ACK/Cmnt: [Pull Request][jammy] Add OVS Internal Port HW Offload to mlx5 driver

[Stefan Bader]

On 04.08.22 15:05, Zachary Tahenakos wrote:
> BugLink: https://bugs.launchpad.net/bugs/1983498 
> <https://bugs.launchpad.net/bugs/1983498>
> 
> SRU Justification
> 
> [Impact]
> 
> A request has been made, originating from 
> https://bugs.launchpad.net/bugs/1980730 
> <https://bugs.launchpad.net/bugs/1980730>, to add support for OVS internal port 
> offload for the mlx5 driver. This feature takes care of certain traffic 
> workloads that are not currently being offloaded to the hardware.
> 
> [Test Case]
> 
> Tested by the requester.
> 
> [Where things could go wrong]
> 
> These changes could potentially introduce regressions in behavior of the mlx5 
> driver.
> 
> The following changes since commit 4cc5e60c9399a10400737bda19312574b5a5c7a2:
> 
>    UBUNTU: Ubuntu-5.15.0-45.48 (2022-07-19 18:57:17 -0300)
> 
> are available in the Git repository at:
> 
> https://git.launchpad.net/~ztahenakos/ubuntu/+source/linux/+git/jammy 
> <https://git.launchpad.net/~ztahenakos/ubuntu/+source/linux/+git/jammy> 
> 5195a9ae669b266eb35cd8ee604923d4c5e5058a
> 
> for you to fetch changes up to 5195a9ae669b266eb35cd8ee604923d4c5e5058a:
> 
>    net/mlx5e: TC, fix decap fallback to uplink when int port not supported 
> (2022-08-03 15:10:10 -0400)
> 
> ----------------------------------------------------------------
> Ariel Levkovich (11):
>        net/mlx5e: Refactor rx handler of represetor device
>        net/mlx5e: Use generic name for the forwarding dev pointer
>        net/mlx5: E-Switch, Add ovs internal port mapping to metadata support
>        net/mlx5e: Accept action skbedit in the tc actions list
>        net/mlx5e: Offload tc rules that redirect to ovs internal port
>        net/mlx5e: Offload internal port as encap route device
>        net/mlx5e: Add indirect tc offload of ovs internal port
>        net/mlx5e: Term table handling of internal port rules
>        net/mlx5: Support internal port as decap route device
>        net/mlx5e: Fix wrong source vport matching on tunnel rule
>        net/mlx5e: TC, fix decap fallback to uplink when int port not supported
> 
> Christophe JAILLET (1):
>        net/mlx5: Fix some error handling paths in 'mlx5e_tc_add_fdb_flow()'
> 
> Dima Chumak (2):
>        net/mlx5e: Enable TC offload for ingress MACVLAN
>        net/mlx5e: Fix nullptr on deleting mirroring rule
> 
> Gal Pressman (1):
>        net/mlx5e: Fix skb memory leak when TC classifier action offloads are 
> disabled
> 
> Roi Dayan (2):
>        net/mlx5e: TC, Fix memory leak with rules with internal port
>        net/mlx5e: Avoid implicit modify hdr for decap drop rule
> 
> Vlad Buslov (1):
>        net/mlx5e: Support accept action
> 
>   drivers/net/ethernet/mellanox/mlx5/core/Makefile   |   2 +-
>   .../net/ethernet/mellanox/mlx5/core/en/rep/tc.c    | 132 ++++--
>   .../net/ethernet/mellanox/mlx5/core/en/rep/tc.h    |  14 +-
>   .../ethernet/mellanox/mlx5/core/en/tc/int_port.c   | 457 +++++++++++++++++++++
>   .../ethernet/mellanox/mlx5/core/en/tc/int_port.h   |  65 +++
>   .../net/ethernet/mellanox/mlx5/core/en/tc_priv.h   |   2 +
>   .../net/ethernet/mellanox/mlx5/core/en/tc_tun.c    |  33 +-
>   .../ethernet/mellanox/mlx5/core/en/tc_tun_encap.c  |  35 ++
>   drivers/net/ethernet/mellanox/mlx5/core/en_rep.c   |  13 +-
>   drivers/net/ethernet/mellanox/mlx5/core/en_rep.h   |   4 +
>   drivers/net/ethernet/mellanox/mlx5/core/en_rx.c    |  22 +-
>   drivers/net/ethernet/mellanox/mlx5/core/en_tc.c    | 205 ++++++++-
>   drivers/net/ethernet/mellanox/mlx5/core/en_tc.h    |  11 +-
>   drivers/net/ethernet/mellanox/mlx5/core/eswitch.h  |  16 +
>   .../ethernet/mellanox/mlx5/core/eswitch_offloads.c |  84 ++--
>   .../mellanox/mlx5/core/eswitch_offloads_termtbl.c  |   7 +-
>   16 files changed, 998 insertions(+), 104 deletions(-)
>   create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c
>   create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.h
> 
> 
All upstream cherry picks and isolated to the mlx5 driver. Which hopefully gives 
us reliable results from verification testing.
I think between me and Zachary we can explain the 2 cherry-picks, the first was 
done by the upstream maintainer (probably to prepare a pull request) and somehow 
when Linus merged this the picked sha1 became part of upstream repository. Now 
you can access it in upstream both ways.

Acked-by: Stefan Bader <stefan.bader at canonical.com>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20220824/374d65ff/attachment.sig>