RE: [EXT] Re: [dpdk-stable] [PATCH v2] test: avoid hang if queues are full and Tx fails

[Rakesh Kudurumalla <rkudurumalla@marvell.com>]

> -----Original Message-----
> From: Thomas Monjalon <thomas@monjalon.net>
> Sent: Monday, November 29, 2021 2:44 PM
> To: Rakesh Kudurumalla <rkudurumalla@marvell.com>; Jerin Jacob
> Kollanukkaran <jerinj@marvell.com>
> Cc: stable@dpdk.org; dev@dpdk.org; david.marchand@redhat.com;
> ferruh.yigit@intel.com; andrew.rybchenko@oktetlabs.ru;
> ajit.khaparde@broadcom.com
> Subject: Re: [EXT] Re: [dpdk-stable] [PATCH v2] test: avoid hang if queues are
> full and Tx fails
> 
> 29/11/2021 09:52, Rakesh Kudurumalla:
> > From: Thomas Monjalon <thomas@monjalon.net>
> > > 22/11/2021 08:59, Rakesh Kudurumalla:
> > > > From: Thomas Monjalon <thomas@monjalon.net>
> > > > > 20/07/2021 18:50, Rakesh Kudurumalla:
> > > > > > Current pmd_perf_autotest() in continuous mode tries to
> > > > > > enqueue MAX_TRAFFIC_BURST completely before starting the test.
> > > > > > Some drivers cannot accept complete MAX_TRAFFIC_BURST even
> > > > > > though
> > > rx+tx
> > > > > > desc
> > > > > count
> > > > > > can fit it.
> > > > >
> > > > > Which driver is failing to do so?
> > > > > Why it cannot enqueue 32 packets?
> > > >
> > > > Octeontx2 driver is failing to enqueue because hardware buffers
> > > > are full
> > > before test.
> 
> Aren't you stopping the support of octeontx2?
> Why do you care now?
>  yes we are not supporting octeontx2,but this  issue is observed in cnxk driver ,current patch fixes the same
> > >
> > > Why hardware buffers are full?
> > Hardware buffers are full because number of number of descriptors in
> > continuous mode Is less than MAX_TRAFFIC_BURST, so if enque fails ,
> > there is no way hardware can drop the Packets . pmd_per_autotest
> > application evaluates performance after enqueueing packets Initially.
> > >
> > > > pmd_perf_autotest() in continuous mode tries to enqueue
> > > > MAX_TRAFFIC_BURST (2048) before starting the test.
> > > >
> > > > > > This patch changes behaviour to stop enqueuing after few retries.
> > > > >
> > > > > If there is a real limitation, there will be issues in more
> > > > > places than this test program.
> > > > > I feel it should be addressed either in the driver or at ethdev level.
> > > > >
> > > > > [...]
> > > > > > @@ -480,10 +483,19 @@ main_loop(__rte_unused void *args)
> > > > > >  			nb_tx = RTE_MIN(MAX_PKT_BURST, num);
> > > > > >  			nb_tx = rte_eth_tx_burst(portid, 0,
> > > > > >  						&tx_burst[idx],
> nb_tx);
> > > > > > +			if (nb_tx == 0)
> > > > > > +				retry_cnt++;
> > > > > >  			num -= nb_tx;
> > > > > >  			idx += nb_tx;
> > > > > > +			if (retry_cnt == MAX_RETRY_COUNT) {
> > > > > > +				retry_cnt = 0;
> > > > > > +				break;
> > > > > > +			}
> 
>