RE: [PATCH] net/iavf: add diagnostic support in TX path

["Zhang, Qi Z" <qi.z.zhang@intel.com>]

> -----Original Message-----
> From: Mingjin Ye <mingjinx.ye@intel.com>
> Sent: Thursday, December 21, 2023 6:12 PM
> To: dev@dpdk.org
> Cc: Yang, Qiming <qiming.yang@intel.com>; Ye, MingjinX
> <mingjinx.ye@intel.com>; Wu, Jingjing <jingjing.wu@intel.com>; Xing, Beilei
> <beilei.xing@intel.com>
> Subject: [PATCH] net/iavf: add diagnostic support in TX path
> 
> The only way to enable diagnostics for TX paths is to modify the application
> source code. Making it difficult to diagnose faults.
> 
> In this patch, the devarg option "mbuf_check" is introduced and the
> parameters are configured to enable the corresponding diagnostics.
> 
> supported cases: mbuf, size, segment, offload, strict.
>  1. mbuf: check for corrupted mbuf.
>  2. size: check min/max packet length according to hw spec.
>  3. segment: check number of mbuf segments not exceed hw limitation.
>  4. offload: check any unsupported offload flag.
>  5. strict: check protocol headers.
> 
> parameter format: mbuf_check=[mbuf,<case1>,<case2>]
> eg: dpdk-testpmd -a 0000:81:01.0,mbuf_check=[mbuf,size] -- -i
> 
> Signed-off-by: Mingjin Ye <mingjinx.ye@intel.com>



> +
> +static uint16_t
> +iavf_xmit_pkts_chain(void *tx_queue, struct rte_mbuf **tx_pkts,
> +uint16_t nb_pkts) {
> +	struct iavf_tx_queue *txq = tx_queue;
> +	struct iavf_adapter *adapter = txq->vsi->adapter;
> +	struct iavf_tx_burst_elem *pos;
> +	struct iavf_tx_burst_elem *save_next;
> +	struct iavf_pkt_burst *item;
> +	uint16_t ret;
> +
> +	item = &ice_rxtx_pkt_burst[adapter->dev_data->port_id];
> +	RTE_TAILQ_FOREACH_SAFE(pos, &item->tx_burst_list, next,
> save_next) {
> +		ret = pos->tx_pkt_burst(tx_queue, tx_pkts, nb_pkts);
> +		if (nb_pkts != ret)

This is not correct,

> +			break;
> +	}
> +
> +	return ret;
> +}
> +
> 
>  /* choose rx function*/
> @@ -3973,9 +4226,6 @@ iavf_set_rx_function(struct rte_eth_dev *dev)
> void  iavf_set_tx_function(struct rte_eth_dev *dev)  {
> -	struct iavf_adapter *adapter =
> -		IAVF_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
> -	int no_poll_on_link_down = adapter-
> >devargs.no_poll_on_link_down;
>  #ifdef RTE_ARCH_X86
>  	struct iavf_tx_queue *txq;
>  	int i;
> @@ -4062,10 +4312,7 @@ iavf_set_tx_function(struct rte_eth_dev *dev)
> #endif
>  		}
> 
> -		if (no_poll_on_link_down) {
> -			adapter->tx_pkt_burst = dev->tx_pkt_burst;
> -			dev->tx_pkt_burst = iavf_xmit_pkts_no_poll;
> -		}
> +		iavf_set_tx_interceptors(dev);

If no_poll or diagnose mode are not enabled, why we still need to interceptor?, 

Btw, its too heavy to use a chain here and I even didn't see an Rx chain implementation.

Why not simplify it as below:

adapter->tx_pkt_burst = dev->tx_pkt_burst;
if (no_poll_on_link_down)
	dev->tx_pkt_burst = iavf_xmit_pkts_no_poll;
else if (diagnose mode)
	dev->tx_pkt_burst = iavf_xmit_pkt_mdd_check.

then iavf_xmit_pkt_mdd_check,  call iavf_xmit_pkts_no_poll or adapter->tx_pkt_burst