On Mon, 30 Jul 2018 16:09:47 +0300
Tariq Toukan <tar...@mellanox.com> wrote:
> On 30/07/2018 4:06 PM, Jesper Dangaard Brouer wrote:
> >
> > On Mon, 30 Jul 2018 14:10:21 +0200 Jesper Dangaard Brouer
> > <bro...@redhat.com> wrote:
> >
> >> On Thu, 26 Jul 2018 15:56:44 -0700 Saeed Mahameed <sae...@mellanox.com>
> >> wrote:
> >>
> >>> +int mlx5e_xdp_xmit(struct net_device *dev, int n, struct xdp_frame
> >>> **frames,
> >>> + u32 flags)
> >>> +{
> > [...]
> >>> + xdpi.dma_addr = dma_map_single(sq->pdev, xdpf->data, xdpf->len,
> >>> + DMA_TO_DEVICE);
> >>> + if (unlikely(dma_mapping_error(sq->pdev, xdpi.dma_addr))) {
> >>> + drops++;
> >>
> >> I think you are missing a xdp_return_frame_rx_napi(xdpf) here.
> >
> > I see it already got applied... I'll send a fixup later...
> >
>
> Yes it is applied, I can do it myself. Just let me know what you prefer.
I can take care of this. I already have a patch ready. And I'm eager
to test these patches and compare the performance to the results Toke
and I got for the article based on your earlier patches[1].
[1]
https://git.kernel.org/pub/scm/linux/kernel/git/hawk/net-next-xdp.git/log/?h=xdp_paper01
--
Best regards,
Jesper Dangaard Brouer
MSc.CS, Principal Kernel Engineer at Red Hat
LinkedIn: http://www.linkedin.com/in/brouer
