Re: [dpdk-dev] [PATCH] net/tap: release port upon close

[Thomas Monjalon <thomas@monjalon.net>]

As you may notice, I have included a slightly modified version
of this patch in my series in order to cover the full picture:
	https://patches.dpdk.org/patch/77566/

Feel free to continue improving your patch in this thread or the other,
as you prefer, as long as the secondary process issue is adressed.

Thanks


13/09/2020 11:16, Thomas Monjalon:
> Hi,
> 
> 28/08/2020 14:37, wangyunjian:
> > @@ -1078,6 +1085,23 @@ tap_dev_close(struct rte_eth_dev *dev)
> > +
> > +	/* mac_addrs must not be freed alone because part of dev_private */
> > +	dev->data->mac_addrs = NULL;
> > +
> > +	internals = dev->data->dev_private;
> > +	TAP_LOG(DEBUG, "Closing %s Ethernet device on numa %u",
> > +		tuntap_types[internals->type], rte_socket_id());
> > +
> > +	if (internals->ioctl_sock != -1) {
> > +		close(internals->ioctl_sock);
> > +		internals->ioctl_sock = -1;
> > +	}
> > +	rte_free(dev->process_private);
> > +	dev->process_private = NULL;
> > +	if (tap_devices_count == 1)
> > +		rte_mp_action_unregister(TAP_MP_KEY);
> > +	tap_devices_count--;
> >  }
> [...]
> > @@ -2446,12 +2466,11 @@ static int
> >  rte_pmd_tap_remove(struct rte_vdev_device *dev)
> >  {
> >  
> >         struct rte_eth_dev *eth_dev = NULL;
> > 
> > -       struct pmd_internals *internals;
> > 
> >         /* find the ethdev entry */
> >         eth_dev = rte_eth_dev_allocated(rte_vdev_device_name(dev));
> >         if (!eth_dev)
> > 
> > -               return -ENODEV;
> > +               return 0;
> > 
> >         /* mac_addrs must not be freed alone because part of dev_private */
> >         eth_dev->data->mac_addrs = NULL;
> 
> The line above can be removed because mac_addrs is not freed
> in secondary process, and it is redundant with tap_dev_close().
> 
> >  
> > 	if (rte_eal_process_type() != RTE_PROC_PRIMARY)
> >  		return rte_eth_dev_release_port(eth_dev);
> 
> There is an inconsistency in secondary process
> if tap_dev_close() is not called from .remove (unplug)
> but can be called from .dev_close (rte_eth_dev_close).
> 
> I think the process type check must be done inside tap_dev_close(),
> so the process private resources can be freed.
> 
> > 	
> >  	tap_dev_close(eth_dev);
> >  
> 
> This blank line can be removed in my opinion.
> 
> > -	internals = eth_dev->data->dev_private;
> > -	TAP_LOG(DEBUG, "Closing %s Ethernet device on numa %u",
> > -		tuntap_types[internals->type], rte_socket_id());
> > -
> > -	close(internals->ioctl_sock);
> > -	rte_free(eth_dev->process_private);
> > -	if (tap_devices_count == 1)
> > -		rte_mp_action_unregister(TAP_MP_KEY);
> > -	tap_devices_count--;
> >  	rte_eth_dev_release_port(eth_dev);
> >  
> >  	return 0;