Re: [dpdk-dev] [PATCH v3 5/6] ixgbe: Config VF RSS

[Vlad Zolotarov <vladz@cloudius-systems.com>]

On 12/25/14 04:43, Ouyang, Changchun wrote:
> Hi,
> Sorry miss some comments, so continue my response below,
>
>> -----Original Message-----
>> From: Vlad Zolotarov [mailto:vladz@cloudius-systems.com]
>> Sent: Wednesday, December 24, 2014 6:40 PM
>> To: Ouyang, Changchun; dev@dpdk.org
>> Subject: Re: [dpdk-dev] [PATCH v3 5/6] ixgbe: Config VF RSS
>>
>>
>> On 12/24/14 07:23, Ouyang Changchun wrote:
>>> It needs config RSS and IXGBE_MRQC and IXGBE_VFPSRTYPE to enable VF
>> RSS.
>>> The psrtype will determine how many queues the received packets will
>>> distribute to, and the value of psrtype should depends on both facet:
>>> max VF rxq number which has been negotiated with PF, and the number of
>> rxq specified in config on guest.
>>> Signed-off-by: Changchun Ouyang <changchun.ouyang@intel.com>
>>> ---
>>>    lib/librte_pmd_ixgbe/ixgbe_pf.c   | 15 +++++++
>>>    lib/librte_pmd_ixgbe/ixgbe_rxtx.c | 92
>> ++++++++++++++++++++++++++++++++++-----
>>>    2 files changed, 97 insertions(+), 10 deletions(-)
>>>
>>> diff --git a/lib/librte_pmd_ixgbe/ixgbe_pf.c
>>> b/lib/librte_pmd_ixgbe/ixgbe_pf.c index cbb0145..9c9dad8 100644
>>> --- a/lib/librte_pmd_ixgbe/ixgbe_pf.c
>>> +++ b/lib/librte_pmd_ixgbe/ixgbe_pf.c
>>> @@ -187,6 +187,21 @@ int ixgbe_pf_host_configure(struct rte_eth_dev
>> *eth_dev)
>>>    	IXGBE_WRITE_REG(hw, IXGBE_MPSAR_LO(hw-
>>> mac.num_rar_entries), 0);
>>>    	IXGBE_WRITE_REG(hw, IXGBE_MPSAR_HI(hw-
>>> mac.num_rar_entries), 0);
>>>
>>> +	/*
>>> +	 * VF RSS can support at most 4 queues for each VF, even if
>>> +	 * 8 queues are available for each VF, it need refine to 4
>>> +	 * queues here due to this limitation, otherwise no queue
>>> +	 * will receive any packet even RSS is enabled.
>> According to Table 7-3 in the 82599 spec RSS is not available when port is
>> configured to have 8 queues per pool. This means that if u see this
>> configuration u may immediately disable RSS flow in your code.
>>
>>> +	 */
>>> +	if (eth_dev->data->dev_conf.rxmode.mq_mode ==
>> ETH_MQ_RX_VMDQ_RSS) {
>>> +		if (RTE_ETH_DEV_SRIOV(eth_dev).nb_q_per_pool == 8) {
>>> +			RTE_ETH_DEV_SRIOV(eth_dev).active =
>> ETH_32_POOLS;
>>> +			RTE_ETH_DEV_SRIOV(eth_dev).nb_q_per_pool = 4;
>>> +			RTE_ETH_DEV_SRIOV(eth_dev).def_pool_q_idx =
>>> +				dev_num_vf(eth_dev) * 4;
>> According to 82599 spec u can't do that since RSS is not allowed when port is
>> configured to have 8 function per-VF. Have u verified that this works? If yes,
>> then spec should be updated.
>>
>>> +		}
>>> +	}
>>> +
>>>    	/* set VMDq map to default PF pool */
>>>    	hw->mac.ops.set_vmdq(hw, 0,
>>> RTE_ETH_DEV_SRIOV(eth_dev).def_vmdq_idx);
>>>
>>> diff --git a/lib/librte_pmd_ixgbe/ixgbe_rxtx.c
>>> b/lib/librte_pmd_ixgbe/ixgbe_rxtx.c
>>> index f69abda..a7c17a4 100644
>>> --- a/lib/librte_pmd_ixgbe/ixgbe_rxtx.c
>>> +++ b/lib/librte_pmd_ixgbe/ixgbe_rxtx.c
>>> @@ -3327,6 +3327,39 @@ ixgbe_alloc_rx_queue_mbufs(struct
>> igb_rx_queue *rxq)
>>>    }
>>>
>>>    static int
>>> +ixgbe_config_vf_rss(struct rte_eth_dev *dev) {
>>> +	struct ixgbe_hw *hw;
>>> +	uint32_t mrqc;
>>> +
>>> +	ixgbe_rss_configure(dev);
>>> +
>>> +	hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
>>> +
>>> +	/* MRQC: enable VF RSS */
>>> +	mrqc = IXGBE_READ_REG(hw, IXGBE_MRQC);
>>> +	mrqc &= ~IXGBE_MRQC_MRQE_MASK;
>>> +	switch (RTE_ETH_DEV_SRIOV(dev).active) {
>>> +	case ETH_64_POOLS:
>>> +		mrqc |= IXGBE_MRQC_VMDQRSS64EN;
>>> +		break;
>>> +
>>> +	case ETH_32_POOLS:
>>> +	case ETH_16_POOLS:
>>> +		mrqc |= IXGBE_MRQC_VMDQRSS32EN;
>> Again, this contradicts with the spec.
> Yes, the spec say the hw can't support vf rss at all, but experiment find that could be done.

I have just realized something - why did u have to experiment at all? U 
work at Intel, don't u? Can't u just ask a HW engineer that have 
designed this NIC? What do u mean by an "experiment" here?
 From my experience u can't just write some random values in the 
registers and conclude that if it worked for like 5 minutes it will 
continue to work for the next minute... There is always a clear 
procedure of how HW should be initialized and used and that's the only 
way it may be used since this was the way the HW has been tested. U 
can't assume anything in regards to reliability if u don't follow specs 
and programmer manuals of HW provider.

Could u clarify, please?

> We can focus on discussing the implementation firstly.
>   
>>> +		break;
>>> +
>>> +	default:
>>> +		PMD_INIT_LOG(ERR, "Invalid pool number in IOV mode");
>>> +		return -EINVAL;
>>> +	}
>>> +
>>> +	IXGBE_WRITE_REG(hw, IXGBE_MRQC, mrqc);
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static int
>>>    ixgbe_dev_mq_rx_configure(struct rte_eth_dev *dev)
>>>    {
>>>    	struct ixgbe_hw *hw =
>>> @@ -3358,24 +3391,38 @@ ixgbe_dev_mq_rx_configure(struct
>> rte_eth_dev *dev)
>>>    			default: ixgbe_rss_disable(dev);
>>>    		}
>>>    	} else {
>>> -		switch (RTE_ETH_DEV_SRIOV(dev).active) {
>>>    		/*
>>>    		 * SRIOV active scheme
>>>    		 * FIXME if support DCB/RSS together with VMDq & SRIOV
>>>    		 */
>>> -		case ETH_64_POOLS:
>>> -			IXGBE_WRITE_REG(hw, IXGBE_MRQC,
>> IXGBE_MRQC_VMDQEN);
>>> +		switch (dev->data->dev_conf.rxmode.mq_mode) {
>>> +		case ETH_MQ_RX_RSS:
>>> +		case ETH_MQ_RX_VMDQ_RSS:
>>> +			ixgbe_config_vf_rss(dev);
>>>    			break;
>>>
>>> -		case ETH_32_POOLS:
>>> -			IXGBE_WRITE_REG(hw, IXGBE_MRQC,
>> IXGBE_MRQC_VMDQRT4TCEN);
>>> -			break;
>>> +		default:
>>> +			switch (RTE_ETH_DEV_SRIOV(dev).active) {
>> Sorry for nitpicking but have u considered taking this encapsulated "switch-
>> case" block into a separate function? This could make the code look a lot
>> nicer. ;)
>>
>>> +			case ETH_64_POOLS:
>>> +				IXGBE_WRITE_REG(hw, IXGBE_MRQC,
>>> +					IXGBE_MRQC_VMDQEN);
>>> +				break;
>>>
>>> -		case ETH_16_POOLS:
>>> -			IXGBE_WRITE_REG(hw, IXGBE_MRQC,
>> IXGBE_MRQC_VMDQRT8TCEN);
>>> +			case ETH_32_POOLS:
>>> +				IXGBE_WRITE_REG(hw, IXGBE_MRQC,
>>> +					IXGBE_MRQC_VMDQRT4TCEN);
>>> +				break;
>>> +
>>> +			case ETH_16_POOLS:
>>> +				IXGBE_WRITE_REG(hw, IXGBE_MRQC,
>>> +					IXGBE_MRQC_VMDQRT8TCEN);
>>> +				break;
>>> +			default:
>>> +				PMD_INIT_LOG(ERR,
>>> +					"invalid pool number in IOV mode");
>>> +				break;
>>> +			}
>>>    			break;
>>> -		default:
>>> -			PMD_INIT_LOG(ERR, "invalid pool number in IOV
>> mode");
>>>    		}
>>>    	}
>>>
>>> @@ -3989,10 +4036,32 @@ ixgbevf_dev_rx_init(struct rte_eth_dev *dev)
>>>    	uint16_t buf_size;
>>>    	uint16_t i;
>>>    	int ret;
>>> +	uint16_t valid_rxq_num;
>>>
>>>    	PMD_INIT_FUNC_TRACE();
>>>    	hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
>>>
>>> +	valid_rxq_num = RTE_MIN(dev->data->nb_rx_queues,
>>> +hw->mac.max_rx_queues);
>>> +
>>> +	/*
>>> +	 * VMDq RSS can't support 3 queues, so config it into 4 queues,
>>> +	 * and give user a hint that some packets may loss if it doesn't
>>> +	 * poll the queue where those packets are distributed to.
>>> +	 */
>>> +	if (valid_rxq_num == 3)
>>> +		valid_rxq_num = 4;
>> Why to configure more queues that requested and not less (2)? Why to
>> configure anything at all and not return an error?
> Sorry, I don't agree this is "anything" you say, because I don't use 5,6,7, 8, ..., 16, 2014, 2015,... etc.
> By considering 2 or 4,
> I prefer 4, the reason is if user need more than 3 queues per vf to do something,
> And pf has also the capability to setup 4 queues per vf, confining to 2 queues is also not good thing,
> So here try to enable 4 queues, and give user hints here.
> Btw, change it into 2 is another way, depends on other guys' more insight here.
>
>>> +
>>> +	if (dev->data->nb_rx_queues > valid_rxq_num) {
>>> +		PMD_INIT_LOG(ERR, "The number of Rx queue invalid, "
>>> +			"it should be equal to or less than %d",
>>> +			valid_rxq_num);
>>> +		return -1;
>>> +	} else if (dev->data->nb_rx_queues < valid_rxq_num)
>>> +		PMD_INIT_LOG(ERR, "The number of Rx queue is less "
>>> +			"than the number of available Rx queues:%d, "
>>> +			"packets in Rx queues(q_id >= %d) may loss.",
>>> +			valid_rxq_num, dev->data->nb_rx_queues);
>> Who ever looks in the "INIT_LOG" if everything "work well" and u make it
>> look so by allowing this call to succeed. And then some packets will just
>> silently not arrive?! And what the used should somehow guess to do?
>> - Look in the "INIT_LOG"?! This is a nightmare!
>>
>>> +
>>>    	/*
>>>    	 * When the VF driver issues a IXGBE_VF_RESET request, the PF
>> driver
>>>    	 * disables the VF receipt of packets if the PF MTU is > 1500.
>>> @@ -4094,6 +4163,9 @@ ixgbevf_dev_rx_init(struct rte_eth_dev *dev)
>>>    			IXGBE_PSRTYPE_IPV6HDR;
>>>    #endif
>>>
>>> +	/* Set RQPL for VF RSS according to max Rx queue */
>>> +	psrtype |= (valid_rxq_num >> 1) <<
>>> +		IXGBE_PSRTYPE_RQPL_SHIFT;
>>>    	IXGBE_WRITE_REG(hw, IXGBE_VFPSRTYPE, psrtype);
>>>
>>>    	if (dev->data->dev_conf.rxmode.enable_scatter) {