On vr, 02 nov 2018 12:32:20 +0100, Oleksij Rempel wrote:
>
> better to know if we do wrong mapping, instead of silently "fix" it
> on this place.
Do we agree that this 'should not occur'?
This is a workaround for a deeper problem, shouldn't we attach the
deeper problem?
>
> Signed-off-by: Oleksij Rempel <o.rempel@xxxxxxxxxxxxxx>
> ---
> net/can/j1939/bus.c | 15 +++++++++++++--
> 1 file changed, 13 insertions(+), 2 deletions(-)
>
> diff --git a/net/can/j1939/bus.c b/net/can/j1939/bus.c
> index 171fbdb2a1a8..9ebaa497139b 100644
> --- a/net/can/j1939/bus.c
> +++ b/net/can/j1939/bus.c
> @@ -51,14 +51,25 @@ static bool j1939_ecu_is_mapped_locked(struct j1939_ecu *ecu)
> /* map ECU to a bus address space */
> static void j1939_ecu_map_locked(struct j1939_ecu *ecu)
> {
> + struct j1939_priv *priv = ecu->priv;
> + struct j1939_addr_ent *ent;
> +
> lockdep_assert_held(&ecu->priv->lock);
>
> if (!j1939_address_is_unicast(ecu->addr))
> return;
>
> + ent = &ecu->priv->ents[ecu->addr];
> +
> + if (ent->ecu) {
> + netdev_warn(priv->ndev, "Trying to map already mapped ECU, addr: 0x%02x, name: 0x%016llx. Skip it.\n",
> + ecu->addr, ecu->name);
> + return;
> + }
> +
> j1939_ecu_get(ecu);
> - ecu->priv->ents[ecu->addr].ecu = ecu;
> - ecu->priv->ents[ecu->addr].nusers += ecu->nusers;
> + ent->ecu = ecu;
> + ent->nusers += ecu->nusers;
> }
>
> /* unmap ECU from a bus address space */
> --
> 2.19.1
>
