diff mbox

[net-next,04/13] i40e: Fix dangling ring pointers upon driver removal

Message ID 1403250588-14356-5-git-send-email-jeffrey.t.kirsher@intel.com
State Accepted, archived
Delegated to: David Miller
Headers show

Commit Message

Kirsher, Jeffrey T June 20, 2014, 7:49 a.m. UTC 
From: Anjali Singhai Jain <anjali.singhai@intel.com>

When we resize the number of queues, the driver needs to disassociate any
qvectors that are no longer in use from the original rings, this way we
do not try to access the rings through these qvectors at the time of freeing
the qvectors.

Change-ID: Ie4eb9fc749f8e12348517fe1560f599c58f4a2a4
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
diff mbox

Patch

diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 95c331a..7d2aeeb 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -3151,8 +3151,12 @@  static void i40e_vsi_map_rings_to_vectors(struct i40e_vsi *vsi)
 
 	/* If we don't have enough vectors for a 1-to-1 mapping, we'll have to
 	 * group them so there are multiple queues per vector.
+	 * It is also important to go through all the vectors available to be
+	 * sure that if we don't use all the vectors, that the remaining vectors
+	 * are cleared. This is especially important when decreasing the
+	 * number of queues in use.
 	 */
-	for (; v_start < q_vectors && qp_remaining; v_start++) {
+	for (; v_start < q_vectors; v_start++) {
 		struct i40e_q_vector *q_vector = vsi->q_vectors[v_start];
 
 		num_ringpairs = DIV_ROUND_UP(qp_remaining, q_vectors - v_start);