[PATCH 19.11] net/bonding: fix flow flush order on close

[Christian Ehrhardt]

On Mon, Nov 14, 2022 at 12:13 PM Ivan Malov <ivan.malov at oktetlabs.ru> wrote:
>
> [ upstream commit df810d1b6e31a3e25085a6abae3be119af3034c1 ]

Thanks, applied to the WIP branch - expect it to be part of 19.11.14
unless some builds stumble over it.

> The current code first removes all back-end devices of
> the bonded device and then invokes flush operation to
> remove flows in such back-end devices, which makes no
> sense. Fix that by re-ordering the steps accordingly.
>
> Fixes: 49dad9028e2a ("net/bonding: support flow API")
>
> Signed-off-by: Ivan Malov <ivan.malov at oktetlabs.ru>
> Reviewed-by: Andrew Rybchenko <andrew.rybchenko at oktetlabs.ru>
> ---
>  drivers/net/bonding/rte_eth_bond_pmd.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c
> index 4a0f6e1b8..7e79bac42 100644
> --- a/drivers/net/bonding/rte_eth_bond_pmd.c
> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
> @@ -2115,6 +2115,10 @@ bond_ethdev_close(struct rte_eth_dev *dev)
>         struct rte_flow_error ferror;
>
>         RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
> +
> +       /* Flush flows in all back-end devices before removing them */
> +       bond_flow_ops.flush(dev, &ferror);
> +
>         while (internals->slave_count != skipped) {
>                 uint16_t port_id = internals->slaves[skipped].port_id;
>
> @@ -2127,7 +2131,6 @@ bond_ethdev_close(struct rte_eth_dev *dev)
>                         skipped++;
>                 }
>         }
> -       bond_flow_ops.flush(dev, &ferror);
>         bond_ethdev_free_queues(dev);
>         rte_bitmap_reset(internals->vlan_filter_bmp);
>  }
> --
> 2.30.2
>


-- 
Christian Ehrhardt
Senior Staff Engineer, Ubuntu Server
Canonical Ltd