[net-next,0/5] igc: ethtool: Check VLAN TCI mask

Message

Kurt Kanzenbach Nov. 28, 2023, 7:48 a.m. UTC

Hi,

currently it is possible to configure receive queue assignment using the VLAN
TCI field with arbitrary masks. However, the hardware only supports steering
either by full TCI or the priority (PCP) field. In case a wrong mask is given by
the user the driver will silently convert it into a PCP filter which is not
desired. Therefore, add a check for it.

Patches #1 to #4 are minor things found along the way.

Kurt Kanzenbach (5):
  igc: Use reverse xmas tree
  igc: Use netdev printing functions for flex filters
  igc: Unify filtering rule fields
  igc: Report VLAN EtherType matching back to user
  igc: Check VLAN TCI mask

 drivers/net/ethernet/intel/igc/igc.h         |  3 +-
 drivers/net/ethernet/intel/igc/igc_ethtool.c | 36 +++++++++++++++++---
 drivers/net/ethernet/intel/igc/igc_main.c    | 21 ++++++------
 3 files changed, 45 insertions(+), 15 deletions(-)

Comments

Vinicius Costa Gomes Nov. 28, 2023, 11:37 p.m. UTC | #1

Kurt Kanzenbach <[e-mail]> writes:

> Hi,
>
> currently it is possible to configure receive queue assignment using the VLAN
> TCI field with arbitrary masks. However, the hardware only supports steering
> either by full TCI or the priority (PCP) field. In case a wrong mask is given by
> the user the driver will silently convert it into a PCP filter which is not
> desired. Therefore, add a check for it.
>
> Patches #1 to #4 are minor things found along the way.
>

Some very minor things: patches 2,3 and 4 have extra long lines in their
commit messages that checkpatch.pl doesn't seem to like.

Patches 4 and 5 read more like fixes to me. I think they could be
proposed to -net, as they contain fixes to user visible issues. Do you
think that makes sense?

As for the code, feel free to add my Ack to the series:

Acked-by: Vinicius Costa Gomes <[e-mail]>


Cheers,

Kurt Kanzenbach Nov. 29, 2023, 11:15 a.m. UTC | #2

On Tue Nov 28 2023, Vinicius Costa Gomes wrote:
> Kurt Kanzenbach <[e-mail]> writes:
>
>> Hi,
>>
>> currently it is possible to configure receive queue assignment using the VLAN
>> TCI field with arbitrary masks. However, the hardware only supports steering
>> either by full TCI or the priority (PCP) field. In case a wrong mask is given by
>> the user the driver will silently convert it into a PCP filter which is not
>> desired. Therefore, add a check for it.
>>
>> Patches #1 to #4 are minor things found along the way.
>>
>
> Some very minor things: patches 2,3 and 4 have extra long lines in their
> commit messages that checkpatch.pl doesn't seem to like.

OK. checkpatch wants 75 chars per line. These patches have 80 set. I'll
adjust it.

>
> Patches 4 and 5 read more like fixes to me. I think they could be
> proposed to -net, as they contain fixes to user visible issues. Do you
> think that makes sense?

Probably yes. I'll sent them to -net instead. Fixes tags would be:

 - Patch 4: 2b477d057e33 ("igc: Integrate flex filter into ethtool ops")
 - Patch 5: 7991487ecb2d ("igc: Allow for Flex Filters to be installed")

>
> As for the code, feel free to add my Ack to the series:
>
> Acked-by: Vinicius Costa Gomes <[e-mail]>
>

Thanks,
Kurt