RE: [EXT] Re: [PATCH 1/1] net: fec: add initial XDP support

[Shenwei Wang]

> -----Original Message-----
> From: Andrew Lunn <andrew@xxxxxxx>
> Sent: Wednesday, September 28, 2022 8:34 PM
> To: Shenwei Wang <shenwei.wang@xxxxxxx>
> Cc: Joakim Zhang <qiangqing.zhang@xxxxxxx>; David S. Miller
> <davem@xxxxxxxxxxxxx>; Eric Dumazet <edumazet@xxxxxxxxxx>; Jakub
> Kicinski <kuba@xxxxxxxxxx>; Paolo Abeni <pabeni@xxxxxxxxxx>; Alexei
> Starovoitov <ast@xxxxxxxxxx>; Daniel Borkmann <daniel@xxxxxxxxxxxxx>;
> Jesper Dangaard Brouer <hawk@xxxxxxxxxx>; John Fastabend
> <john.fastabend@xxxxxxxxx>; netdev@xxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx; imx@xxxxxxxxxxxxxxx
> Subject: [EXT] Re: [PATCH 1/1] net: fec: add initial XDP support
> 
> Caution: EXT Email
> 
> On Wed, Sep 28, 2022 at 10:25:09AM -0500, Shenwei Wang wrote:
> > This patch adds the initial XDP support to Freescale driver. It
> > supports XDP_PASS, XDP_DROP and XDP_REDIRECT actions. Upcoming
> patches
> > will add support for XDP_TX and Zero Copy features.
> >
> > This patch also optimizes the RX buffers by using the page pool, which
> > uses one frame per page for easy management. In the future, it can be
> > further improved to use two frames per page.
> 
> Please could you split this patch up. It is rather large and hard to review. I think
> you can first add support for the page pool, and then add XDP support, for
> example.
> 

Will do it in the next version. Thanks for the review.

> I would be interesting to see how the page pool helps performance for normal
> traffic, since that is what most people use it for. And for a range of devices,
> since we need to make sure it does not cause any regressions for older devices.
> 

I actually did some compare testing regarding the page pool for normal traffic. 
So far I don't see significant improvement in the current implementation. The performance for large packets improves a little, and the performance for small packets get a little worse.

Thanks,
Shenwei

>             Andrew