[3/7] net/hns3: fix packets's offload features flags in Rx

Message ID 20200317091206.34928-4-huwei013@chinasoftinc.com (mailing list archive)
State Accepted, archived
Delegated to: Ferruh Yigit
Headers
Series misc updates for hns3 PMD driver |

Checks

Context Check Description
ci/checkpatch success coding style OK

Commit Message

Wei Hu (Xavier) March 17, 2020, 9:12 a.m. UTC
  From: Chengwen Feng <fengchengwen@huawei.com>

Currently there is a certain probability of the unexpected ol_flag of the
Rx packets's rte_mbuf when receiving packets.

The root cause as below:
1. The member variable named ol_flag of the structure named rte_mbuf is not
   properly initialized to zero in the '.rx_pkt_burst' ops implementation
   function named hns3_recv_pkts.
2. When multi-segment rte_mbufs are needed for long packet in Rx operation,
   the driver should assign value to the ol_flag of the first segment, not
   to the ol_flag of the last segment.

This patch fixes it with the following modification in the '.rx_pkt_burst'
ops implementation function named hns3_recv_pkts.
1. Where the first write operation in the '.rx_pkt_burst' ops
   implementation function, assign PKT_RX_RSS_HASH to ol_flags directly
   using '=' operation instead of '|=' operation.
2. In the static function named hns3_rx_set_cksum_flag, the last rte_mbuf's
   ol_flags should be assigned when processing multi-segment. We fix it by
   passing first_seg variable to the function instead of rxm(the last
   segment's address).

Fixes: bba636698316 ("net/hns3: support Rx/Tx and related operations")
Fixes: ad7cf94823e8 ("net/hns3: fix offload flag for RSS hash")
Cc: stable@dpdk.org

Signed-off-by: Chengwen Feng <fengchengwen@huawei.com>
Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
---
 drivers/net/hns3/hns3_rxtx.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
  

Patch

diff --git a/drivers/net/hns3/hns3_rxtx.c b/drivers/net/hns3/hns3_rxtx.c
index ec6d19f58..0c965b1b8 100644
--- a/drivers/net/hns3/hns3_rxtx.c
+++ b/drivers/net/hns3/hns3_rxtx.c
@@ -1582,7 +1582,7 @@  hns3_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
 		first_seg->pkt_len = pkt_len;
 		first_seg->port = rxq->port_id;
 		first_seg->hash.rss = rte_le_to_cpu_32(rxd.rx.rss_hash);
-		first_seg->ol_flags |= PKT_RX_RSS_HASH;
+		first_seg->ol_flags = PKT_RX_RSS_HASH;
 		if (unlikely(hns3_get_bit(bd_base_info, HNS3_RXD_LUM_B))) {
 			first_seg->hash.fdir.hi =
 				rte_le_to_cpu_32(rxd.rx.fd_id);
@@ -1599,7 +1599,8 @@  hns3_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
 								  ol_info);
 
 		if (bd_base_info & BIT(HNS3_RXD_L3L4P_B))
-			hns3_rx_set_cksum_flag(rxm, first_seg->packet_type,
+			hns3_rx_set_cksum_flag(first_seg,
+					       first_seg->packet_type,
 					       cksum_err);
 
 		first_seg->vlan_tci = rte_le_to_cpu_16(rxd.rx.vlan_tag);