From: "Xia, Chenbo" <chenbo.xia@intel.com>
To: David Christensen <drc@linux.vnet.ibm.com>,
"maxime.coquelin@redhat.com" <maxime.coquelin@redhat.com>,
"Wang, Zhihong" <zhihong.wang@intel.com>,
"dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] [PATCH] net/vhost: fix xstats wrong after clearing stats
Date: Fri, 11 Sep 2020 07:44:03 +0000 [thread overview]
Message-ID: <MN2PR11MB406373E9A4F25F731805C14A9C240@MN2PR11MB4063.namprd11.prod.outlook.com> (raw)
In-Reply-To: <20200902170309.16513-1-drc@linux.vnet.ibm.com>
Hi David,
Thanks for working on this. Comments inline.
> -----Original Message-----
> From: David Christensen <drc@linux.vnet.ibm.com>
> Sent: Thursday, September 3, 2020 1:03 AM
> To: maxime.coquelin@redhat.com; Xia, Chenbo <chenbo.xia@intel.com>; Wang,
> Zhihong <zhihong.wang@intel.com>; dev@dpdk.org
> Cc: David Christensen <drc@linux.vnet.ibm.com>
> Subject: [PATCH] net/vhost: fix xstats wrong after clearing stats
>
> The PMD API allows stats and xstats values to be cleared separately.
> This is a problem for the vhost PMD since some of the xstats values are
> derived from existing stats values. For example:
>
> testpmd> show port xstats all
> ...
> tx_unicast_packets: 17562959
> ...
> testpmd> clear port stats all
> ...
> show port xstats all
> ...
> tx_unicast_packets: 18446744073709551615
> ...
>
> Modify the driver so that stats and xstats values are stored, updated,
> and cleared separately.
I think it's fix patch. So please add 'Fixes:XXX' and cc to stable@dpdk.org
in your commit message.
>
> Signed-off-by: David Christensen <drc@linux.vnet.ibm.com>
> ---
> drivers/net/vhost/rte_eth_vhost.c | 54 ++++++++++++++++++-------------
> 1 file changed, 32 insertions(+), 22 deletions(-)
>
> diff --git a/drivers/net/vhost/rte_eth_vhost.c
> b/drivers/net/vhost/rte_eth_vhost.c
> index e55278af6..4e72cc2ca 100644
> --- a/drivers/net/vhost/rte_eth_vhost.c
> +++ b/drivers/net/vhost/rte_eth_vhost.c
> @@ -73,6 +73,9 @@ enum vhost_xstats_pkts {
> VHOST_BROADCAST_PKT,
> VHOST_MULTICAST_PKT,
> VHOST_UNICAST_PKT,
> + VHOST_PKT,
> + VHOST_BYTE,
> + VHOST_MISSED_PKT,
> VHOST_ERRORS_PKT,
> VHOST_ERRORS_FRAGMENTED,
> VHOST_ERRORS_JABBER,
> @@ -149,11 +152,11 @@ struct vhost_xstats_name_off {
> /* [rx]_is prepended to the name string here */
> static const struct vhost_xstats_name_off vhost_rxport_stat_strings[] = {
> {"good_packets",
> - offsetof(struct vhost_queue, stats.pkts)},
> + offsetof(struct vhost_queue, stats.xstats[VHOST_PKT])},
> {"total_bytes",
> - offsetof(struct vhost_queue, stats.bytes)},
> + offsetof(struct vhost_queue, stats.xstats[VHOST_BYTE])},
> {"missed_pkts",
> - offsetof(struct vhost_queue, stats.missed_pkts)},
> + offsetof(struct vhost_queue, stats.xstats[VHOST_MISSED_PKT])},
> {"broadcast_packets",
> offsetof(struct vhost_queue, stats.xstats[VHOST_BROADCAST_PKT])},
> {"multicast_packets",
> @@ -189,11 +192,11 @@ static const struct vhost_xstats_name_off
> vhost_rxport_stat_strings[] = {
> /* [tx]_ is prepended to the name string here */
> static const struct vhost_xstats_name_off vhost_txport_stat_strings[] = {
> {"good_packets",
> - offsetof(struct vhost_queue, stats.pkts)},
> + offsetof(struct vhost_queue, stats.xstats[VHOST_PKT])},
> {"total_bytes",
> - offsetof(struct vhost_queue, stats.bytes)},
> + offsetof(struct vhost_queue, stats.xstats[VHOST_BYTE])},
> {"missed_pkts",
> - offsetof(struct vhost_queue, stats.missed_pkts)},
> + offsetof(struct vhost_queue, stats.xstats[VHOST_MISSED_PKT])},
> {"broadcast_packets",
> offsetof(struct vhost_queue, stats.xstats[VHOST_BROADCAST_PKT])},
> {"multicast_packets",
> @@ -291,18 +294,11 @@ vhost_dev_xstats_get(struct rte_eth_dev *dev, struct
> rte_eth_xstat *xstats,
> vq = dev->data->rx_queues[i];
> if (!vq)
> continue;
> - vq->stats.xstats[VHOST_UNICAST_PKT] = vq->stats.pkts
> - - (vq->stats.xstats[VHOST_BROADCAST_PKT]
> - + vq->stats.xstats[VHOST_MULTICAST_PKT]);
> }
Why not delete the for loop here?
> for (i = 0; i < dev->data->nb_tx_queues; i++) {
> vq = dev->data->tx_queues[i];
> if (!vq)
> continue;
> - vq->stats.xstats[VHOST_UNICAST_PKT] = vq->stats.pkts
> - + vq->stats.missed_pkts
> - - (vq->stats.xstats[VHOST_BROADCAST_PKT]
> - + vq->stats.xstats[VHOST_MULTICAST_PKT]);
> }
Ditto.
> for (t = 0; t < VHOST_NB_XSTATS_RXPORT; t++) {
> xstats[count].value = 0;
> @@ -346,20 +342,27 @@ vhost_count_multicast_broadcast(struct vhost_queue
> *vq,
> pstats->xstats[VHOST_BROADCAST_PKT]++;
> else
> pstats->xstats[VHOST_MULTICAST_PKT]++;
> + } else {
> + pstats->xstats[VHOST_UNICAST_PKT]++;
As this function also count unicast pkts now. The function name should better
be changed. Besides, in 'eth_vhost_tx' which calls this function, there's a
comment about why we calls the function. I think that should also be updated.
Thanks!
Chenbo
> }
> }
>
> static void
> -vhost_update_packet_xstats(struct vhost_queue *vq,
> - struct rte_mbuf **bufs,
> - uint16_t count)
> +vhost_update_packet_xstats(struct vhost_queue *vq, struct rte_mbuf **bufs,
> + uint16_t count, uint64_t nb_bytes,
> + uint64_t nb_missed)
> {
> uint32_t pkt_len = 0;
> uint64_t i = 0;
> uint64_t index;
> struct vhost_stats *pstats = &vq->stats;
>
> + pstats->xstats[VHOST_BYTE] += nb_bytes;
> + pstats->xstats[VHOST_MISSED_PKT] += nb_missed;
> + pstats->xstats[VHOST_UNICAST_PKT] += nb_missed;
> +
> for (i = 0; i < count ; i++) {
> + pstats->xstats[VHOST_PKT]++;
> pkt_len = bufs[i]->pkt_len;
> if (pkt_len == 64) {
> pstats->xstats[VHOST_64_PKT]++;
> @@ -385,6 +388,7 @@ eth_vhost_rx(void *q, struct rte_mbuf **bufs, uint16_t
> nb_bufs)
> struct vhost_queue *r = q;
> uint16_t i, nb_rx = 0;
> uint16_t nb_receive = nb_bufs;
> + uint64_t nb_bytes = 0;
>
> if (unlikely(rte_atomic32_read(&r->allow_queuing) == 0))
> return 0;
> @@ -419,10 +423,11 @@ eth_vhost_rx(void *q, struct rte_mbuf **bufs,
> uint16_t nb_bufs)
> if (r->internal->vlan_strip)
> rte_vlan_strip(bufs[i]);
>
> - r->stats.bytes += bufs[i]->pkt_len;
> + nb_bytes += bufs[i]->pkt_len;
> }
>
> - vhost_update_packet_xstats(r, bufs, nb_rx);
> + r->stats.bytes += nb_bytes;
> + vhost_update_packet_xstats(r, bufs, nb_rx, nb_bytes, 0);
>
> out:
> rte_atomic32_set(&r->while_queuing, 0);
> @@ -436,6 +441,8 @@ eth_vhost_tx(void *q, struct rte_mbuf **bufs, uint16_t
> nb_bufs)
> struct vhost_queue *r = q;
> uint16_t i, nb_tx = 0;
> uint16_t nb_send = 0;
> + uint64_t nb_bytes = 0;
> + uint64_t nb_missed = 0;
>
> if (unlikely(rte_atomic32_read(&r->allow_queuing) == 0))
> return 0;
> @@ -476,13 +483,16 @@ eth_vhost_tx(void *q, struct rte_mbuf **bufs,
> uint16_t nb_bufs)
> break;
> }
>
> + for (i = 0; likely(i < nb_tx); i++)
> + nb_bytes += bufs[i]->pkt_len;
> +
> + nb_missed = nb_bufs - nb_tx;
> +
> r->stats.pkts += nb_tx;
> + r->stats.bytes += nb_bytes;
> r->stats.missed_pkts += nb_bufs - nb_tx;
>
> - for (i = 0; likely(i < nb_tx); i++)
> - r->stats.bytes += bufs[i]->pkt_len;
> -
> - vhost_update_packet_xstats(r, bufs, nb_tx);
> + vhost_update_packet_xstats(r, bufs, nb_tx, nb_bytes, nb_missed);
>
> /* According to RFC2863 page42 section ifHCOutMulticastPkts and
> * ifHCOutBroadcastPkts, the counters "multicast" and "broadcast"
> --
> 2.18.4
next prev parent reply other threads:[~2020-09-11 7:44 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-02 17:03 David Christensen
2020-09-11 7:44 ` Xia, Chenbo [this message]
2020-09-23 8:07 ` Maxime Coquelin
2020-10-05 17:43 ` David Christensen
2020-10-06 7:32 ` Maxime Coquelin
2020-10-06 21:23 ` [dpdk-dev] [PATCH v2] " David Christensen
2020-10-09 3:13 ` Xia, Chenbo
2020-10-15 17:49 ` [dpdk-dev] [PATCH v3] " David Christensen
2020-10-16 1:38 ` Xia, Chenbo
2020-10-23 10:54 ` Maxime Coquelin
2020-10-23 11:22 ` Maxime Coquelin
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=MN2PR11MB406373E9A4F25F731805C14A9C240@MN2PR11MB4063.namprd11.prod.outlook.com \
--to=chenbo.xia@intel.com \
--cc=dev@dpdk.org \
--cc=drc@linux.vnet.ibm.com \
--cc=maxime.coquelin@redhat.com \
--cc=zhihong.wang@intel.com \
/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).