[Intel-wired-lan] [PATCH iwl-net v3] ice: use DSN instead of PCI BDF for ice_adapter index

Simon Horman horms at kernel.org
Thu Apr 10 15:43:53 UTC 2025


On Tue, Apr 08, 2025 at 03:46:55PM +0200, Przemek Kitszel wrote:
> Use Device Serial Number instead of PCI bus/device/function for
> index of struct ice_adapter.
> Functions on the same physical device should point to the very same
> ice_adapter instance.
> 
> This is not only simplification, but also fixes things up when PF
> is passed to VM (and thus has a random BDF).

Maybe it's just me but "fixes things up" seems a bit vague for
a fix for net. Could something more specific go here?

> 
> Suggested-by: Jacob Keller <jacob.e.keller at intel.com>
> Suggested-by: Jakub Kicinski <kuba at kernel.org>
> Suggested-by: Jiri Pirko <jiri at resnulli.us>
> Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov at intel.com>
> Signed-off-by: Przemek Kitszel <przemyslaw.kitszel at intel.com>
> ---
> CC: Karol Kolacinski <karol.kolacinski at intel.com>
> CC: Grzegorz Nitka <grzegorz.nitka at intel.com>
> CC: Michal Schmidt <mschmidt at redhat.com>
> CC: Sergey Temerkhanov <sergey.temerkhanov at intel.com>
> CC: Michal Kubiak <michal.kubiak at intel.com>
> 
> v3:
>  - Add fixes tag (Michal K)

The fixes tag seems to have got lost in transit.

I believe it should be [1]:

Fixes: 0e2bddf9e5f9 ("ice: add ice_adapter for shared data across PFs on the same NIC")

[1] https://lore.kernel.org/intel-wired-lan/7f700a89-7058-4c16-b53a-2e84bbed8542@intel.com/

>  - add missing braces (lkp bot), turns out it's hard to purge C++ from your mind
>  - (no changes in the collision handling on 32bit systems)
> 
> v2:
> https://lore.kernel.org/intel-wired-lan/20250407112005.85468-1-przemyslaw.kitszel@intel.com/
>  - target to -net (Jiri)
>  - mix both halves of u64 DSN on 32bit systems (Jiri)
>  - (no changes in terms of fallbacks for pre-prod HW)
>  - warn when there is DSN collision after reducing to 32bit
> 
> v1:
> https://lore.kernel.org/netdev/20250306211159.3697-2-przemyslaw.kitszel@intel.com

...


More information about the Intel-wired-lan mailing list