[PATCH 0/6] fix the set flow actions

[PATCH 0/6] fix the set flow actions

[Chaoyong He]

The former logic of set flow actions don't consider the mask filed
of control message passed to the firmware, and this caused the
firmware skip the set action logic wrongly.
This patch series aims to fix this problem.

Chaoyong He (6):
  net/nfp: fix the set MAC flow action
  net/nfp: fix the set IPv4 flow action
  net/nfp: fix the set IPv6 flow action
  net/nfp: fix the set TP flow action
  net/nfp: fix the set TTL flow action
  net/nfp: fix the set IPv4/v6 DSCP flow action

 drivers/net/nfp/nfp_flow.c | 29 +++++++++++++++++++++++------
 1 file changed, 23 insertions(+), 6 deletions(-)

-- 
2.29.3

[PATCH 1/6] net/nfp: fix the set MAC flow action

[Chaoyong He]

The former logic of set MAC source/destination address flow action don't
consider the mask filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right MAC address as expected.

Fixes: 4f6983154570 ("net/nfp: support MAC source flow action")
Fixes: eecc7ca3088a ("net/nfp: support MAC destination flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index 0c38925701..f373171d7e 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2066,6 +2066,7 @@ nfp_flow_action_set_mac(char *act_data,
 		bool mac_src_flag,
 		bool mac_set_flag)
 {
+	uint8_t i;
 	size_t act_size;
 	struct nfp_fl_act_set_eth *set_eth;
 	const struct rte_flow_action_set_mac *set_mac;
@@ -2084,9 +2085,13 @@ nfp_flow_action_set_mac(char *act_data,
 	if (mac_src_flag) {
 		rte_memcpy(&set_eth->eth_addr[RTE_ETHER_ADDR_LEN],
 				set_mac->mac_addr, RTE_ETHER_ADDR_LEN);
+		for (i = 0; i < RTE_ETHER_ADDR_LEN; i++)
+			set_eth->eth_addr_mask[RTE_ETHER_ADDR_LEN + i] = 0xff;
 	} else {
 		rte_memcpy(&set_eth->eth_addr[0],
 				set_mac->mac_addr, RTE_ETHER_ADDR_LEN);
+		for (i = 0; i < RTE_ETHER_ADDR_LEN; i++)
+			set_eth->eth_addr_mask[i] = 0xff;
 	}
 }
 
-- 
2.29.3

[PATCH 2/6] net/nfp: fix the set IPv4 flow action

[Chaoyong He]

The former logic of set IPv4 source/destination address flow action don't
consider the mask filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right IPv4 address as expected.

Fixes: ab80975b0a14 ("net/nfp: support IPv4 source flow action")
Fixes: f85c1e05bff3 ("net/nfp: support IPv4 destination flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index f373171d7e..39493ce98d 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2132,10 +2132,13 @@ nfp_flow_action_set_ip(char *act_data,
 	set_ip->reserved     = 0;
 
 	set_ipv4 = (const struct rte_flow_action_set_ipv4 *)action->conf;
-	if (ip_src_flag)
+	if (ip_src_flag) {
 		set_ip->ipv4_src = set_ipv4->ipv4_addr;
-	else
+		set_ip->ipv4_src_mask = RTE_BE32(0xffffffff);
+	} else {
 		set_ip->ipv4_dst = set_ipv4->ipv4_addr;
+		set_ip->ipv4_dst_mask = RTE_BE32(0xffffffff);
+	}
 }
 
 static void
-- 
2.29.3

[PATCH 3/6] net/nfp: fix the set IPv6 flow action

[Chaoyong He]

The former logic of set IPv6 source/destination address flow action don't
consider the mask filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right IPv6 address as expected.

Fixes: 596ae2217214 ("net/nfp: support IPv6 source flow action")
Fixes: 51384f79b264 ("net/nfp: support IPv6 destination flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index 39493ce98d..a4561b08e8 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2163,8 +2163,10 @@ nfp_flow_action_set_ipv6(char *act_data,
 	set_ip->head.len_lw = act_size >> NFP_FL_LW_SIZ;
 	set_ip->reserved = 0;
 
-	for (i = 0; i < 4; i++)
-		set_ip->ipv6[i].exact = set_ipv6->ipv6_addr[i];
+	for (i = 0; i < 4; i++) {
+		set_ip->ipv6[i].exact = *(const rte_be32_t *)&set_ipv6->ipv6_addr[i * 4];
+		set_ip->ipv6[i].mask = RTE_BE32(0xffffffff);
+	}
 }
 
 static void
-- 
2.29.3

[PATCH 4/6] net/nfp: fix the set TP flow action

[Chaoyong He]

The former logic of set TP source/destination flow action don't
consider the mask filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right TP as expected.

Fixes: fc185097bbe6 ("net/nfp: support TP source flow action")
Fixes: 87986df09d75 ("net/nfp: support TP destination flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index a4561b08e8..f22e89e1ea 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2190,10 +2190,13 @@ nfp_flow_action_set_tp(char *act_data,
 	set_tp->reserved     = 0;
 
 	set_tp_conf = (const struct rte_flow_action_set_tp *)action->conf;
-	if (tp_src_flag)
+	if (tp_src_flag) {
 		set_tp->src_port = set_tp_conf->port;
-	else
+		set_tp->src_port_mask = RTE_BE16(0xffff);
+	} else {
 		set_tp->dst_port = set_tp_conf->port;
+		set_tp->dst_port_mask = RTE_BE16(0xffff);
+	}
 }
 
 static int
-- 
2.29.3

[PATCH 5/6] net/nfp: fix the set TTL flow action

[Chaoyong He]

The former logic of set TTL flow action don't consider the mask
filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right TTL as expected.

Fixes: ac12e126c482 ("net/nfp: support TTL flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index f22e89e1ea..c6e19d8cac 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2252,6 +2252,7 @@ nfp_flow_action_set_ttl(char *act_data,
 
 	ttl_conf = (const struct rte_flow_action_set_ttl *)action->conf;
 	ttl_tos->ipv4_ttl = ttl_conf->ttl_value;
+	ttl_tos->ipv4_ttl_mask = 0xff;
 	ttl_tos->reserved = 0;
 }
 
@@ -2275,6 +2276,7 @@ nfp_flow_action_set_hl(char *act_data,
 
 	ttl_conf = (const struct rte_flow_action_set_ttl *)action->conf;
 	tc_hl->ipv6_hop_limit = ttl_conf->ttl_value;
+	tc_hl->ipv6_hop_limit_mask = 0xff;
 	tc_hl->reserved = 0;
 }
 
-- 
2.29.3

[PATCH 6/6] net/nfp: fix the set IPv4/v6 DSCP flow action

[Chaoyong He]

The former logic of set IPv4/v6 DSCP flow action don't consider the
mask filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right IPv4/v6 DSCP as expected.

Fixes: 9c665d70f7a7 ("net/nfp: support IPv4 DSCP flow action")
Fixes: 3202b003bec1 ("net/nfp: support IPv6 DSCP flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index c6e19d8cac..c8c57bc363 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2300,6 +2300,7 @@ nfp_flow_action_set_tos(char *act_data,
 
 	tos_conf = (const struct rte_flow_action_set_dscp *)action->conf;
 	ttl_tos->ipv4_tos = tos_conf->dscp;
+	ttl_tos->ipv4_tos_mask = 0xff;
 	ttl_tos->reserved = 0;
 }
 
@@ -2323,6 +2324,7 @@ nfp_flow_action_set_tc(char *act_data,
 
 	tos_conf = (const struct rte_flow_action_set_dscp *)action->conf;
 	tc_hl->ipv6_tc = tos_conf->dscp;
+	tc_hl->ipv6_tc_mask = 0xff;
 	tc_hl->reserved = 0;
 }
 
-- 
2.29.3

[PATCH v2 0/6] fix the set flow actions

[Chaoyong He]

The former logic of set flow actions don't consider the mask filed
of control message passed to the firmware, and this caused the
firmware skip the set action logic wrongly.
This patch series aims to fix this problem.

--
v2:
* fix the compile warning.
--

Chaoyong He (6):
  net/nfp: fix the set MAC flow action
  net/nfp: fix the set IPv4 flow action
  net/nfp: fix the set IPv6 flow action
  net/nfp: fix the set TP flow action
  net/nfp: fix the set TTL flow action
  net/nfp: fix the set IPv4/v6 DSCP flow action

 drivers/net/nfp/nfp_flow.c | 29 +++++++++++++++++++++++------
 1 file changed, 23 insertions(+), 6 deletions(-)

-- 
2.29.3

[PATCH v2 1/6] net/nfp: fix the set MAC flow action

[Chaoyong He]

The former logic of set MAC source/destination address flow action don't
consider the mask filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right MAC address as expected.

Fixes: 4f6983154570 ("net/nfp: support MAC source flow action")
Fixes: eecc7ca3088a ("net/nfp: support MAC destination flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index 0c38925701..f373171d7e 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2066,6 +2066,7 @@ nfp_flow_action_set_mac(char *act_data,
 		bool mac_src_flag,
 		bool mac_set_flag)
 {
+	uint8_t i;
 	size_t act_size;
 	struct nfp_fl_act_set_eth *set_eth;
 	const struct rte_flow_action_set_mac *set_mac;
@@ -2084,9 +2085,13 @@ nfp_flow_action_set_mac(char *act_data,
 	if (mac_src_flag) {
 		rte_memcpy(&set_eth->eth_addr[RTE_ETHER_ADDR_LEN],
 				set_mac->mac_addr, RTE_ETHER_ADDR_LEN);
+		for (i = 0; i < RTE_ETHER_ADDR_LEN; i++)
+			set_eth->eth_addr_mask[RTE_ETHER_ADDR_LEN + i] = 0xff;
 	} else {
 		rte_memcpy(&set_eth->eth_addr[0],
 				set_mac->mac_addr, RTE_ETHER_ADDR_LEN);
+		for (i = 0; i < RTE_ETHER_ADDR_LEN; i++)
+			set_eth->eth_addr_mask[i] = 0xff;
 	}
 }
 
-- 
2.29.3

[PATCH v2 2/6] net/nfp: fix the set IPv4 flow action

[Chaoyong He]

The former logic of set IPv4 source/destination address flow action don't
consider the mask filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right IPv4 address as expected.

Fixes: ab80975b0a14 ("net/nfp: support IPv4 source flow action")
Fixes: f85c1e05bff3 ("net/nfp: support IPv4 destination flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index f373171d7e..39493ce98d 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2132,10 +2132,13 @@ nfp_flow_action_set_ip(char *act_data,
 	set_ip->reserved     = 0;
 
 	set_ipv4 = (const struct rte_flow_action_set_ipv4 *)action->conf;
-	if (ip_src_flag)
+	if (ip_src_flag) {
 		set_ip->ipv4_src = set_ipv4->ipv4_addr;
-	else
+		set_ip->ipv4_src_mask = RTE_BE32(0xffffffff);
+	} else {
 		set_ip->ipv4_dst = set_ipv4->ipv4_addr;
+		set_ip->ipv4_dst_mask = RTE_BE32(0xffffffff);
+	}
 }
 
 static void
-- 
2.29.3

[PATCH v2 3/6] net/nfp: fix the set IPv6 flow action

[Chaoyong He]

The former logic of set IPv6 source/destination address flow action don't
consider the mask filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right IPv6 address as expected.

Fixes: 596ae2217214 ("net/nfp: support IPv6 source flow action")
Fixes: 51384f79b264 ("net/nfp: support IPv6 destination flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index 39493ce98d..d1fe744758 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2163,8 +2163,10 @@ nfp_flow_action_set_ipv6(char *act_data,
 	set_ip->head.len_lw = act_size >> NFP_FL_LW_SIZ;
 	set_ip->reserved = 0;
 
-	for (i = 0; i < 4; i++)
-		set_ip->ipv6[i].exact = set_ipv6->ipv6_addr[i];
+	for (i = 0; i < 4; i++) {
+		set_ip->ipv6[i].exact = set_ipv6->ipv6_addr[i * 4];
+		set_ip->ipv6[i].mask = RTE_BE32(0xffffffff);
+	}
 }
 
 static void
-- 
2.29.3

[PATCH v2 4/6] net/nfp: fix the set TP flow action

[Chaoyong He]

The former logic of set TP source/destination flow action don't
consider the mask filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right TP as expected.

Fixes: fc185097bbe6 ("net/nfp: support TP source flow action")
Fixes: 87986df09d75 ("net/nfp: support TP destination flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index d1fe744758..5be4c4f18a 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2190,10 +2190,13 @@ nfp_flow_action_set_tp(char *act_data,
 	set_tp->reserved     = 0;
 
 	set_tp_conf = (const struct rte_flow_action_set_tp *)action->conf;
-	if (tp_src_flag)
+	if (tp_src_flag) {
 		set_tp->src_port = set_tp_conf->port;
-	else
+		set_tp->src_port_mask = RTE_BE16(0xffff);
+	} else {
 		set_tp->dst_port = set_tp_conf->port;
+		set_tp->dst_port_mask = RTE_BE16(0xffff);
+	}
 }
 
 static int
-- 
2.29.3

[PATCH v2 5/6] net/nfp: fix the set TTL flow action

[Chaoyong He]

The former logic of set TTL flow action don't consider the mask
filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right TTL as expected.

Fixes: ac12e126c482 ("net/nfp: support TTL flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index 5be4c4f18a..77f249e2b9 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2252,6 +2252,7 @@ nfp_flow_action_set_ttl(char *act_data,
 
 	ttl_conf = (const struct rte_flow_action_set_ttl *)action->conf;
 	ttl_tos->ipv4_ttl = ttl_conf->ttl_value;
+	ttl_tos->ipv4_ttl_mask = 0xff;
 	ttl_tos->reserved = 0;
 }
 
@@ -2275,6 +2276,7 @@ nfp_flow_action_set_hl(char *act_data,
 
 	ttl_conf = (const struct rte_flow_action_set_ttl *)action->conf;
 	tc_hl->ipv6_hop_limit = ttl_conf->ttl_value;
+	tc_hl->ipv6_hop_limit_mask = 0xff;
 	tc_hl->reserved = 0;
 }
 
-- 
2.29.3

[PATCH v2 6/6] net/nfp: fix the set IPv4/v6 DSCP flow action

[Chaoyong He]

The former logic of set IPv4/v6 DSCP flow action don't consider the
mask filed of control message passed to the firmware.
This caused the firmware skip the set action logic, and the offloaded
packets don't have the right IPv4/v6 DSCP as expected.

Fixes: 9c665d70f7a7 ("net/nfp: support IPv4 DSCP flow action")
Fixes: 3202b003bec1 ("net/nfp: support IPv6 DSCP flow action")
Cc: stable@dpdk.org

Signed-off-by: Chaoyong He <chaoyong.he@corigine.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@corigine.com>
---
 drivers/net/nfp/nfp_flow.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/nfp/nfp_flow.c b/drivers/net/nfp/nfp_flow.c
index 77f249e2b9..53029d00c9 100644
--- a/drivers/net/nfp/nfp_flow.c
+++ b/drivers/net/nfp/nfp_flow.c
@@ -2300,6 +2300,7 @@ nfp_flow_action_set_tos(char *act_data,
 
 	tos_conf = (const struct rte_flow_action_set_dscp *)action->conf;
 	ttl_tos->ipv4_tos = tos_conf->dscp;
+	ttl_tos->ipv4_tos_mask = 0xff;
 	ttl_tos->reserved = 0;
 }
 
@@ -2323,6 +2324,7 @@ nfp_flow_action_set_tc(char *act_data,
 
 	tos_conf = (const struct rte_flow_action_set_dscp *)action->conf;
 	tc_hl->ipv6_tc = tos_conf->dscp;
+	tc_hl->ipv6_tc_mask = 0xff;
 	tc_hl->reserved = 0;
 }
 
-- 
2.29.3

Re: [PATCH v2 0/6] fix the set flow actions

[Ferruh Yigit]

On 2/21/2023 10:35 AM, Chaoyong He wrote:
> The former logic of set flow actions don't consider the mask filed
> of control message passed to the firmware, and this caused the
> firmware skip the set action logic wrongly.
> This patch series aims to fix this problem.
> 
> --
> v2:
> * fix the compile warning.
> --
> 
> Chaoyong He (6):
>   net/nfp: fix the set MAC flow action
>   net/nfp: fix the set IPv4 flow action
>   net/nfp: fix the set IPv6 flow action
>   net/nfp: fix the set TP flow action
>   net/nfp: fix the set TTL flow action
>   net/nfp: fix the set IPv4/v6 DSCP flow action


Series applied to dpdk-next-net/main, thanks.