[PATCH 3/4] eventdev: have ethernet Tx adapter appropriately report idle

[Naga Harish K, S V]

> -----Original Message-----
> From: Jayatheerthan, Jay <jay.jayatheerthan at intel.com>
> Sent: Tuesday, October 11, 2022 12:41 PM
> To: mattias.ronnblom <mattias.ronnblom at ericsson.com>; Carrillo, Erik G
> <erik.g.carrillo at intel.com>; Gujjar, Abhinandan S
> <abhinandan.gujjar at intel.com>; Jerin Jacob <jerinj at marvell.com>; Naga
> Harish K, S V <s.v.naga.harish.k at intel.com>
> Cc: dev at dpdk.org; Van Haaren, Harry <harry.van.haaren at intel.com>;
> hofors at lysator.liu.se; mattias.ronnblom <mattias.ronnblom at ericsson.com>
> Subject: RE: [PATCH 3/4] eventdev: have ethernet Tx adapter appropriately
> report idle
> 
> @Harish, could you review the patch ?
> 
> -Jay
> 
> 
> 
> > -----Original Message-----
> > From: Mattias Rönnblom <mattias.ronnblom at ericsson.com>
> > Sent: Monday, October 10, 2022 8:24 PM
> > To: Jayatheerthan, Jay <jay.jayatheerthan at intel.com>; Carrillo, Erik G
> > <erik.g.carrillo at intel.com>; Gujjar, Abhinandan S
> > <abhinandan.gujjar at intel.com>; Jerin Jacob <jerinj at marvell.com>
> > Cc: dev at dpdk.org; Van Haaren, Harry <harry.van.haaren at intel.com>;
> > hofors at lysator.liu.se; mattias.ronnblom
> > <mattias.ronnblom at ericsson.com>
> > Subject: [PATCH 3/4] eventdev: have ethernet Tx adapter appropriately
> > report idle
> >
> > Update the Event Ethernet Tx Adapter's service function to report as
> > idle (i.e., return -EAGAIN) in case no events were dequeued from the
> > event device and no Ethernet frames were sent out on the wire.
> >
> > Signed-off-by: Mattias Rönnblom <mattias.ronnblom at ericsson.com>
> > ---
> >  lib/eventdev/rte_event_eth_tx_adapter.c | 13 +++++++++----
> >  1 file changed, 9 insertions(+), 4 deletions(-)
> >
> > diff --git a/lib/eventdev/rte_event_eth_tx_adapter.c
> > b/lib/eventdev/rte_event_eth_tx_adapter.c
> > index 7f7d86f683..c2a848103b 100644
> > --- a/lib/eventdev/rte_event_eth_tx_adapter.c
> > +++ b/lib/eventdev/rte_event_eth_tx_adapter.c
> > @@ -639,6 +639,7 @@ txa_service_func(void *args)
> >  	struct txa_service_data *txa = args;
> >  	uint8_t dev_id;
> >  	uint8_t port;
> > +	int ret = -EAGAIN;
> >  	uint16_t n;
> >  	uint32_t nb_tx, max_nb_tx;
> >  	struct rte_event ev[TXA_BATCH_SIZE]; @@ -648,10 +649,10 @@
> > txa_service_func(void *args)
> >  	port = txa->port_id;
> >
> >  	if (txa->nb_queues == 0)
> > -		return 0;
> > +		return ret;
> >
> >  	if (!rte_spinlock_trylock(&txa->tx_lock))
> > -		return 0;
> > +		return ret;
> >
> >  	for (nb_tx = 0; nb_tx < max_nb_tx; nb_tx += n) {
> >
> > @@ -659,6 +660,7 @@ txa_service_func(void *args)
> >  		if (!n)
> >  			break;
> >  		txa_service_tx(txa, ev, n);
> > +		ret = 0;
> >  	}
> >
> >  	if ((txa->loop_cnt++ & (TXA_FLUSH_THRESHOLD - 1)) == 0) { @@ -
> 692,10
> > +694,13 @@ txa_service_func(void *args)
> >  			}
> >  		}
> >
> > -		txa->stats.tx_packets += nb_tx;
> > +		if (likely(nb_tx > 0)) {

How are the performance numbers before and after this patch?
Trying to understand the performance impact, as new condition is added to the service function Datapath.

> > +			txa->stats.tx_packets += nb_tx;
> > +			ret = 0;
> > +		}
> >  	}
> >  	rte_spinlock_unlock(&txa->tx_lock);
> > -	return 0;
> > +	return ret;
> >  }
> >
> >  static int
> > --
> > 2.34.1