DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Zhang, Qi Z" <qi.z.zhang@intel.com>
To: "Jiang, JunyuX" <junyux.jiang@intel.com>, "dev@dpdk.org" <dev@dpdk.org>
Cc: "Yang, Qiming" <qiming.yang@intel.com>, "Su, Simei" <simei.su@intel.com>
Subject: Re: [dpdk-dev] [PATCH v2] net/ice: support based RSS configure
Date: Mon, 22 Jun 2020 14:36:33 +0000
Message-ID: <039ED4275CED7440929022BC67E7061154842F8A@SHSMSX103.ccr.corp.intel.com> (raw)
In-Reply-To: <20200622053312.46719-1-junyux.jiang@intel.com>



> -----Original Message-----
> From: Jiang, JunyuX <junyux.jiang@intel.com>
> Sent: Monday, June 22, 2020 1:33 PM
> To: dev@dpdk.org
> Cc: Yang, Qiming <qiming.yang@intel.com>; Zhang, Qi Z
> <qi.z.zhang@intel.com>; Su, Simei <simei.su@intel.com>; Jiang, JunyuX
> <junyux.jiang@intel.com>
> Subject: [PATCH v2] net/ice: support based RSS configure
The title is misleading, how about "initialize and update RSS based on user request"

> 
> Enable/disable RSS for corresponding flow base on the user's requirement

Initialize and update RSS configure based on user request (rte_eth_rss_conf) from dev_configure and .rss_hash_update ops.
All previous default configure has been removed.
.
> 
> Signed-off-by: Junyu Jiang <junyux.jiang@intel.com>
> 
> ---
> v1->v2:
> remove gtpu and pppoe/pppod configuration from rss init
> ---
>  drivers/net/ice/ice_ethdev.c | 162 +++++++++++++++++++++--------------
>  1 file changed, 96 insertions(+), 66 deletions(-)
> 
> diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c index
> 5a89a1955..cbe59a40e 100644
> --- a/drivers/net/ice/ice_ethdev.c
> +++ b/drivers/net/ice/ice_ethdev.c
> @@ -2441,6 +2441,87 @@ ice_dev_uninit(struct rte_eth_dev *dev)
>  	return 0;
>  }
> 
> +static void
> +ice_rss_hash_set(struct ice_pf *pf, uint64_t rss_hf) {
> +	struct ice_hw *hw = ICE_PF_TO_HW(pf);
> +	struct ice_vsi *vsi = pf->main_vsi;
> +	int ret;
> +
> +	/**
> +	 * configure RSS for IPv4 with input set IPv4 src/dst
> +	 * configure RSS for IPv6 with input set IPv6 src/dst

The comment looks redundant, please merge into one line.
	/* Configure RSS for IP with src/dst address as input set */

> +	 */
> +	if (rss_hf & ETH_RSS_IP) {

This is not correct, it is possible user only want IPv4 but not IPv6.

For ice, I think we can do like below

If (rss_hf & ETH_RSS_IPV4) {
	Ice_add_rss_cfg (... ICE_FLOW_HASH_IPV4 ...)
}
If (rss_hf & ETH_RSS_IPV6) {
	Ice_add_rss_cfg (... ICE_FLOW_HASH_IPv6 ...)
}
We can just ignore ETH_RSS_FRAG_IPV4 and ETH_RSS_NONFRAG_IPV4_OTHER and same for UDP/TCP and SCTP.


> +		ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV4,
> +				      ICE_FLOW_SEG_HDR_IPV4, 0);

You need ICE_FLOW_SEG_HDR_IPV_OTHER, see patch 
http://patchwork.dpdk.org/patch/71584/


> +		if (ret)
> +			PMD_DRV_LOG(ERR, "%s IPV4 rss flow fail %d",
> +				    __func__, ret);
> +
> +		ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV6,
> +				      ICE_FLOW_SEG_HDR_IPV6, 0);
> +		if (ret)
> +			PMD_DRV_LOG(ERR, "%s IPV6 rss flow fail %d",
> +				    __func__, ret);
> +	}
> +	/**
> +	 *configure RSS for udp6 with input set IPv6 src/dst, UDP src/dst
> +	 *configure RSS for udp4 with input set IP src/dst, UDP src/dst
> +	 */
> +	if (rss_hf & ETH_RSS_UDP) {
> +		ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_UDP_IPV6,
> +				      ICE_FLOW_SEG_HDR_UDP |
> +				      ICE_FLOW_SEG_HDR_IPV6, 0);
> +		if (ret)
> +			PMD_DRV_LOG(ERR, "%s UDP_IPV6 rss flow fail %d",
> +				    __func__, ret);
> +
> +		ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_UDP_IPV4,
> +				      ICE_FLOW_SEG_HDR_UDP |
> +				      ICE_FLOW_SEG_HDR_IPV4, 0);
> +		if (ret)
> +			PMD_DRV_LOG(ERR, "%s UDP_IPV4 rss flow fail %d",
> +				    __func__, ret);
> +	}
> +	/**
> +	 * configure RSS for tcp6 with input set IPv6 src/dst, TCP src/dst
> +	 * configure RSS for tcp4 with input set IP src/dst, TCP src/dst
> +	 */
> +	if (rss_hf & ETH_RSS_TCP) {
> +		ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_TCP_IPV6,
> +				      ICE_FLOW_SEG_HDR_TCP |
> +				      ICE_FLOW_SEG_HDR_IPV6, 0);
> +		if (ret)
> +			PMD_DRV_LOG(ERR, "%s TCP_IPV6 rss flow fail %d",
> +				    __func__, ret);
> +		ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_TCP_IPV4,
> +				      ICE_FLOW_SEG_HDR_TCP |
> +				      ICE_FLOW_SEG_HDR_IPV4, 0);
> +		if (ret)
> +			PMD_DRV_LOG(ERR, "%s TCP_IPV4 rss flow fail %d",
> +				    __func__, ret);
> +	}
> +	/**
> +	 * configure RSS for sctp6 with input set IPv6 src/dst
> +	 * configure RSS for sctp4 with input set IP src/dst
> +	 */
> +	if (rss_hf & ETH_RSS_SCTP) {
> +		ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV6,
> +				      ICE_FLOW_SEG_HDR_SCTP |
> +				      ICE_FLOW_SEG_HDR_IPV6, 0);
> +		if (ret)
> +			PMD_DRV_LOG(ERR, "%s SCTP_IPV6 rss flow fail %d",
> +				    __func__, ret);
> +		ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV4,
> +				      ICE_FLOW_SEG_HDR_SCTP |
> +				      ICE_FLOW_SEG_HDR_IPV4, 0);
> +		if (ret)
> +			PMD_DRV_LOG(ERR, "%s SCTP_IPV4 rss flow fail %d",
> +				    __func__, ret);
> +	}
> +}
> +
>  static int ice_init_rss(struct ice_pf *pf)  {
>  	struct ice_hw *hw = ICE_PF_TO_HW(pf);
> @@ -2501,72 +2582,9 @@ static int ice_init_rss(struct ice_pf *pf)
>  		(1 << VSIQF_HASH_CTL_HASH_SCHEME_S);
>  	ICE_WRITE_REG(hw, VSIQF_HASH_CTL(vsi->vsi_id), reg);
> 
> -	/* configure RSS for IPv4 with input set IPv4 src/dst */
> -	ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV4,
> -			      ICE_FLOW_SEG_HDR_IPV4, 0);
> -	if (ret)
> -		PMD_DRV_LOG(ERR, "%s IPV4 rss flow fail %d", __func__, ret);
> -
> -	/* configure RSS for IPv6 with input set IPv6 src/dst */
> -	ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV6,
> -			      ICE_FLOW_SEG_HDR_IPV6, 0);
> -	if (ret)
> -		PMD_DRV_LOG(ERR, "%s IPV6 rss flow fail %d", __func__, ret);
> -
> -	/* configure RSS for tcp6 with input set IPv6 src/dst, TCP src/dst */
> -	ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_TCP_IPV6,
> -			      ICE_FLOW_SEG_HDR_TCP | ICE_FLOW_SEG_HDR_IPV6,
> 0);
> -	if (ret)
> -		PMD_DRV_LOG(ERR, "%s TCP_IPV6 rss flow fail %d", __func__, ret);
> -
> -	/* configure RSS for udp6 with input set IPv6 src/dst, UDP src/dst */
> -	ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_UDP_IPV6,
> -			      ICE_FLOW_SEG_HDR_UDP | ICE_FLOW_SEG_HDR_IPV6,
> 0);
> -	if (ret)
> -		PMD_DRV_LOG(ERR, "%s UDP_IPV6 rss flow fail %d", __func__, ret);
> -
> -	/* configure RSS for sctp6 with input set IPv6 src/dst */
> -	ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV6,
> -			      ICE_FLOW_SEG_HDR_SCTP | ICE_FLOW_SEG_HDR_IPV6,
> 0);
> -	if (ret)
> -		PMD_DRV_LOG(ERR, "%s SCTP_IPV6 rss flow fail %d",
> -				__func__, ret);
> -
> -	/* configure RSS for tcp4 with input set IP src/dst, TCP src/dst */
> -	ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_TCP_IPV4,
> -			      ICE_FLOW_SEG_HDR_TCP | ICE_FLOW_SEG_HDR_IPV4,
> 0);
> -	if (ret)
> -		PMD_DRV_LOG(ERR, "%s TCP_IPV4 rss flow fail %d", __func__, ret);
> -
> -	/* configure RSS for udp4 with input set IP src/dst, UDP src/dst */
> -	ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_UDP_IPV4,
> -			      ICE_FLOW_SEG_HDR_UDP | ICE_FLOW_SEG_HDR_IPV4,
> 0);
> -	if (ret)
> -		PMD_DRV_LOG(ERR, "%s UDP_IPV4 rss flow fail %d", __func__, ret);
> -
> -	/* configure RSS for sctp4 with input set IP src/dst */
> -	ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV4,
> -			      ICE_FLOW_SEG_HDR_SCTP | ICE_FLOW_SEG_HDR_IPV4,
> 0);
> -	if (ret)
> -		PMD_DRV_LOG(ERR, "%s SCTP_IPV4 rss flow fail %d",
> -				__func__, ret);
> -
> -	/* configure RSS for gtpu with input set TEID */
> -	ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_GTP_U_IPV4_TEID,
> -				ICE_FLOW_SEG_HDR_GTPU_IP, 0);
> -	if (ret)
> -		PMD_DRV_LOG(ERR, "%s GTPU_TEID rss flow fail %d",
> -				__func__, ret);
> 
> -	/**
> -	 * configure RSS for pppoe/pppod with input set
> -	 * Source MAC and Session ID
> -	 */
> -	ret = ice_add_rss_cfg(hw, vsi->idx,
> ICE_FLOW_HASH_PPPOE_SESS_ID_ETH,
> -				ICE_FLOW_SEG_HDR_PPPOE, 0);
> -	if (ret)
> -		PMD_DRV_LOG(ERR, "%s PPPoE/PPPoD_SessionID rss flow fail %d",
> -				__func__, ret);
> +	/* RSS hash configuration */
> +	ice_rss_hash_set(pf, rss_conf->rss_hf);
> 
>  	return 0;
>  }
> @@ -3680,6 +3698,7 @@ ice_rss_hash_update(struct rte_eth_dev *dev,  {
>  	enum ice_status status = ICE_SUCCESS;
>  	struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
> +	struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
>  	struct ice_vsi *vsi = pf->main_vsi;
> 
>  	/* set hash key */
> @@ -3687,7 +3706,18 @@ ice_rss_hash_update(struct rte_eth_dev *dev,
>  	if (status)
>  		return status;
> 
> -	/* TODO: hash enable config, ice_add_rss_cfg */
> +	if (rss_conf->rss_hf == 0)
> +		return -EINVAL;
> +
> +	status = ice_rem_vsi_rss_cfg(hw, vsi->idx);
> +	if (status != ICE_SUCCESS) {
> +		PMD_DRV_LOG(ERR, "Failed to remove rss cfg!");
> +		return status;
> +	}

No need to remove exist configure, 
we just need to make sure all the configure from rss_hf has been applied.

> +
> +	/* RSS hash configuration */
> +	ice_rss_hash_set(pf, rss_conf->rss_hf);
> +
>  	return 0;
>  }
> 
> --
> 2.17.1


  reply	other threads:[~2020-06-22 14:36 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-10  6:33 [dpdk-dev] [PATCH] " Junyu Jiang
2020-06-22  5:33 ` [dpdk-dev] [PATCH v2] " Junyu Jiang
2020-06-22 14:36   ` Zhang, Qi Z [this message]
2020-06-23  2:47     ` Jiang, JunyuX
2020-07-03  8:16   ` Zhao1, Wei
2020-07-03  8:46     ` Zhao1, Wei
2020-06-23  8:33 ` [dpdk-dev] [PATCH v3] net/ice: initialize and update RSS based on user request Junyu Jiang
2020-06-23  9:40   ` Zhang, Qi Z
2020-06-24  2:09 ` [dpdk-dev] [PATCH v4] " Junyu Jiang
2020-06-24  7:30   ` 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=039ED4275CED7440929022BC67E7061154842F8A@SHSMSX103.ccr.corp.intel.com \
    --to=qi.z.zhang@intel.com \
    --cc=dev@dpdk.org \
    --cc=junyux.jiang@intel.com \
    --cc=qiming.yang@intel.com \
    --cc=simei.su@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

DPDK patches and discussions

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://inbox.dpdk.org/dev/0 dev/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 dev dev/ https://inbox.dpdk.org/dev \
		dev@dpdk.org
	public-inbox-index dev

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


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