On 2023/10/16 16:47, David Marchand wrote: > If a bonding port gets released, a link status alarm callback still > referenced the ethdev port that may be reused later. > Cancel this callback when stopping the port. > > Bugzilla ID: 1301 > Fixes: a45b288ef21a ("bond: support link status polling") > Cc: sta...@dpdk.org > > Signed-off-by: David Marchand <david.march...@redhat.com> > --- > Note: this issue was made apparent now that we release the ethdev port > shared mem, see 36c46e738120 ("ethdev: cleanup shared data with ..."). > > --- > drivers/net/bonding/rte_eth_bond_pmd.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c > b/drivers/net/bonding/rte_eth_bond_pmd.c > index 122b1187fd..b8ee8be50f 100644 > --- a/drivers/net/bonding/rte_eth_bond_pmd.c > +++ b/drivers/net/bonding/rte_eth_bond_pmd.c > @@ -2159,6 +2159,10 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) > eth_dev->data->dev_link.link_status = RTE_ETH_LINK_DOWN; > eth_dev->data->dev_started = 0; > > + if (internals->link_status_polling_enabled) { > + > rte_eal_alarm_cancel(bond_ethdev_member_link_status_change_monitor, > + (void *)&rte_eth_devices[internals->port_id]); > + } > internals->link_status_polling_enabled = 0;
How about move this line in above {}. Anyway, both OK for me Acked-by: Chengwen Feng <fengcheng...@huawei.com> > for (i = 0; i < internals->member_count; i++) { > uint16_t member_id = internals->members[i].port_id; >