[dpdk-dev] [PATCH] net/mlx5: allow multi probing

[Ophir Munk]

v2 sent

> -----Original Message-----
> From: Yongseok Koh
> Sent: Friday, October 05, 2018 4:42 AM
> To: Ophir Munk <ophirmu at mellanox.com>
> Cc: dev at dpdk.org; Asaf Penso <asafp at mellanox.com>; Shahaf Shuler
> <shahafs at mellanox.com>; Thomas Monjalon <thomas at monjalon.net>;
> Olga Shern <olgas at mellanox.com>
> Subject: Re: [PATCH] net/mlx5: allow multi probing
> 
> On Wed, Oct 03, 2018 at 01:01:23AM -0700, Ophir Munk wrote:
> > Implement probing of a device multiple times, see [1].
> > Consecutive probing requests with a devargs string may contain
> > repetitive master and representors devices for which eth device should
> > be created. If an eth device already exists - silently ignore it.
> >
> > [1]
> > Serie: ("eal: allow hotplug to skip an already probed device")
> 
> A typo? Series?
> 
> For the title, either 'multi-probing' or 'multiple probing'.
> And I suggest
> 
> 	net/mlx5: allow multiple probing for representor
> 
> And I have only nitpickings below ;-)
> 
> > https://patches.dpdk.org/project/dpdk/list/?series=1580
> >
> > Signed-off-by: Ophir Munk <ophirmu at mellanox.com>
> > ---
> >  drivers/net/mlx5/mlx5.c | 22 +++++++++++++---------
> >  1 file changed, 13 insertions(+), 9 deletions(-)
> >
> > diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c index
> > b2b0aaa..16a8b9d 100644
> > --- a/drivers/net/mlx5/mlx5.c
> > +++ b/drivers/net/mlx5/mlx5.c
> > @@ -736,6 +736,7 @@
> >  	struct ether_addr mac;
> >  	char name[RTE_ETH_NAME_MAX_LEN];
> >  	int own_domain_id = 0;
> > +	uint16_t port_id;
> >  	unsigned int i;
> >
> >  	/* Determine if this port representor is supposed to be spawned. */
> > @@ -758,6 +759,17 @@
> >  			return NULL;
> >  		}
> >  	}
> > +	/* Build device name */
> 
> Better to put a period (.), if it is a sentence.
> 
> > +	if (!switch_info->representor)
> > +		rte_strlcpy(name, dpdk_dev->name, sizeof(name));
> > +	else
> > +		snprintf(name, sizeof(name), "%s_representor_%u",
> > +			 dpdk_dev->name, switch_info->port_name);
> > +	/* if dev (master or representor) is already spawned - return NULL
> > +*/
> 
> How about,
> 	/* Check if the device is already spawned. */
> 
> "return NULL" is so obvious from the code.
> 
> > +	if (rte_eth_dev_get_port_by_name(name, &port_id) == 0) {
> > +		rte_errno = EBUSY;
> 
> Semantically, shouldn't it be EEXIST and add the condition in probe()?
> 
> > +		return NULL;
> > +	}
> >  	/* Prepare shared data between primary and secondary process. */
> >  	mlx5_prepare_shared_data();
> >  	errno = 0;
> > @@ -864,11 +876,6 @@
> >  		DEBUG("ibv_query_device_ex() failed");
> >  		goto error;
> >  	}
> > -	if (!switch_info->representor)
> > -		rte_strlcpy(name, dpdk_dev->name, sizeof(name));
> > -	else
> > -		snprintf(name, sizeof(name), "%s_representor_%u",
> > -			 dpdk_dev->name, switch_info->port_name);
> >  	DRV_LOG(DEBUG, "naming Ethernet device \"%s\"", name);
> >  	if (rte_eal_process_type() == RTE_PROC_SECONDARY) {
> >  		eth_dev = rte_eth_dev_attach_secondary(name);
> > @@ -1298,9 +1305,6 @@ struct mlx5_dev_spawn_data {
> >  	assert(pci_drv == &mlx5_driver);
> >  	errno = 0;
> >
> > -	if (rte_dev_is_probed(&pci_dev->device))
> > -		return -EEXIST;
> > -
> >  	ibv_list = mlx5_glue->get_device_list(&ret);
> >  	if (!ibv_list) {
> >  		rte_errno = errno ? errno : ENOSYS; @@ -1412,7 +1416,7
> @@ struct
> > mlx5_dev_spawn_data {
> >  		if (!list[i].eth_dev) {
> >  			if (rte_errno != EBUSY)
> 
> I meant this to be,
> 			if (rte_errno != EBUSY && rte_errno != EEXIST)
> 
> 
> Thanks,
> Yongseok
> 
> >  				break;
> > -			/* Device is disabled, ignore it. */
> > +			/* Device is disabled or already spawned. Ignore it.
> */
> >  			continue;
> >  		}
> >  		restore = list[i].eth_dev->data->dev_flags;
> > --
> > 1.8.3.1
> >