25/09/2020 10:52, Ferruh Yigit:
> On 9/24/2020 10:58 PM, Thomas Monjalon wrote:
> > 23/09/2020 22:47, Thomas Monjalon:
> >> 23/09/2020 18:44, Ferruh Yigit:
> >>> On 9/13/2020 11:07 PM, Thomas Monjalon wrote:
> >>>> The flag RTE_ETH_DEV_CLOSE_REMOVE is set so all port resources
> >>>> can be freed by rte_eth_dev_close().
> >>>>
> >>>> Signed-off-by: Thomas Monjalon <tho...@monjalon.net>
> >>>> ---
> >>>> --- a/drivers/net/null/rte_eth_null.c
> >>>> +++ b/drivers/net/null/rte_eth_null.c
> >>>> +static int
> >>>> +eth_dev_close(struct rte_eth_dev *dev)
> >>>> +{
> >>>> + PMD_LOG(INFO, "Closing null ethdev on NUMA socket %u",
> >>>> + rte_socket_id());
> >>>> +
> >>>> + /* mac_addrs must not be freed alone because part of
> >>>> dev_private */
> >>>> + dev->data->mac_addrs = NULL;
> >>>> +
> >>>> + return 0;
> >>>> +}
> >>>
> >>> should check 'RTE_PROC_PRIMARY' in 'eth_dev_close()'?
> >>
> >> Yes, looks to be a miss for this new function.
> >
> > Sorry no, this function is not freeing any shared data,
> > so no restriction on secondary process.
>
> "dev->data->mac_addrs = NULL", won't this change the shared data?
Absolutely, you're right.
Will fix in v2.
