Re: [dpdk-dev] [PATCH v2] net/i40e: fix Tx fn selection when using new ethdev offloads

[Bruce Richardson <bruce.richardson@intel.com>]

On Tue, May 01, 2018 at 06:52:18PM +0100, Ananyev, Konstantin wrote:
> 
> 
> > -----Original Message-----
> > From: Richardson, Bruce
> > Sent: Tuesday, May 1, 2018 3:14 PM
> > To: Xing, Beilei <beilei.xing@intel.com>; Zhang, Qi Z <qi.z.zhang@intel.com>
> > Cc: dev@dpdk.org; Yigit, Ferruh <ferruh.yigit@intel.com>; Ananyev, Konstantin <konstantin.ananyev@intel.com>; Richardson,
> > Bruce <bruce.richardson@intel.com>
> > Subject: [PATCH v2] net/i40e: fix Tx fn selection when using new ethdev offloads
> > 
> > The Tx function selection code in the driver only used the older txq
> > flags values to check whether the scalar or vector functions should be
> > used. This caused performance regressions with testpmd io-fwd as the
> > scalar path rather than the vector one was being used in the default
> > case. Fix this by changing the code to take account of new offloads and
> > deleting the defines used for the old ones.
> > 
> > Fixes: 7497d3e2f777 ("net/i40e: convert to new Tx offloads API")
> > 
> > Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
> > ---
> >  drivers/net/i40e/i40e_rxtx.c | 39 ++++++++++++++++++---------------------
> >  1 file changed, 18 insertions(+), 21 deletions(-)
> > 
> > diff --git a/drivers/net/i40e/i40e_rxtx.c b/drivers/net/i40e/i40e_rxtx.c
> > index ec1ce54ca..006f5b846 100644
> > --- a/drivers/net/i40e/i40e_rxtx.c
> > +++ b/drivers/net/i40e/i40e_rxtx.c
> > @@ -40,9 +40,6 @@
> >  /* Base address of the HW descriptor ring should be 128B aligned. */
> >  #define I40E_RING_BASE_ALIGN	128
> > 
> > -#define I40E_SIMPLE_FLAGS ((uint32_t)ETH_TXQ_FLAGS_NOMULTSEGS | \
> > -					ETH_TXQ_FLAGS_NOOFFLOADS)
> > -
> >  #define I40E_TXD_CMD (I40E_TX_DESC_CMD_EOP | I40E_TX_DESC_CMD_RS)
> > 
> >  #ifdef RTE_LIBRTE_IEEE1588
> > @@ -2108,11 +2105,9 @@ i40e_dev_tx_queue_setup_runtime(struct rte_eth_dev *dev,
> >  				 dev->data->nb_tx_queues)) {
> >  		/**
> >  		 * If it is the first queue to setup,
> > -		 * set all flags to default and call
> > +		 * set all flags and call
> >  		 * i40e_set_tx_function.
> >  		 */
> > -		ad->tx_simple_allowed = true;
> > -		ad->tx_vec_allowed = true;
> >  		i40e_set_tx_function_flag(dev, txq);
> >  		i40e_set_tx_function(dev);
> >  		return 0;
> > @@ -2128,9 +2123,8 @@ i40e_dev_tx_queue_setup_runtime(struct rte_eth_dev *dev,
> >  	}
> >  	/* check simple tx conflict */
> >  	if (ad->tx_simple_allowed) {
> > -		if (((txq->txq_flags & I40E_SIMPLE_FLAGS) !=
> > -		     I40E_SIMPLE_FLAGS) ||
> > -		    txq->tx_rs_thresh < RTE_PMD_I40E_TX_MAX_BURST) {
> > +		if (txq->offloads != 0 ||
> > +				txq->tx_rs_thresh < RTE_PMD_I40E_TX_MAX_BURST) {
> >  			PMD_DRV_LOG(ERR, "No-simple tx is required.");
> >  			return -EINVAL;
> >  		}
> > @@ -3080,18 +3074,21 @@ i40e_set_tx_function_flag(struct rte_eth_dev *dev, struct i40e_tx_queue *txq)
> >  		I40E_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
> > 
> >  	/* Use a simple Tx queue (no offloads, no multi segs) if possible */
> > -	if (((txq->txq_flags & I40E_SIMPLE_FLAGS) == I40E_SIMPLE_FLAGS)
> > -			&& (txq->tx_rs_thresh >= RTE_PMD_I40E_TX_MAX_BURST)) {
> > -		if (txq->tx_rs_thresh <= RTE_I40E_TX_MAX_FREE_BUF_SZ) {
> > -			PMD_INIT_LOG(DEBUG, "Vector tx"
> > -				     " can be enabled on this txq.");
> > -
> > -		} else {
> > -			ad->tx_vec_allowed = false;
> > -		}
> > -	} else {
> > -		ad->tx_simple_allowed = false;
> > -	}
> > +	ad->tx_simple_allowed = (txq->offloads == 0 &&
> > +			txq->tx_rs_thresh >= RTE_PMD_I40E_TX_MAX_BURST);
> 
> Actually after another thought - who setup txq->offloads?
> I did a quick scan, through i40e code and seems no one does.
> So now it seems not possible to enable TX offloads at all.
> Konstantin
> 
> BTW, seems like rxq->offloads are not properly initialised too.
> 
The offloads value should come from the app, no?