[dpdk-dev] [PATCH 1/2] net/mlx5: fix setting offsets for SW parser

[dpdk-dev] [PATCH 1/2] net/mlx5: fix setting offsets for SW parser

[Yongseok Koh]

Since ConnectX-5, SW parser just complements HW parser. SW parser starts to
engage only if HW parser can't reach a header. For the older devices, HW
parser will not kick in if any of SWP offsets is set. Therefore, all of the
L3 offsets should be set regardless of HW offload. As IPv6 doesn't have
header checksum, the mbuf can't have PKT_TX_[OUTER_]IP_CKSUM if outer or
inner L3 is IPv6.

And if inner packet isn't IP, the inner offsets shouldn't be set.

Fixes: 5f8ba81c4228 ("net/mlx5: support generic tunnel offloading")

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5_rxtx.h | 43 +++++++++++++++++++++++++++----------------
 1 file changed, 27 insertions(+), 16 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index a6017f0dd..1e4b2fdb9 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -655,10 +655,13 @@ txq_mbuf_to_swp(struct mlx5_txq_data *txq, struct rte_mbuf *buf,
 		 uint8_t *offsets, uint8_t *swp_types)
 {
 	uint64_t tunnel = buf->ol_flags & PKT_TX_TUNNEL_MASK;
-	uint16_t idx;
-	uint16_t off;
+	const uint64_t csum_flags = buf->ol_flags & PKT_TX_L4_MASK;
+	const uint64_t inner_ip =
+		buf->ol_flags & (PKT_TX_IPV4 | PKT_TX_IPV6);
 	const uint64_t ol_flags_mask = PKT_TX_L4_MASK | PKT_TX_IPV6 |
 				       PKT_TX_OUTER_IPV6;
+	uint16_t idx;
+	uint16_t off;
 
 	if (likely(!tunnel || !txq->swp_en ||
 		   (tunnel != PKT_TX_TUNNEL_UDP && tunnel != PKT_TX_TUNNEL_IP)))
@@ -674,20 +677,28 @@ txq_mbuf_to_swp(struct mlx5_txq_data *txq, struct rte_mbuf *buf,
 	if (tunnel == PKT_TX_TUNNEL_UDP)
 		idx |= 1 << 9;
 	*swp_types = mlx5_swp_types_table[idx];
-	/* swp offsets. */
-	off = buf->outer_l2_len + (vlan ? 4 : 0); /* Outer L3 offset. */
-	if (tso || (buf->ol_flags & PKT_TX_OUTER_IP_CKSUM))
-		offsets[1] = off >> 1;
-	off += buf->outer_l3_len; /* Outer L4 offset. */
-	if (tunnel == PKT_TX_TUNNEL_UDP)
-		offsets[0] = off >> 1;
-	off += buf->l2_len; /* Inner L3 offset. */
-	if (tso || (buf->ol_flags & PKT_TX_IP_CKSUM))
-		offsets[3] = off >> 1;
-	off += buf->l3_len; /* Inner L4 offset. */
-	if (tso || ((buf->ol_flags & PKT_TX_L4_MASK) == PKT_TX_TCP_CKSUM) ||
-	    ((buf->ol_flags & PKT_TX_L4_MASK) == PKT_TX_UDP_CKSUM))
-		offsets[2] = off >> 1;
+	/*
+	 * Set offsets for SW parser. Since ConnectX-5, SW parser just
+	 * complements HW parser. SW parser starts to engage only if HW parser
+	 * can't reach a header. For the older devices, HW parser will not kick
+	 * in if any of SWP offsets is set. Therefore, all of the L3 offsets
+	 * should be set regardless of HW offload.
+	 */
+	off = buf->outer_l2_len + (vlan ? 4 : 0);
+	offsets[1] = off >> 1; /* Outer L3 offset. */
+	if (tunnel == PKT_TX_TUNNEL_UDP) {
+		off += buf->outer_l3_len;
+		offsets[0] = off >> 1; /* Outer L4 offset. */
+	}
+	if (inner_ip) {
+		off += buf->l2_len;
+		offsets[3] = off >> 1; /* Inner L3 offset. */
+		if (csum_flags == PKT_TX_TCP_CKSUM || tso ||
+		    csum_flags == PKT_TX_UDP_CKSUM) {
+			off += buf->l3_len;
+			offsets[2] = off >> 1; /* Inner L4 offset. */
+		}
+	}
 }
 
 /**
-- 
2.11.0

[dpdk-dev] [PATCH 2/2] net/mlx5: remove redundant checks

[Yongseok Koh]

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
---
 drivers/net/mlx5/mlx5_rxtx.c |  6 ++----
 drivers/net/mlx5/mlx5_rxtx.h | 14 +++++++-------
 2 files changed, 9 insertions(+), 11 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index cdd373e3e..527859461 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -522,7 +522,6 @@ mlx5_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 		uint16_t ehdr;
 		uint8_t cs_flags;
 		uint8_t tso = txq->tso_en && (buf->ol_flags & PKT_TX_TCP_SEG);
-		uint8_t is_vlan = !!(buf->ol_flags & PKT_TX_VLAN_PKT);
 		uint32_t swp_offsets = 0;
 		uint8_t swp_types = 0;
 		uint16_t tso_segsz = 0;
@@ -566,11 +565,10 @@ mlx5_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 			rte_prefetch0(
 			    rte_pktmbuf_mtod(*(pkts + 1), volatile void *));
 		cs_flags = txq_ol_cksum_to_cs(buf);
-		txq_mbuf_to_swp(txq, buf, tso, is_vlan,
-				(uint8_t *)&swp_offsets, &swp_types);
+		txq_mbuf_to_swp(txq, buf, (uint8_t *)&swp_offsets, &swp_types);
 		raw = ((uint8_t *)(uintptr_t)wqe) + 2 * MLX5_WQE_DWORD_SIZE;
 		/* Replace the Ethernet type by the VLAN if necessary. */
-		if (is_vlan) {
+		if (buf->ol_flags & PKT_TX_VLAN_PKT) {
 			uint32_t vlan = rte_cpu_to_be_32(0x81000000 |
 							 buf->vlan_tci);
 			unsigned int len = 2 * ETHER_ADDR_LEN - 2;
diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index 1e4b2fdb9..5d9a7dffd 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -651,10 +651,10 @@ mlx5_tx_dbrec(struct mlx5_txq_data *txq, volatile struct mlx5_wqe *wqe)
  */
 static __rte_always_inline void
 txq_mbuf_to_swp(struct mlx5_txq_data *txq, struct rte_mbuf *buf,
-		 uint8_t tso, uint64_t vlan,
-		 uint8_t *offsets, uint8_t *swp_types)
+		uint8_t *offsets, uint8_t *swp_types)
 {
-	uint64_t tunnel = buf->ol_flags & PKT_TX_TUNNEL_MASK;
+	const uint64_t vlan = buf->ol_flags & PKT_TX_VLAN_PKT;
+	const uint64_t tunnel = buf->ol_flags & PKT_TX_TUNNEL_MASK;
 	const uint64_t csum_flags = buf->ol_flags & PKT_TX_L4_MASK;
 	const uint64_t inner_ip =
 		buf->ol_flags & (PKT_TX_IPV4 | PKT_TX_IPV6);
@@ -663,8 +663,8 @@ txq_mbuf_to_swp(struct mlx5_txq_data *txq, struct rte_mbuf *buf,
 	uint16_t idx;
 	uint16_t off;
 
-	if (likely(!tunnel || !txq->swp_en ||
-		   (tunnel != PKT_TX_TUNNEL_UDP && tunnel != PKT_TX_TUNNEL_IP)))
+	if (likely(!txq->swp_en || (tunnel != PKT_TX_TUNNEL_UDP &&
+				    tunnel != PKT_TX_TUNNEL_IP)))
 		return;
 	/*
 	 * The index should have:
@@ -684,7 +684,7 @@ txq_mbuf_to_swp(struct mlx5_txq_data *txq, struct rte_mbuf *buf,
 	 * in if any of SWP offsets is set. Therefore, all of the L3 offsets
 	 * should be set regardless of HW offload.
 	 */
-	off = buf->outer_l2_len + (vlan ? 4 : 0);
+	off = buf->outer_l2_len + (vlan ? sizeof(struct vlan_hdr) : 0);
 	offsets[1] = off >> 1; /* Outer L3 offset. */
 	if (tunnel == PKT_TX_TUNNEL_UDP) {
 		off += buf->outer_l3_len;
@@ -693,7 +693,7 @@ txq_mbuf_to_swp(struct mlx5_txq_data *txq, struct rte_mbuf *buf,
 	if (inner_ip) {
 		off += buf->l2_len;
 		offsets[3] = off >> 1; /* Inner L3 offset. */
-		if (csum_flags == PKT_TX_TCP_CKSUM || tso ||
+		if (csum_flags == PKT_TX_TCP_CKSUM ||
 		    csum_flags == PKT_TX_UDP_CKSUM) {
 			off += buf->l3_len;
 			offsets[2] = off >> 1; /* Inner L4 offset. */
-- 
2.11.0