patches for DPDK stable branches
 help / color / mirror / Atom feed
From: Kevin Traynor <ktraynor@redhat.com>
To: Alvin Zhang <alvinx.zhang@intel.com>,
	beilei.xing@intel.com, junfeng.guo@intel.com
Cc: dev@dpdk.org, stable@dpdk.org, "Zhang,
	Qi Z" <qi.z.zhang@intel.com>,
	Ferruh Yigit <ferruh.yigit@intel.com>
Subject: Re: [dpdk-stable] [dpdk-dev] [PATCH] net/i40e: fix Rx packet statistics
Date: Mon, 27 Sep 2021 17:00:17 +0100
Message-ID: <863055cd-d436-6bdd-e6fe-2aa4d57db19a@redhat.com> (raw)
In-Reply-To: <20210926075757.15116-1-alvinx.zhang@intel.com>

On 26/09/2021 08:57, Alvin Zhang wrote:
> Some packets are discarded by the NIC because they are larger than
> the MTU, these packets should be counted as "RX error" instead of
> "RX packet".
> 
> The register 'GL_RXERR1' can count above discarded packets.
> This patch adds reading and calculation of the 'GL_RXERR1' counter
> when reporting DPDK statistics.
> 
> Fixes: f4a91c38b4ad ("i40e: add extended stats")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Alvin Zhang <alvinx.zhang@intel.com>
> ---
>   drivers/net/i40e/i40e_ethdev.c | 16 +++++++++++++---
>   drivers/net/i40e/i40e_ethdev.h | 10 ++++++++++
>   2 files changed, 23 insertions(+), 3 deletions(-)
> 

It's a bit hard to understand the code for someone not familiar with the 
i40e stats. I think it needs careful review from i40e maintainers. A few 
questions below,

Did you test this with testpmd? Can you show an example of a test where 
these packets are now correctly accounted for?

I see there is also an RXERR2 register that catches other errors, does 
it need to be considered as well?

> diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c
> index 7a2a828..30a2cdf 100644
> --- a/drivers/net/i40e/i40e_ethdev.c
> +++ b/drivers/net/i40e/i40e_ethdev.c
> @@ -532,7 +532,7 @@ static int i40e_sw_tunnel_filter_insert(struct i40e_pf *pf,
>   /* store statistics names and its offset in stats structure */
>   struct rte_i40e_xstats_name_off {
>   	char name[RTE_ETH_XSTATS_NAME_SIZE];
> -	unsigned offset;
> +	int offset;

It is unusual to see you changing an offset to an int. You are expecting 
negative offsets?

>   };
>   
>   static const struct rte_i40e_xstats_name_off rte_i40e_stats_strings[] = {
> @@ -542,6 +542,8 @@ struct rte_i40e_xstats_name_off {
>   	{"rx_dropped_packets", offsetof(struct i40e_eth_stats, rx_discards)},
>   	{"rx_unknown_protocol_packets", offsetof(struct i40e_eth_stats,
>   		rx_unknown_protocol)},
> +	{"rx_err1", offsetof(struct i40e_pf, rx_err1) -
> +		    offsetof(struct i40e_pf, stats)},

rx_err1 is correct by datasheet but meaningless to a user. Suggest to 
find a more descriptive name, or document what it is, or tell the user 
to reference the datasheet.

>   	{"tx_unicast_packets", offsetof(struct i40e_eth_stats, tx_unicast)},
>   	{"tx_multicast_packets", offsetof(struct i40e_eth_stats, tx_multicast)},
>   	{"tx_broadcast_packets", offsetof(struct i40e_eth_stats, tx_broadcast)},
> @@ -3238,6 +3240,10 @@ void i40e_flex_payload_reg_set_default(struct i40e_hw *hw)
>   			    pf->offset_loaded,
>   			    &os->eth.rx_unknown_protocol,
>   			    &ns->eth.rx_unknown_protocol);
> +	i40e_stat_update_48(hw, I40E_GL_RXERR1_H(hw->pf_id + I40E_MAX_VF),
> +			    I40E_GL_RXERR1_L(hw->pf_id + I40E_MAX_VF),
> +			    pf->offset_loaded, &pf->rx_err1_offset,
> +			    &pf->rx_err1);
>   	i40e_stat_update_48_in_64(hw, I40E_GLPRT_GOTCH(hw->port),
>   				  I40E_GLPRT_GOTCL(hw->port),
>   				  pf->offset_loaded, &os->eth.tx_bytes,
> @@ -3437,7 +3443,8 @@ void i40e_flex_payload_reg_set_default(struct i40e_hw *hw)
>   	stats->ipackets = pf->main_vsi->eth_stats.rx_unicast +
>   			pf->main_vsi->eth_stats.rx_multicast +
>   			pf->main_vsi->eth_stats.rx_broadcast -
> -			pf->main_vsi->eth_stats.rx_discards;
> +			pf->main_vsi->eth_stats.rx_discards -
> +			pf->rx_err1;
>   	stats->opackets = ns->eth.tx_unicast +
>   			ns->eth.tx_multicast +
>   			ns->eth.tx_broadcast;
> @@ -3451,7 +3458,8 @@ void i40e_flex_payload_reg_set_default(struct i40e_hw *hw)
>   			pf->main_vsi->eth_stats.rx_discards;
>   	stats->ierrors  = ns->crc_errors +
>   			ns->rx_length_errors + ns->rx_undersize +
> -			ns->rx_oversize + ns->rx_fragments + ns->rx_jabber;
> +			ns->rx_oversize + ns->rx_fragments + ns->rx_jabber +
> +			pf->rx_err1;
>   
>   	if (pf->vfs) {
>   		for (i = 0; i < pf->vf_num; i++) {
> @@ -6232,6 +6240,8 @@ struct i40e_vsi *
>   	memset(&pf->stats_offset, 0, sizeof(struct i40e_hw_port_stats));
>   	memset(&pf->internal_stats, 0, sizeof(struct i40e_eth_stats));
>   	memset(&pf->internal_stats_offset, 0, sizeof(struct i40e_eth_stats));
> +	pf->rx_err1 = 0;
> +	pf->rx_err1_offset = 0;
>   
>   	ret = i40e_pf_get_switch_config(pf);
>   	if (ret != I40E_SUCCESS) {
> diff --git a/drivers/net/i40e/i40e_ethdev.h b/drivers/net/i40e/i40e_ethdev.h
> index cd6deab..846c8d4 100644
> --- a/drivers/net/i40e/i40e_ethdev.h
> +++ b/drivers/net/i40e/i40e_ethdev.h
> @@ -19,6 +19,13 @@
>   #include "base/i40e_type.h"
>   #include "base/virtchnl.h"
>   
> +#define I40E_GL_RXERR1_H(_i)	(0x00318004 + ((_i) * 8))
> +/**
> + * _i=0...143,
> + * counters 0-127 are for the 128 VFs,
> + * counters 128-143 are for the 16 PFs
> + */
> +
>   #define I40E_VLAN_TAG_SIZE        4
>   
>   #define I40E_AQ_LEN               32
> @@ -1134,6 +1141,9 @@ struct i40e_pf {
>   
>   	struct i40e_hw_port_stats stats_offset;
>   	struct i40e_hw_port_stats stats;
> +	u64 rx_err1;	/* rxerr1 */
> +	u64 rx_err1_offset;
> +
>   	/* internal packet statistics, it should be excluded from the total */
>   	struct i40e_eth_stats internal_stats_offset;
>   	struct i40e_eth_stats internal_stats;
> 


  reply	other threads:[~2021-09-27 16:00 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-26  7:57 [dpdk-stable] " Alvin Zhang
2021-09-27 16:00 ` Kevin Traynor [this message]
2021-09-28  2:12   ` [dpdk-stable] [dpdk-dev] " Zhang, AlvinX
2021-09-28  8:53     ` Kevin Traynor
2021-09-28  3:21 ` [dpdk-stable] [PATCH v2] " Alvin Zhang
2021-09-28  3:22 ` Alvin Zhang
2021-09-28 11:16   ` [dpdk-stable] [dpdk-dev] " Zhang, Qi Z
2021-09-29  1:33     ` Zhang, AlvinX
2021-09-30  5:19   ` [dpdk-stable] [PATCH v3] " Alvin Zhang
2021-10-08  6:17     ` Zhang, Qi Z

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=863055cd-d436-6bdd-e6fe-2aa4d57db19a@redhat.com \
    --to=ktraynor@redhat.com \
    --cc=alvinx.zhang@intel.com \
    --cc=beilei.xing@intel.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=junfeng.guo@intel.com \
    --cc=qi.z.zhang@intel.com \
    --cc=stable@dpdk.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

patches for DPDK stable branches

This inbox may be cloned and mirrored by anyone:

	git clone --mirror http://inbox.dpdk.org/stable/0 stable/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 stable stable/ http://inbox.dpdk.org/stable \
		stable@dpdk.org
	public-inbox-index stable

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.stable


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git