On 10/21/2021 9:03 AM, Eli Britstein wrote:
RX/TX functions (rte_eth_rx_burst/rte_eth_tx_burst) get 'nb_pkts'
argument, which specifies the maximum number to receive/transmit.
It can be 0..nb_pkts, meaning nb_pkts+1 options.
Enlarge the spread stats array by one cell to fix the possible out of
range memory access.
This looks a valid problem, I guess problem is able to hide this long
because of big MAX_PKT_BURST value in testpmd.
Can you please briefly describe what 'pkt_burst_spread[]' is in the
commit log? That makes easy to understand the problem.
Fixes: af75078fece3 ("first public release")
Cc: sta...@dpdk.org
Signed-off-by: Eli Britstein <el...@nvidia.com>
Reviewed-by: Matan Azrad <ma...@mellanox.com>
Acked-by: Ferruh Yigit <ferruh.yi...@intel.com>
