From: "Stefan Lässer" <stefan.laesser@omicronenergy.com>
To: Stephen Hemminger <stephen@networkplumber.org>
Cc: "dev@dpdk.org" <dev@dpdk.org>
Subject: RE: [PATCH v2] net/af_packet: provide packet drop stats
Date: Wed, 22 Jan 2025 07:03:12 +0000 [thread overview]
Message-ID: <AM0PR03MB6275A20F54A11C666ABD843196E12@AM0PR03MB6275.eurprd03.prod.outlook.com> (raw)
In-Reply-To: <20250121074815.536fbaf9@hermes.local>
> -----Original Message-----
> From: Stephen Hemminger <stephen@networkplumber.org>
> Sent: Tuesday, January 21, 2025 4:48 PM
> To: Stefan Lässer <stefan.laesser@omicronenergy.com>
> Cc: dev@dpdk.org
> Subject: Re: [PATCH v2] net/af_packet: provide packet drop stats
>
> On Tue, 21 Jan 2025 13:26:16 +0100
> Stefan Laesser <stefan.laesser@omicronenergy.com> wrote:
>
> > + /* query dropped packets counter from socket */
> > + if (internal->rx_queue[i].sockfd != -1 &&
> > + getsockopt(internal->rx_queue[i].sockfd, SOL_PACKET,
> > + PACKET_STATISTICS,
> &iface_stats,
> > + &iface_stats_len) > -1) {
> > + /*
> > + * keep total because each call to getsocketopt with
> PACKET_STATISTICS
> > + * reset the counter of the socket
> > + */
> > + internal->rx_queue[i].rx_dropped_pkts +=
> iface_stats.tp_drops;
> > + }
>
> Since reading the value clears it, maybe stats_reset should call as well?
>
> Also, queues greater that the RTE_ETHDEV_STAT_CNTRS should still count
> against the total. And the two loops could be combined.
>
> Maybe something like this?
That looks very reasonable - I will update my patch.
>
> diff --git a/drivers/net/af_packet/rte_eth_af_packet.c
> b/drivers/net/af_packet/rte_eth_af_packet.c
> index 7033910df8..af08e893f1 100644
> --- a/drivers/net/af_packet/rte_eth_af_packet.c
> +++ b/drivers/net/af_packet/rte_eth_af_packet.c
> @@ -417,50 +417,51 @@ eth_dev_info(struct rte_eth_dev *dev, struct
> rte_eth_dev_info *dev_info)
> return 0;
> }
>
> +/* query dropped packets counter from socket */ static unsigned int
> +packet_drop_count(int sockfd) {
> + struct tpacket_stats pkt_stats;
> + socklen_t pkt_stats_len = sizeof(struct tpacket_stats);
> +
> + if (sockfd == -1)
> + return 0;
> +
> + if (getsockopt(sockfd, SOL_PACKET, PACKET_STATISTICS, &pkt_stats,
> &pkt_stats_len) < -1)
> + return 0;
> +
> + return pkt_stats.tp_drops;
> +}
> +
> static int
> eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) {
> - unsigned i, imax;
> + unsigned int i;
> unsigned long rx_total = 0, rx_dropped_total = 0, rx_nombuf_total = 0;
> unsigned long tx_total = 0, tx_err_total = 0;
> unsigned long rx_bytes_total = 0, tx_bytes_total = 0;
> const struct pmd_internals *internal = dev->data->dev_private;
>
> - imax = (internal->nb_queues < RTE_ETHDEV_QUEUE_STAT_CNTRS ?
> - internal->nb_queues : RTE_ETHDEV_QUEUE_STAT_CNTRS);
> - for (i = 0; i < imax; i++) {
> - struct tpacket_stats iface_stats;
> - socklen_t iface_stats_len = sizeof(struct tpacket_stats);
> -
> - /* query dropped packets counter from socket */
> - if (internal->rx_queue[i].sockfd != -1 &&
> - getsockopt(internal->rx_queue[i].sockfd, SOL_PACKET,
> - PACKET_STATISTICS,
> &iface_stats,
> - &iface_stats_len) > -1) {
> - /*
> - * keep total because each call to getsocketopt with
> PACKET_STATISTICS
> - * reset the counter of the socket
> - */
> - internal->rx_queue[i].rx_dropped_pkts +=
> iface_stats.tp_drops;
> - }
> -
> - stats->q_ipackets[i] = internal->rx_queue[i].rx_pkts;
> - stats->q_ibytes[i] = internal->rx_queue[i].rx_bytes;
> + for (i = 0; i < internal->nb_queues; i++) {
> + /* reading drop count clears the value */
> + internal->rx_queue[i].rx_dropped_pkts +=
> + packet_drop_count(internal->rx_queue[i].sockfd);
>
> rx_total += stats->q_ipackets[i];
> rx_bytes_total += stats->q_ibytes[i];
> rx_dropped_total += internal->rx_queue[i].rx_dropped_pkts;
> rx_nombuf_total += internal->rx_queue[i].rx_nombuf;
> - }
>
> - imax = (internal->nb_queues < RTE_ETHDEV_QUEUE_STAT_CNTRS ?
> - internal->nb_queues : RTE_ETHDEV_QUEUE_STAT_CNTRS);
> - for (i = 0; i < imax; i++) {
> - stats->q_opackets[i] = internal->tx_queue[i].tx_pkts;
> - stats->q_obytes[i] = internal->tx_queue[i].tx_bytes;
> tx_total += stats->q_opackets[i];
> tx_err_total += internal->tx_queue[i].err_pkts;
> tx_bytes_total += stats->q_obytes[i];
> +
> + if (i < RTE_ETHDEV_QUEUE_STAT_CNTRS) {
> + stats->q_ipackets[i] = internal->rx_queue[i].rx_pkts;
> + stats->q_ibytes[i] = internal->rx_queue[i].rx_bytes;
> + stats->q_opackets[i] = internal->tx_queue[i].tx_pkts;
> + stats->q_obytes[i] = internal->tx_queue[i].tx_bytes;
> + }
> }
>
> stats->ipackets = rx_total;
> @@ -480,6 +481,7 @@ eth_stats_reset(struct rte_eth_dev *dev)
> struct pmd_internals *internal = dev->data->dev_private;
>
> for (i = 0; i < internal->nb_queues; i++) {
> + packet_drop_count(internal->rx_queue[i].sockfd);
> internal->rx_queue[i].rx_pkts = 0;
> internal->rx_queue[i].rx_bytes = 0;
> internal->rx_queue[i].rx_nombuf = 0;
next prev parent reply other threads:[~2025-01-22 7:03 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-01-16 16:17 [PATCH] " Stefan Laesser
2025-01-16 16:24 ` Stephen Hemminger
2025-01-17 7:20 ` Stefan Lässer
2025-01-17 16:43 ` Stephen Hemminger
2025-01-21 12:26 ` [PATCH v2] " Stefan Laesser
2025-01-21 15:48 ` Stephen Hemminger
2025-01-22 7:03 ` Stefan Lässer [this message]
2025-01-22 7:28 ` [PATCH v3] " Stefan Laesser
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=AM0PR03MB6275A20F54A11C666ABD843196E12@AM0PR03MB6275.eurprd03.prod.outlook.com \
--to=stefan.laesser@omicronenergy.com \
--cc=dev@dpdk.org \
--cc=stephen@networkplumber.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).