netdev - [PATCH iwl-next 0/2] net: intel: cleanup RSS hash configuration bits

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20250430-jk-hash-ena-refactor-v1-0-8310a4785472@intel.com>
Date: Wed, 30 Apr 2025 10:11:51 -0700
From: Jacob Keller <jacob.e.keller@...el.com>
To: netdev <netdev@...r.kernel.org>, 
 Anthony Nguyen <anthony.l.nguyen@...el.com>, 
 Intel Wired LAN <intel-wired-lan@...ts.osuosl.org>
Cc: Jacob Keller <jacob.e.keller@...el.com>, 
 Aleksandr Loktionov <aleksandr.loktionov@...el.com>, 
 Przemek Kitszel <przemyslaw.kitszel@...el.com>
Subject: [PATCH iwl-next 0/2] net: intel: cleanup RSS hash configuration
 bits

The virtchnl API does not define the RSS Hash configuration bits directly
in virtchnl.h, but instead implicitly relies on the hardware definitions
for X710 and X722 in the i40e driver.

This results in duplicating the same bit definitions across 3 drivers. The
actual virtchnl.h header makes no mention of this, and its very unclear
what the bits mean without deep knowledge of the way RSS configuration
works over virtchnl.

In addition, the use of the term 'hena' is confusing. It comes from the
I40E_PFQF_HENA registers, indicating which hash types are enabled.

Rename the 'hena' fields and related functions to use 'hashcfg' as a
shorthand for hash configuration.

We could define the enumeration of packet types in virtchnl.h. Indeed, this
is what the out-of-tree releases of virtchnl.h do. However, this is
somewhat confusing for i40e. The X710 and X722 hardware use these bits
directly with PF hardware registers. It looks confusing to use "VIRTCHNL_*"
names for such access.

Instead, we move these definitions to libie as part of new pctype.h header
file. This allows us to remove all duplicate definitions and have a single
place for Linux to define the bit meanings. The virtchnl.h header can point
to this enumeration to clarify where the values are defined.

Signed-off-by: Jacob Keller <jacob.e.keller@...el.com>
---
Jacob Keller (2):
      net: intel: rename 'hena' to 'hashcfg' for clarity
      net: intel: move RSS packet classifier types to libie

 drivers/net/ethernet/intel/i40e/i40e_txrx.h        | 43 ++++++------
 drivers/net/ethernet/intel/i40e/i40e_type.h        | 32 ---------
 drivers/net/ethernet/intel/iavf/iavf.h             | 10 +--
 drivers/net/ethernet/intel/iavf/iavf_txrx.h        | 40 ++++++-----
 drivers/net/ethernet/intel/iavf/iavf_type.h        | 32 ---------
 drivers/net/ethernet/intel/ice/ice_flow.h          | 68 ++++++------------
 drivers/net/ethernet/intel/ice/ice_virtchnl.h      |  4 +-
 include/linux/avf/virtchnl.h                       | 23 +++---
 include/linux/net/intel/libie/pctype.h             | 44 ++++++++++++
 drivers/net/ethernet/intel/i40e/i40e_ethtool.c     | 81 +++++++++++-----------
 drivers/net/ethernet/intel/i40e/i40e_main.c        | 25 +++----
 drivers/net/ethernet/intel/i40e/i40e_txrx.c        | 25 +++----
 drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 46 ++++++------
 drivers/net/ethernet/intel/iavf/iavf_main.c        | 17 ++---
 drivers/net/ethernet/intel/iavf/iavf_virtchnl.c    | 33 ++++-----
 drivers/net/ethernet/intel/ice/ice_flow.c          | 45 ++++++------
 drivers/net/ethernet/intel/ice/ice_lib.c           |  2 +-
 drivers/net/ethernet/intel/ice/ice_virtchnl.c      | 44 ++++++------
 .../ethernet/intel/ice/ice_virtchnl_allowlist.c    |  2 +-
 19 files changed, 289 insertions(+), 327 deletions(-)
---
base-commit: deeed351e982ac4d521598375b34b071304533b0
change-id: 20250423-jk-hash-ena-refactor-9ac1e2b83a5e

Best regards,
-- 
Jacob Keller <jacob.e.keller@...el.com>