DPDK patches and discussions
 help / color / Atom feed
From: "Zhang, Xiao" <xiao.zhang@intel.com>
To: "Xing, Beilei" <beilei.xing@intel.com>, "dev@dpdk.org" <dev@dpdk.org>
Cc: "Zhang, Qi Z" <qi.z.zhang@intel.com>,
	"stable@dpdk.org" <stable@dpdk.org>
Subject: Re: [dpdk-dev] [v2] net/i40e: fix vf runtime queues rss config
Date: Fri, 9 Aug 2019 03:25:36 +0000
Message-ID: <AF0377F445CB2540BB46FF359C1C1BBE0118A9A0@SHSMSX105.ccr.corp.intel.com> (raw)
In-Reply-To: <94479800C636CB44BD422CB454846E013CE222D5@SHSMSX101.ccr.corp.intel.com>



> -----Original Message-----
> From: Xing, Beilei
> Sent: Friday, August 9, 2019 10:44 AM
> To: Zhang, Xiao <xiao.zhang@intel.com>; dev@dpdk.org
> Cc: Zhang, Qi Z <qi.z.zhang@intel.com>; stable@dpdk.org
> Subject: RE: [v2] net/i40e: fix vf runtime queues rss config
> 
> 
> > -----Original Message-----
> > From: Zhang, Xiao
> > Sent: Friday, August 9, 2019 12:41 AM
> > To: dev@dpdk.org
> > Cc: Xing, Beilei <beilei.xing@intel.com>; Zhang, Qi Z
> > <qi.z.zhang@intel.com>; Zhang, Xiao <xiao.zhang@intel.com>;
> > stable@dpdk.org
> > Subject: [v2] net/i40e: fix vf runtime queues rss config
> >
> > I40evf queue can not work properly with kernel pf driver. Eg. when
> > configure
> > 8 queues pair, only 4 queues can receive packets, and half packets
> > will be lost if using 2 queues pair.
> > This issue is caused by misconfiguration of look up table, use aq
> > command to setup the lut to make it work properly.
> >
> > Fixes: cea7a51c1750 ("i40evf: support RSS")
> > Cc: stable@dpdk.org
> >
> > Signed-off-by: Xiao Zhang <xiao.zhang@intel.com>
> > ---
> > v2 change for loop format to avoid build patch issue.
> > ---
> >  drivers/net/i40e/i40e_ethdev_vf.c | 36
> > +++++++++++++++++++++++++++++---
> > ----
> >  1 file changed, 29 insertions(+), 7 deletions(-)
> >
> > diff --git a/drivers/net/i40e/i40e_ethdev_vf.c
> > b/drivers/net/i40e/i40e_ethdev_vf.c
> > index 308fb98..2d3cf3e 100644
> > --- a/drivers/net/i40e/i40e_ethdev_vf.c
> > +++ b/drivers/net/i40e/i40e_ethdev_vf.c
> > @@ -2574,6 +2574,26 @@ i40evf_hw_rss_hash_set(struct i40e_vf *vf,
> > struct rte_eth_rss_conf *rss_conf)
> >  	if (ret)
> >  		return ret;
> >
> > +	if (vf->flags & I40E_FLAG_RSS_AQ_CAPABLE) {
> > +		uint8_t *lut;
> > +		uint32_t rss_lut_size = (I40E_VFQF_HLUT1_MAX_INDEX + 1) *
> > 4;
> > +		uint32_t i;
> > +		lut = rte_zmalloc("i40e_rss_lut", rss_lut_size, 0);
> > +		if (!lut) {
> > +			PMD_DRV_LOG(ERR, "No memory can be allocated");
> > +			return -ENOMEM;
> > +		}
> > +
> > +		for (i = 0; i < rss_lut_size; i++)
> > +			lut[i] = i % vf->num_queue_pairs;
> > +
> > +		ret = i40evf_set_rss_lut(&vf->vsi, lut,
> > +					 rss_lut_size);
> > +		rte_free(lut);
> > +		if (ret)
> > +			return ret;
> > +	}
> > +
> 
> Why not moving the LUT configuration to the following i40evf_config_rss
> function?

Yes, it's better to move the above code to i40evf_config_rss since the lagecy 
LUT configuration is also there. It will be more clear.

Thanks,
Xiao
> 
> >  	hena = i40e_config_hena(vf->adapter, rss_conf->rss_hf);
> >  	i40e_write_rx_ctl(hw, I40E_VFQF_HENA(0), (uint32_t)hena);
> >  	i40e_write_rx_ctl(hw, I40E_VFQF_HENA(1), (uint32_t)(hena >> 32));
> @@
> > -2607,13 +2627,15 @@ i40evf_config_rss(struct i40e_vf *vf)
> >  	}
> >
> >  	num = RTE_MIN(vf->dev_data->nb_rx_queues,
> > I40E_MAX_QP_NUM_PER_VF);
> > -	/* Fill out the look up table */
> > -	for (i = 0, j = 0; i < nb_q; i++, j++) {
> > -		if (j >= num)
> > -			j = 0;
> > -		lut = (lut << 8) | j;
> > -		if ((i & 3) == 3)
> > -			I40E_WRITE_REG(hw, I40E_VFQF_HLUT(i >> 2), lut);
> > +	if (!(vf->flags & I40E_FLAG_RSS_AQ_CAPABLE)) {
> > +		/* Fill out the look up table */
> > +		for (i = 0, j = 0; i < nb_q; i++, j++) {
> > +			if (j >= num)
> > +				j = 0;
> > +			lut = (lut << 8) | j;
> > +			if ((i & 3) == 3)
> > +				I40E_WRITE_REG(hw, I40E_VFQF_HLUT(i >>
> > 2), lut);
> > +		}
> >  	}
> >
> >  	rss_conf = vf->dev_data->dev_conf.rx_adv_conf.rss_conf;
> > --
> > 2.7.4


  reply index

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-08-07  0:09 [dpdk-dev] " Xiao Zhang
2019-08-08 16:41 ` [dpdk-dev] [v2] " Xiao Zhang
2019-08-09  2:44   ` Xing, Beilei
2019-08-09  3:25     ` Zhang, Xiao [this message]
2019-08-12 10:29   ` [dpdk-dev] [v3] " Xiao Zhang
2019-08-12 17:43     ` [dpdk-dev] [v4] " Xiao Zhang
2019-08-12 12:07       ` Aaron Conole
2019-08-13  1:43       ` Xing, Beilei
2019-08-13 10:40       ` [dpdk-dev] [v5] " Xiao Zhang
2019-08-13  2:21         ` Xing, Beilei
2019-08-13  6:28         ` Ye Xiaolong
2019-08-13  7:24           ` Zhang, Xiao
2019-08-13  7:37             ` Ye Xiaolong
2019-08-13  7:59               ` Zhang, Xiao
2019-08-13 22:17         ` [dpdk-dev] [v6] " Xiao Zhang
2019-08-13 14:55           ` Ye Xiaolong

Reply instructions:

You may reply publically 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=AF0377F445CB2540BB46FF359C1C1BBE0118A9A0@SHSMSX105.ccr.corp.intel.com \
    --to=xiao.zhang@intel.com \
    --cc=beilei.xing@intel.com \
    --cc=dev@dpdk.org \
    --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

DPDK patches and discussions

Archives are clonable:
	git clone --mirror http://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/ http://inbox.dpdk.org/dev \
		dev@dpdk.org
	public-inbox-index dev


Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.dev


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