XFRM core defines to same
values but with MLX5_* prefix, cache the input values as is by making
sure that the proper storage objects are used.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

As a preparation for future extension of IPsec hardware object to allow
configuration of packet offload mode, extend the XFRM validator to check
replay window values.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Add needed capabilities check to determine if device supports IPsec
packet offload mode.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Add all needed bits to support IPsec packet offload mode.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

There is no need in hiding returned ASO WQE type by providing void*,
use the real type instead. Do it together with zeroing that memory,
so ASO WQE will be ready to use immediately.
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Leon Romanovsky says:

============
The following series extends XFRM core code to handle a new type of IPsec
offload - packet offload.

In this mode, the HW is going to be responsible for the whole data path,
so both policy and state should be offloaded.

IPsec packet offload is an improved version of IPsec crypto mode,
In packet mode, HW is responsible to trim/add headers in addition to
decrypt/encrypt. In this mode, the packet arrives to the stack as already
decrypted and vice versa for TX (exits to HW as not-encrypted).

Devices that implement IPsec packet offload mode offload policies too.
In the RX path, it causes the situation that HW can't effectively
handle mixed SW and HW priorities unless users make sure that HW offloaded
policies have higher priorities.

It means that we don't need to perform any search of inexact policies
and/or priority checks if HW policy was discovered. In such situation,
the HW will catch the packets anyway and HW can still implement inexact
lookups.

In case specific policy is not found, we will continue with packet lookup
and check for existence of HW policies in inexact list.

HW policies are added to the head of SPD to ensure fast lookup, as XFRM
iterates over all policies in the loop.

This simple solution allows us to achieve same benefits of separate HW/SW
policies databases without over-engineering the code to iterate and manage
two databases at the same path.

To not over-engineer the code, HW policies are treated as SW ones and
don't take into account netdev to allow reuse of the same priorities for
policies databases without over-engineering the code to iterate and manage
two databases at the same path.

To not over-engineer the code, HW policies are treated as SW ones and
don't take into account netdev to allow reuse of the same priorities for
different devices.
 * No software fallback
 * Fragments are dropped, both in RX and TX
 * No sockets policies
 * Only IPsec transport mode is implemented

================================================================================
Rekeying:

In order to support rekeying, as XFRM core is skipped, the HW/driver should
do the following:
 * Count the handled packets
 * Raise event that limits are reached
 * Drop packets once hard limit is occurred.

The XFRM core calls to newly introduced xfrm_dev_state_update_curlft()
function in order to perform sync between device statistics and internal
structures. On HW limit event, driver calls to xfrm_state_check_expire()
to allow XFRM core take relevant decisions.

This separation between control logic (in XFRM) and data plane allows us
to packet reuse SW stack.

================================================================================
Configuration:

iproute2: https://lore.kernel.org/netdev/cover.1652179360.git.leonro@nvidia.com/

Packet offload mode:
  ip xfrm state offload packet dev <if-name> dir <in|out>
  ip xfrm policy .... offload packet dev <if-name>
Crypto offload mode:
  ip xfrm state offload crypto dev <if-name> dir <in|out>
or (backward compatibility)
  ip xfrm state offload dev <if-name> dir <in|out>

================================================================================
Performance results:

TCP multi-stream, using iperf3 instance per-CPU.
+----------------------+--------+--------+--------+--------+---------+---------+
|                      | 1 CPU  | 2 CPUs | 4 CPUs | 8 CPUs | 16 CPUs | 32 CPUs |
|                      +--------+--------+--------+--------+---------+---------+
|                      |                   BW (Gbps)                           |
+----------------------+--------+--------+-------+---------+---------+---------+
| Baseline             | 27.9   | 59     | 93.1  | 92.8    | 93.7    | 94.4    |
+----------------------+--------+--------+-------+---------+---------+---------+
| Software IPsec       | 6      | 11.9   | 23.3  | 45.9    | 83.8    | 91.8    |
+----------------------+--------+--------+-------+---------+---------+---------+
| IPsec crypto offload | 15     | 29.7   | 58.5  | 89.6    | 90.4    | 90.8    |
+----------------------+--------+--------+-------+---------+---------+---------+
| IPsec packet offload | 28     | 57     | 90.7  | 91      | 91.3    | 91.9    |
+----------------------+--------+--------+-------+---------+---------+---------+

IPsec packet offload mode behaves as baseline and reaches linerate with same amount
of CPUs.

Setups details (similar for both sides):
* NIC: ConnectX6-DX dual port, 100 Gbps each.
  Single port used in the tests.
* CPU: Intel(R) Xeon(R) Platinum 8380 CPU @ 2.30GHz

================================================================================
Series together with mlx5 part:
https://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git/log/?h=xfrm-next

================================================================================
Changelog:

v10:
 * Added forgotten xdo_dev_state_del. Patch #4.
 * Moved changelog in cover letter to the end.
 * Added "if (xs->xso.type != XFRM_DEV_OFFLOAD_CRYPTO) {" line to newly
   added netronome IPsec support. Patch #2.
v9: https://lore.kernel.org/all/cover.1669547603.git.leonro@nvidia.com
 * Added acquire support
v8: https://lore.kernel.org/all/cover.1668753030.git.leonro@nvidia.com
 * Removed not-related blank line
 * Fixed typos in documentation
v7: https://lore.kernel.org/all/cover.1667997522.git.leonro@nvidia.com
As was discussed in IPsec workshop:
 * Renamed "full offload" to be "packet offload".
 * Added check that offloaded SA and policy have same device while sending packet
 * Added to SAD same optimization as was done for SPD to speed-up lookups.
v6: https://lore.kernel.org/all/cover.1666692948.git.leonro@nvidia.com
 * Fixed misplaced "!" in sixth patch.
v5: https://lore.kernel.org/all/cover.1666525321.git.leonro@nvidia.com
 * Rebased to latest ipsec-next.
 * Replaced HW priority patch with solution which mimics separated SPDs
   for SW and HW. See more description in this cover letter.
 * Dropped RFC tag, usecase, API and implementation are clear.
v4: https://lore.kernel.org/all/cover.1662295929.git.leonro@nvidia.com
 * Changed title from "PATCH" to "PATCH RFC" per-request.
 * Added two new patches: one to update hard/soft limits and another
   initial take on documentation.
 * Added more info about lifetime/rekeying flow to cover letter, see
   relevant section.
 * perf traces for crypto mode will come later.
v3: https://lore.kernel.org/all/cover.1661260787.git.leonro@nvidia.com
 * I didn't hear any suggestion what term to use instead of
   "packet offload", so left it as is. It is used in commit messages
   and documentation only and easy to rename.
 * Added performance data and background info to cover letter
 * Reused xfrm_output_resume() function to support multiple XFRM transformations
 * Add PMTU check in addition to driver .xdo_dev_offload_ok validation
 * Documentation is in progress, but not part of this series yet.
v2: https://lore.kernel.org/all/cover.1660639789.git.leonro@nvidia.com
 * Rebased to latest 6.0-rc1
 * Add an extra check in TX datapath patch to validate packets before
   forwarding to HW.
 * Added policy cleanup logic in case of netdev down event
v1: https://lore.kernel.org/all/cover.1652851393.git.leonro@nvidia.com
 * Moved comment to be before if (...) in third patch.
v0: https://lore.kernel.org/all/cover.1652176932.git.leonro@nvidia.com
-----------------------------------------------------------------------
============
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Extend XFRM device offload API description with newly
added packet offload mode.
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Both in RX and TX, the traffic that performs IPsec packet offload
transformation is accounted by HW. It is needed to properly handle
hard limits that require to drop the packet.

It means that XFRM core needs to update internal counters with the one
that accounted by the HW, so new callbacks are introduced in this patch.

In case of soft or hard limit is occurred, the driver should call to
xfrm_state_check_expire() that will perform key rekeying exactly as
done by XFRM core.
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Devices that implement IPsec packet offload mode should offload SA and
policies too. In RX path, it causes to the situation that HW will always
have higher priority over any SW policies.

It means that we don't need to perform any search of inexact policies
and/or priority checks if HW policy was discovered. In such situation,
the HW will catch the packets anyway and HW can still implement inexact
lookups.

In case specific policy is not found, we will continue with packet lookup and
check for existence of HW policies in inexact list.

HW policies are added to the head of SPD to ensure fast lookup, as XFRM
iterates over all policies in the loop.

The same solution of adding HW SAs at the begging of the list is applied
to SA database too. However, we don't need to change lookups as they are
sorted by insertion order and not priority.
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Traffic received by device with enabled IPsec packet offload should
be forwarded to the stack only after decryption, packet headers and
trailers removed.

Such packets are expected to be seen as normal (non-XFRM) ones, while
not-supported packets should be dropped by the HW.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

In IPsec packet mode, the device is going to encrypt and encapsulate
packets that are associated with offloaded policy. After successful
policy lookup to indicate if packets should be offloaded or not,
the stack forwards packets to the device to do the magic.
Signed-off-by: Raed Salem <raeds@nvidia.com>
Signed-off-by: Huy Nguyen <huyn@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Extend netlink interface to add and delete XFRM policy from the device.
This functionality is a first step to implement packet IPsec offload solution.
Signed-off-by: Raed Salem <raeds@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

Allow users to configure xfrm states with packet offload mode.
The packet mode must be requested both for policy and state, and
such requires us to do not implement fallback.

We explicitly return an error if requested packet mode can't
be configured.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>

In the next patches, the xfrm core code will be extended to support
new type of offload - packet offload. In that mode, both policy and state
should be specially configured in order to perform whole offloaded data
path.

Full offload takes care of encryption, decryption, encapsulation and
other operations with headers.

As this mode is new for XFRM policy flow, we can "start fresh" with flag
bits and release first and second bit for future use.
Reviewed-by: Raed Salem <raeds@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>