> On Sun, 16 Feb 2020 22:07:32 +0100
> Lorenzo Bianconi <lorenzo@xxxxxxxxxx> wrote:
>
> > @@ -2033,6 +2050,7 @@ mvneta_xdp_submit_frame(struct mvneta_port *pp, struct mvneta_tx_queue *txq,
> > u64_stats_update_begin(&stats->syncp);
> > stats->es.ps.tx_bytes += xdpf->len;
> > stats->es.ps.tx_packets++;
> > + stats->es.ps.xdp_tx++;
> > u64_stats_update_end(&stats->syncp);
>
> I find it confusing that this ethtool stats is named "xdp_tx".
> Because you use it as an "xmit" counter and not for the action XDP_TX.
>
> Both XDP_TX and XDP_REDIRECT out this device will increment this
> "xdp_tx" counter. I don't think end-users will comprehend this...
>
> What about naming it "xdp_xmit" ?
Hi Jesper,
yes, I think it is definitely better. So to follow up:
- rename current "xdp_tx" counter in "xdp_xmit" and increment it for
XDP_TX verdict and for ndo_xdp_xmit
- introduce a new "xdp_tx" counter only for XDP_TX verdict.
If we agree I can post a follow-up patch.
Regards,
Lorenzo
>
>
> > mvneta_txq_inc_put(txq);
> > @@ -2114,6 +2132,7 @@ mvneta_run_xdp(struct mvneta_port *pp, struct mvneta_rx_queue *rxq,
> >
> > switch (act) {
> > case XDP_PASS:
> > + stats->xdp_pass++;
> > return MVNETA_XDP_PASS;
> > case XDP_REDIRECT: {
> > int err;
> > @@ -2126,6 +2145,7 @@ mvneta_run_xdp(struct mvneta_port *pp, struct mvneta_rx_queue *rxq,
> > len, true);
> > } else {
> > ret = MVNETA_XDP_REDIR;
> > + stats->xdp_redirect++;
> > }
> > break;
> > }
> > @@ -2147,6 +2167,7 @@ mvneta_run_xdp(struct mvneta_port *pp, struct mvneta_rx_queue *rxq,
> > virt_to_head_page(xdp->data),
> > len, true);
> > ret = MVNETA_XDP_DROPPED;
> > + stats->xdp_drop++;
> > break;
> > }
>
>
> --
> Best regards,
> Jesper Dangaard Brouer
> MSc.CS, Principal Kernel Engineer at Red Hat
> LinkedIn: http://www.linkedin.com/in/brouer
>
Attachment:
signature.asc
Description: PGP signature
