[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <41eae08f-5f77-4099-bcd4-ccc7bbcf6426@intel.com>
Date: Fri, 22 Aug 2025 11:09:12 +0200
From: Przemek Kitszel <przemyslaw.kitszel@...el.com>
To: <mheib@...hat.com>, Jiri Pirko <jiri@...nulli.us>
CC: <anthony.l.nguyen@...el.com>, <intel-wired-lan@...ts.osuosl.org>,
<netdev@...r.kernel.org>, Jacob Keller <jacob.e.keller@...el.com>, "Simon
Horman" <horms@...nel.org>
Subject: Re: [PATCH net] i40e: add devlink param to control VF MAC address
limit
On 8/22/25 01:39, mheib@...hat.com wrote:
> From: Mohammad Heib <mheib@...hat.com>
>
> This patch introduces a new devlink runtime parameter
> to control whether the VF MAC filter limit feature is
> enabled or disabled.
>
> When the parameter is set to non-zero, the driver enforces the per-VF MAC
> filter limit calculated from the number of allocated VFs and ports.
> When the parameter is unset (zero), no limit is applied and behavior
> remains as before commit cfb1d572c986
> ("i40e: Add ensurance of MacVlan resources for every trusted VF").
>
> This implementation allows us to toggle the feature through devlink
> while preserving old behavior. In the future, the parameter can be
> extended to represent a configurable "max MACs per VF" value, but for
> now it acts as a simple on/off switch.
>
> Example command to enable per-vf mac limit:
> - devlink dev param set pci/0000:3b:00.0 name max_mac_per_vf \
> value 1 \
> cmode runtime
>
> Fixes: cfb1d572c986 ("i40e: Add ensurance of MacVlan resources for every trusted VF")
> Signed-off-by: Mohammad Heib <mheib@...hat.com>
thank you for the patch, I have a few questions/objections
1. it git-conflicts with [1], please post your next revision based on
Tony's (fixes) tree dev-queue branch [2]
2a. it is good practice to link to the previous discussions, and CC
individuals involved (Jake, Simon)
2b. for changes that utilize given subsystem (devlink), you need to CC
respective maintainers (Jiri)
3. it would really be better to treat not-zero values as strict limit
4. this idea is not limited to i40e, the parameter should be global
(for all drivers to implement), as it seems generic enough
5. when someone will make a per-given-VF param, this one will not be
deprecated but will still work as a cap (max) value. (Leaving it at
zero will be ofc perfectly fine).
[1]
https://lore.kernel.org/intel-wired-lan/20250813104552.61027-9-przemyslaw.kitszel@intel.com/T/#mac68de249365b8c4fa83054592dd98f0f789fab4
[2]
https://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue.git/log/?h=dev-queue
> ---
> Documentation/networking/devlink/i40e.rst | 19 +++++++
> drivers/net/ethernet/intel/i40e/i40e.h | 4 ++
> .../net/ethernet/intel/i40e/i40e_devlink.c | 56 ++++++++++++++++++-
> .../ethernet/intel/i40e/i40e_virtchnl_pf.c | 14 ++++-
> 4 files changed, 89 insertions(+), 4 deletions(-)
>
Powered by blists - more mailing lists