Re: [Intel-wired-lan] [PATCH net-next 09/10] bonding: fill IPsec state validation failure reason

[Leon Romanovsky <leon@kernel.org>]

On Mon, Jan 23, 2023 at 03:01:32PM -0800, Jay Vosburgh wrote:
> Leon Romanovsky <leon@kernel.org> wrote:
> 
> >From: Leon Romanovsky <leonro@nvidia.com>
> >
> >Rely on extack to return failure reason.
> >
> >Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
> >Signed-off-by: Leon Romanovsky <leon@kernel.org>
> >---
> > drivers/net/bonding/bond_main.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> >diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> >index 686b2a6fd674..00646aa315c3 100644
> >--- a/drivers/net/bonding/bond_main.c
> >+++ b/drivers/net/bonding/bond_main.c
> >@@ -444,7 +444,7 @@ static int bond_ipsec_add_sa(struct xfrm_state *xs,
> > 	if (!slave->dev->xfrmdev_ops ||
> > 	    !slave->dev->xfrmdev_ops->xdo_dev_state_add ||
> > 	    netif_is_bond_master(slave->dev)) {
> >-		slave_warn(bond_dev, slave->dev, "Slave does not support ipsec offload\n");
> >+		NL_SET_ERR_MSG_MOD(extack, "Slave does not support ipsec offload");
> > 		rcu_read_unlock();
> > 		return -EINVAL;
> > 	}
> 
> 	Why only this one, and not include the other similar
> slave_warn() calls in the bond_ipsec_* functions?  

Which functions did you have in mind?

The extack was added to XFRM .xdo_dev_state_add() call, which is
translated to bond_ipsec_add_sa() with only one slave_warn() print.

If you are talking about bond_ipsec_add_sa_all(), that function isn't
directly connected to netlink and doesn't have extack pointer to fill.

If you are talking about bond_ipsec_del_sai*() and slave_warn() there, it
will be better to be deleted/changed to make sure what ipsec_list have
only valid devices.

Thanks


> That would seem to make some failures show up in dmesg,
> and others returned to the caller via extack.
> 
> 	-J
> 
> ---
> 	-Jay Vosburgh, jay.vosburgh@canonical.com