RE: [PATCH] common/idpf: fix Tx checksum offload

RE: [PATCH] common/idpf: fix Tx checksum offload

[Huang, ZhiminX]

> -----Original Message-----
> From: beilei.xing@intel.com <beilei.xing@intel.com>
> Sent: Friday, September 22, 2023 11:30 PM
> To: Wu, Jingjing <jingjing.wu@intel.com>
> Cc: dev@dpdk.org; Xing, Beilei <beilei.xing@intel.com>; stable@dpdk.org
> Subject: [PATCH] common/idpf: fix Tx checksum offload
>
> From: Beilei Xing <beilei.xing@intel.com>
>
> For multi-segment packets, the Tx checksum offload doesn't work except the
> last segment, because other segments don't enable HW checksum offload
> successfully.
>
> Fixes: ef47d95e9031 ("net/idpf: fix TSO")
> Fixes: 8c6098afa075 ("common/idpf: add Rx/Tx data path")
> Cc: stable@dpdk.org
>
> Signed-off-by: Beilei Xing <beilei.xing@intel.com>
> ---
Verified scalar path + single queue and split queue, tested pass for TX checksum on TSO offload.
Tested-by: Zhimin Huang <zhiminx.huang@intel.com >

[PATCH] common/idpf: fix Tx checksum offload

[beilei.xing]

From: Beilei Xing <beilei.xing@intel.com>

For multi-segment packets, the Tx checksum offload doesn't work
except the last segment, because other segments don't enable HW
checksum offload successfully.

Fixes: ef47d95e9031 ("net/idpf: fix TSO")
Fixes: 8c6098afa075 ("common/idpf: add Rx/Tx data path")
Cc: stable@dpdk.org

Signed-off-by: Beilei Xing <beilei.xing@intel.com>
---
 drivers/common/idpf/idpf_common_rxtx.c | 17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

diff --git a/drivers/common/idpf/idpf_common_rxtx.c b/drivers/common/idpf/idpf_common_rxtx.c
index e6d2486272..83b131ef28 100644
--- a/drivers/common/idpf/idpf_common_rxtx.c
+++ b/drivers/common/idpf/idpf_common_rxtx.c
@@ -871,6 +871,7 @@ idpf_dp_splitq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
 	uint16_t nb_to_clean;
 	uint16_t nb_tx = 0;
 	uint64_t ol_flags;
+	uint8_t cmd_dtype;
 	uint16_t nb_ctx;
 
 	if (unlikely(txq == NULL) || unlikely(!txq->q_started))
@@ -902,6 +903,7 @@ idpf_dp_splitq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
 		if (txq->nb_free < tx_pkt->nb_segs)
 			break;
 
+		cmd_dtype = 0;
 		ol_flags = tx_pkt->ol_flags;
 		tx_offload.l2_len = tx_pkt->l2_len;
 		tx_offload.l3_len = tx_pkt->l3_len;
@@ -911,6 +913,9 @@ idpf_dp_splitq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
 		nb_ctx = idpf_calc_context_desc(ol_flags);
 		nb_used = tx_pkt->nb_segs + nb_ctx;
 
+		if (ol_flags & IDPF_TX_CKSUM_OFFLOAD_MASK)
+			cmd_dtype = IDPF_TXD_FLEX_FLOW_CMD_CS_EN;
+
 		/* context descriptor */
 		if (nb_ctx != 0) {
 			volatile union idpf_flex_tx_ctx_desc *ctx_desc =
@@ -933,8 +938,8 @@ idpf_dp_splitq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
 			/* Setup TX descriptor */
 			txd->buf_addr =
 				rte_cpu_to_le_64(rte_mbuf_data_iova(tx_pkt));
-			txd->qw1.cmd_dtype =
-				rte_cpu_to_le_16(IDPF_TX_DESC_DTYPE_FLEX_FLOW_SCHE);
+			cmd_dtype |= IDPF_TX_DESC_DTYPE_FLEX_FLOW_SCHE;
+			txd->qw1.cmd_dtype = cmd_dtype;
 			txd->qw1.rxr_bufsize = tx_pkt->data_len;
 			txd->qw1.compl_tag = sw_id;
 			tx_id++;
@@ -948,8 +953,6 @@ idpf_dp_splitq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
 		/* fill the last descriptor with End of Packet (EOP) bit */
 		txd->qw1.cmd_dtype |= IDPF_TXD_FLEX_FLOW_CMD_EOP;
 
-		if (ol_flags & IDPF_TX_CKSUM_OFFLOAD_MASK)
-			txd->qw1.cmd_dtype |= IDPF_TXD_FLEX_FLOW_CMD_CS_EN;
 		txq->nb_free = (uint16_t)(txq->nb_free - nb_used);
 		txq->nb_used = (uint16_t)(txq->nb_used + nb_used);
 
@@ -1424,6 +1427,9 @@ idpf_dp_singleq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
 			}
 		}
 
+		if (ol_flags & IDPF_TX_CKSUM_OFFLOAD_MASK)
+			td_cmd |= IDPF_TX_FLEX_DESC_CMD_CS_EN;
+
 		if (nb_ctx != 0) {
 			/* Setup TX context descriptor if required */
 			volatile union idpf_flex_tx_ctx_desc *ctx_txd =
@@ -1487,9 +1493,6 @@ idpf_dp_singleq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
 			txq->nb_used = 0;
 		}
 
-		if (ol_flags & IDPF_TX_CKSUM_OFFLOAD_MASK)
-			td_cmd |= IDPF_TX_FLEX_DESC_CMD_CS_EN;
-
 		txd->qw1 |= rte_cpu_to_le_16(td_cmd << IDPF_TXD_QW1_CMD_S);
 	}
 
-- 
2.34.1

RE: [PATCH] common/idpf: fix Tx checksum offload

[Zhang, Qi Z]

> -----Original Message-----
> From: Huang, ZhiminX <zhiminx.huang@intel.com>
> Sent: Friday, September 22, 2023 4:24 PM
> To: Xing, Beilei <beilei.xing@intel.com>; Wu, Jingjing <jingjing.wu@intel.com>
> Cc: dev@dpdk.org; Xing, Beilei <beilei.xing@intel.com>; stable@dpdk.org
> Subject: RE: [PATCH] common/idpf: fix Tx checksum offload
>
> > -----Original Message-----
> > From: beilei.xing@intel.com <beilei.xing@intel.com>
> > Sent: Friday, September 22, 2023 11:30 PM
> > To: Wu, Jingjing <jingjing.wu@intel.com>
> > Cc: dev@dpdk.org; Xing, Beilei <beilei.xing@intel.com>;
> > stable@dpdk.org
> > Subject: [PATCH] common/idpf: fix Tx checksum offload
> >
> > From: Beilei Xing <beilei.xing@intel.com>
> >
> > For multi-segment packets, the Tx checksum offload doesn't work except
> > the last segment, because other segments don't enable HW checksum
> > offload successfully.
> >
> > Fixes: ef47d95e9031 ("net/idpf: fix TSO")
> > Fixes: 8c6098afa075 ("common/idpf: add Rx/Tx data path")
> > Cc: stable@dpdk.org
> >
> > Signed-off-by: Beilei Xing <beilei.xing@intel.com>
> > ---
> Verified scalar path + single queue and split queue, tested pass for TX
> checksum on TSO offload.
> Tested-by: Zhimin Huang <zhiminx.huang@intel.com >

Applied to dpdk-next-net-intel.

Thanks
Qi