Re: [PATCH net-next] net: mvneta: align xdp stats naming scheme to mlx5 driver

[Jesper Dangaard Brouer <brouer@redhat.com>]

On Wed, 19 Feb 2020 10:57:37 +0100
Lorenzo Bianconi <lorenzo@kernel.org> wrote:

> Introduce "rx" prefix in the name scheme for xdp counters
> on rx path.
> Differentiate between XDP_TX and ndo_xdp_xmit counters
> 
> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>

Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>

> ---
> Changes since RFC:
> - rename rx_xdp_tx_xmit in rx_xdp_tx
> - move tx stats accounting in mvneta_xdp_xmit_back/mvneta_xdp_xmit

I like how you managed to bulk update the ndo_xdp_xmit counters.

> ---
>  drivers/net/ethernet/marvell/mvneta.c | 52 ++++++++++++++++++---------
>  1 file changed, 36 insertions(+), 16 deletions(-)
> 
> diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
> index b7045b6a15c2..8e1feb678cea 100644
> --- a/drivers/net/ethernet/marvell/mvneta.c
> +++ b/drivers/net/ethernet/marvell/mvneta.c
> @@ -344,6 +344,7 @@ enum {
>  	ETHTOOL_XDP_REDIRECT,
>  	ETHTOOL_XDP_PASS,
>  	ETHTOOL_XDP_DROP,
> +	ETHTOOL_XDP_XMIT,
>  	ETHTOOL_XDP_TX,
>  	ETHTOOL_MAX_STATS,
>  };
> @@ -399,10 +400,11 @@ static const struct mvneta_statistic mvneta_statistics[] = {
>  	{ ETHTOOL_STAT_EEE_WAKEUP, T_SW, "eee_wakeup_errors", },
>  	{ ETHTOOL_STAT_SKB_ALLOC_ERR, T_SW, "skb_alloc_errors", },
>  	{ ETHTOOL_STAT_REFILL_ERR, T_SW, "refill_errors", },
> -	{ ETHTOOL_XDP_REDIRECT, T_SW, "xdp_redirect", },
> -	{ ETHTOOL_XDP_PASS, T_SW, "xdp_pass", },
> -	{ ETHTOOL_XDP_DROP, T_SW, "xdp_drop", },
> -	{ ETHTOOL_XDP_TX, T_SW, "xdp_tx", },
> +	{ ETHTOOL_XDP_REDIRECT, T_SW, "rx_xdp_redirect", },
> +	{ ETHTOOL_XDP_PASS, T_SW, "rx_xdp_pass", },
> +	{ ETHTOOL_XDP_DROP, T_SW, "rx_xdp_drop", },
> +	{ ETHTOOL_XDP_TX, T_SW, "rx_xdp_tx", },
> +	{ ETHTOOL_XDP_XMIT, T_SW, "tx_xdp_xmit", },
>  };

I'm fine with these names.

Based on the code, the "rx_xdp_tx" counter is counting the successful
XDP_TX operations, which I actually like.  I could use this to debug,
if my XDP_TX operations are drop packets (due to TX-queue full), by
using fexit or XDP-prog counter to deduct this.  I think it would be
valuable to provide a "rx_xdp_tx_err" counter, that would detect this,
it makes it easier to detect this for end-users (as a followup patch).


>  struct mvneta_stats {
> @@ -414,6 +416,7 @@ struct mvneta_stats {
>  	u64	xdp_redirect;
>  	u64	xdp_pass;
>  	u64	xdp_drop;
> +	u64	xdp_xmit;
>  	u64	xdp_tx;
>  };
>  
> @@ -2012,7 +2015,6 @@ static int
>  mvneta_xdp_submit_frame(struct mvneta_port *pp, struct mvneta_tx_queue *txq,
>  			struct xdp_frame *xdpf, bool dma_map)
>  {
> -	struct mvneta_pcpu_stats *stats = this_cpu_ptr(pp->stats);
>  	struct mvneta_tx_desc *tx_desc;
>  	struct mvneta_tx_buf *buf;
>  	dma_addr_t dma_addr;
> @@ -2047,12 +2049,6 @@ mvneta_xdp_submit_frame(struct mvneta_port *pp, struct mvneta_tx_queue *txq,
>  	tx_desc->buf_phys_addr = dma_addr;
>  	tx_desc->data_size = xdpf->len;
>  
> -	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);
> -
>  	mvneta_txq_inc_put(txq);
>  	txq->pending++;
>  	txq->count++;
> @@ -2079,8 +2075,17 @@ mvneta_xdp_xmit_back(struct mvneta_port *pp, struct xdp_buff *xdp)
>  
>  	__netif_tx_lock(nq, cpu);
>  	ret = mvneta_xdp_submit_frame(pp, txq, xdpf, false);
> -	if (ret == MVNETA_XDP_TX)
> +	if (ret == MVNETA_XDP_TX) {
> +		struct mvneta_pcpu_stats *stats = this_cpu_ptr(pp->stats);
> +
> +		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);
> +
>  		mvneta_txq_pend_desc_add(pp, txq, 0);
> +	}
>  	__netif_tx_unlock(nq);
>  
>  	return ret;
> @@ -2091,10 +2096,11 @@ mvneta_xdp_xmit(struct net_device *dev, int num_frame,
>  		struct xdp_frame **frames, u32 flags)
>  {
>  	struct mvneta_port *pp = netdev_priv(dev);
> +	struct mvneta_pcpu_stats *stats = this_cpu_ptr(pp->stats);
> +	int i, nxmit_byte = 0, nxmit = num_frame;
>  	int cpu = smp_processor_id();
>  	struct mvneta_tx_queue *txq;
>  	struct netdev_queue *nq;
> -	int i, drops = 0;
>  	u32 ret;
>  
>  	if (unlikely(flags & ~XDP_XMIT_FLAGS_MASK))
> @@ -2106,9 +2112,11 @@ mvneta_xdp_xmit(struct net_device *dev, int num_frame,
>  	__netif_tx_lock(nq, cpu);
>  	for (i = 0; i < num_frame; i++) {
>  		ret = mvneta_xdp_submit_frame(pp, txq, frames[i], true);
> -		if (ret != MVNETA_XDP_TX) {
> +		if (ret == MVNETA_XDP_TX) {
> +			nxmit_byte += frames[i]->len;
> +		} else {
>  			xdp_return_frame_rx_napi(frames[i]);
> -			drops++;
> +			nxmit--;
>  		}
>  	}
>  
> @@ -2116,7 +2124,13 @@ mvneta_xdp_xmit(struct net_device *dev, int num_frame,
>  		mvneta_txq_pend_desc_add(pp, txq, 0);
>  	__netif_tx_unlock(nq);
>  
> -	return num_frame - drops;
> +	u64_stats_update_begin(&stats->syncp);
> +	stats->es.ps.tx_bytes += nxmit_byte;
> +	stats->es.ps.tx_packets += nxmit;
> +	stats->es.ps.xdp_xmit += nxmit;
> +	u64_stats_update_end(&stats->syncp);
> +
> +	return nxmit;
>  }

Updating xmit counter per bulk, nice.

>  
[...]

-- 
Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  LinkedIn: http://www.linkedin.com/in/brouer