On 9/30/2026 4:48 AM, Anurag Mandal wrote:
ice_vsi_dis_outer_insertion() treats a 0 return from ice_update_vsi() as a failure, but a return value 0 means success, so the error and success paths are swapped. Removing a port VLAN in double VLAN mode logs an incorrect error, returns invalid argument, and leaves a stale ICE_AQ_VSI_OUTER_VLAN_PORT_BASED_INSERT bit in the cached VSI context that later VSI updates replay into hardware.Inverted the check to avoid this bug. Fixes: de5da9d16430 ("net/ice: support double VLAN") Cc: [email protected] Signed-off-by: Anurag Mandal <[email protected]> --- drivers/net/intel/ice/ice_ethdev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/intel/ice/ice_ethdev.c b/drivers/net/intel/ice/ice_ethdev.c index 163304ba1e..2c5ffe07f3 100644 --- a/drivers/net/intel/ice/ice_ethdev.c +++ b/drivers/net/intel/ice/ice_ethdev.c @@ -6447,7 +6447,7 @@ static int ice_vsi_dis_outer_insertion(struct ice_vsi *vsi, struct ice_vsi_vlan_ ICE_AQ_VSI_OUTER_VLAN_TX_MODE_M);status = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);- if (!status) { + if (status) { PMD_DRV_LOG(ERR, "update VSI for disabling outer VLAN insertion failed, err %d", status);
I'd have preferred we do not treat integer values as booleans. This particular function returns the ice status code rather than int so perhaps we should check for `if (status != ICE_SUCCESS)` instead.
However, this isn't a big deal and can probably be fixed on apply, so Acked-by: Anatoly Burakov <[email protected]> -- Thanks, Anatoly
