[xdp-hints] Re: [PATCH bpf-next v7 15/17] net/mlx5e: Introduce wrapper for xdp_buff

[Tariq Toukan <ttoukan.linux@gmail.com>]

On 13/01/2023 0:29, Toke Høiland-Jørgensen wrote:
> Stanislav Fomichev <sdf@google.com> writes:
> 
>> On Thu, Jan 12, 2023 at 1:55 PM Toke Høiland-Jørgensen <toke@redhat.com> wrote:
>>>
>>> Toke Høiland-Jørgensen <toke@redhat.com> writes:
>>>
>>>> Stanislav Fomichev <sdf@google.com> writes:
>>>>
>>>>> On Thu, Jan 12, 2023 at 12:07 AM Tariq Toukan <ttoukan.linux@gmail.com> wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>> On 12/01/2023 2:32, Stanislav Fomichev wrote:
>>>>>>> From: Toke Høiland-Jørgensen <toke@redhat.com>
>>>>>>>
>>>>>>> Preparation for implementing HW metadata kfuncs. No functional change.
>>>>>>>
>>>>>>> Cc: Tariq Toukan <tariqt@nvidia.com>
>>>>>>> Cc: Saeed Mahameed <saeedm@nvidia.com>
>>>>>>> Cc: John Fastabend <john.fastabend@gmail.com>
>>>>>>> Cc: David Ahern <dsahern@gmail.com>
>>>>>>> Cc: Martin KaFai Lau <martin.lau@linux.dev>
>>>>>>> Cc: Jakub Kicinski <kuba@kernel.org>
>>>>>>> Cc: Willem de Bruijn <willemb@google.com>
>>>>>>> Cc: Jesper Dangaard Brouer <brouer@redhat.com>
>>>>>>> Cc: Anatoly Burakov <anatoly.burakov@intel.com>
>>>>>>> Cc: Alexander Lobakin <alexandr.lobakin@intel.com>
>>>>>>> Cc: Magnus Karlsson <magnus.karlsson@gmail.com>
>>>>>>> Cc: Maryam Tahhan <mtahhan@redhat.com>
>>>>>>> Cc: xdp-hints@xdp-project.net
>>>>>>> Cc: netdev@vger.kernel.org
>>>>>>> Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
>>>>>>> Signed-off-by: Stanislav Fomichev <sdf@google.com>
>>>>>>> ---
>>>>>>>    drivers/net/ethernet/mellanox/mlx5/core/en.h  |  1 +
>>>>>>>    .../net/ethernet/mellanox/mlx5/core/en/xdp.c  |  3 +-
>>>>>>>    .../net/ethernet/mellanox/mlx5/core/en/xdp.h  |  6 +-
>>>>>>>    .../ethernet/mellanox/mlx5/core/en/xsk/rx.c   | 25 ++++----
>>>>>>>    .../net/ethernet/mellanox/mlx5/core/en_rx.c   | 58 +++++++++----------
>>>>>>>    5 files changed, 50 insertions(+), 43 deletions(-)
>>>>>>>
>>>>>>> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
>>>>>>> index 2d77fb8a8a01..af663978d1b4 100644
>>>>>>> --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
>>>>>>> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
>>>>>>> @@ -469,6 +469,7 @@ struct mlx5e_txqsq {
>>>>>>>    union mlx5e_alloc_unit {
>>>>>>>        struct page *page;
>>>>>>>        struct xdp_buff *xsk;
>>>>>>> +     struct mlx5e_xdp_buff *mxbuf;
>>>>>>
>>>>>> In XSK files below you mix usage of both alloc_units[page_idx].mxbuf and
>>>>>> alloc_units[page_idx].xsk, while both fields share the memory of a union.
>>>>>>
>>>>>> As struct mlx5e_xdp_buff wraps struct xdp_buff, I think that you just
>>>>>> need to change the existing xsk field type from struct xdp_buff *xsk
>>>>>> into struct mlx5e_xdp_buff *xsk and align the usage.
>>>>>
>>>>> Hmmm, good point. I'm actually not sure how it works currently.
>>>>> mlx5e_alloc_unit.mxbuf doesn't seem to be initialized anywhere? Toke,
>>>>> am I missing something?
>>>>
>>>> It's initialised piecemeal in different places; but yeah, we're mixing
>>>> things a bit...
>>>>
>>>>> I'm thinking about something like this:
>>
>> Seems more invasive? I don't care much tbf, but what's wrong with
>> keeping 'xdp_buff xsk' member and use it consistently?
> 
> Yeah, it's more invasive, but it's also more consistent with the non-xsk
> path where every usage of struct xdp_buff is replaced with the wrapping
> struct?
> 
> Both will work, I suppose (in fact I think the resulting code will be
> more or less identical), so it's more a matter of which one is easier to
> read and where we put the type-safety-breaking casts.
> 
> I can live with either one (just note you'll have to move the
> 'mxbuf->rq' initialisation next to the one for mxbuf->cqe for yours, but
> that's probably fine too). Let's see which way Tariq prefers...
> 
> -Toke
> 

I think both of the above will still have issues. See my comments in the 
email that I've just sent.