[dpdk-dev] [PATCH 11/25] net/bnxt: fix crash in VF rep queue selection
Ajit Khaparde
ajit.khaparde at broadcom.com
Fri Sep 11 03:55:49 CEST 2020
From: Somnath Kotur <somnath.kotur at broadcom.com>
Instead of bounds checking against max possible rings while selecting
queue index for the VF rep, do it against the configured number of rings.
Fixes: 6dc83230b43b ("net/bnxt: support port representor data path")
Signed-off-by: Somnath Kotur <somnath.kotur at broadcom.com>
Reviewed-by: Venkat Duvvuru <venkatkumar.duvvuru at broadcom.com>
---
drivers/net/bnxt/bnxt_reps.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/drivers/net/bnxt/bnxt_reps.c b/drivers/net/bnxt/bnxt_reps.c
index 9f535c0ce..6fa867a8a 100644
--- a/drivers/net/bnxt/bnxt_reps.c
+++ b/drivers/net/bnxt/bnxt_reps.c
@@ -45,9 +45,12 @@ bnxt_vfr_recv(uint16_t port_id, uint16_t queue_id, struct rte_mbuf *mbuf)
vfr_eth_dev = &rte_eth_devices[port_id];
vfr_bp = vfr_eth_dev->data->dev_private;
- /* If rxq_id happens to be > max rep_queue, use rxq0 */
- que = queue_id < BNXT_MAX_VF_REP_RINGS ? queue_id : 0;
+ /* If rxq_id happens to be > nr_rings, use ring 0 */
+ que = queue_id < vfr_bp->rx_nr_rings ? queue_id : 0;
rep_rxq = vfr_bp->rx_queues[que];
+ /* Ideally should not happen now, paranoid check */
+ if (!rep_rxq)
+ return 1;
rep_rxr = rep_rxq->rx_ring;
mask = rep_rxr->rx_ring_struct->ring_mask;
--
2.21.1 (Apple Git-122.3)
More information about the dev
mailing list