If an unsupported interface is requested to be validated then there is no need in further capabilities and-ing since the local array will be left initialized with zeros. Let's explicitly return EINVAL error in that case in order to inform the caller about invalid link-state interface. In any case the phylink_validate_mac_and_pcs() would terminate with error further link-state validation so the suggested update won't change the validation procedure semantics. Signed-off-by: Serge Semin <fancer.lancer@xxxxxxxxx> --- drivers/net/pcs/pcs-xpcs.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/net/pcs/pcs-xpcs.c b/drivers/net/pcs/pcs-xpcs.c index 92c47da61db4..46afeb5510c0 100644 --- a/drivers/net/pcs/pcs-xpcs.c +++ b/drivers/net/pcs/pcs-xpcs.c @@ -613,14 +613,15 @@ static int xpcs_validate(struct phylink_pcs *pcs, unsigned long *supported, xpcs = phylink_pcs_to_xpcs(pcs); compat = xpcs_find_compat(xpcs->id, state->interface); + if (!compat) + return -EINVAL; /* Populate the supported link modes for this PHY interface type. * FIXME: what about the port modes and autoneg bit? This masks * all those away. */ - if (compat) - for (i = 0; compat->supported[i] != __ETHTOOL_LINK_MODE_MASK_NBITS; i++) - set_bit(compat->supported[i], xpcs_supported); + for (i = 0; compat->supported[i] != __ETHTOOL_LINK_MODE_MASK_NBITS; i++) + set_bit(compat->supported[i], xpcs_supported); linkmode_and(supported, supported, xpcs_supported); -- 2.42.1