From: Brett Creeley brett.creeley@intel.com
[ Upstream commit e25f9152bc07de534b2b590ce6c052ea25dd8900 ]
Currently there can be a case where a DCB map is applied and there are more interrupt vectors (vsi->num_q_vectors) than Rx queues (vsi->num_rxq) and Tx queues (vsi->num_txq). If we try to set coalesce settings in this case it will report a false failure. Fix this by checking if vector index is valid with respect to the number of Tx and Rx queues configured.
Signed-off-by: Brett Creeley brett.creeley@intel.com Tested-by: Andrew Bowers andrewx.bowers@intel.com Signed-off-by: Jeff Kirsher jeffrey.t.kirsher@intel.com Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/net/ethernet/intel/ice/ice_ethtool.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_ethtool.c b/drivers/net/ethernet/intel/ice/ice_ethtool.c index 7e23034df955..1fe9f6050635 100644 --- a/drivers/net/ethernet/intel/ice/ice_ethtool.c +++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c @@ -3368,10 +3368,17 @@ __ice_set_coalesce(struct net_device *netdev, struct ethtool_coalesce *ec, struct ice_vsi *vsi = np->vsi;
if (q_num < 0) { - int i; + int v_idx; + + ice_for_each_q_vector(vsi, v_idx) { + /* In some cases if DCB is configured the num_[rx|tx]q + * can be less than vsi->num_q_vectors. This check + * accounts for that so we don't report a false failure + */ + if (v_idx >= vsi->num_rxq && v_idx >= vsi->num_txq) + goto set_complete;
- ice_for_each_q_vector(vsi, i) { - if (ice_set_q_coalesce(vsi, ec, i)) + if (ice_set_q_coalesce(vsi, ec, v_idx)) return -EINVAL; } goto set_complete;