Skip to contentSkip to conversation

net/mlx5: fix IPv6 next proto validation

PATCH New Dekel Peled 5 participants 7 messages Oct 20, 2020 – Oct 22, 2020 Network drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net/mlx5: fix IPv6 next proto validation

Previous patch added validation of the IPv6 next proto field, in order to overcome a known limitation.
One of the values checked is IPPROTO_HOPOPTS, which is currently defined as value 0.
The same value 0 is received when next proto field is not specified for matching, or has mask 0.

This patch updates the validation, to make sure next proto is not 0 before validating it.

Fixes 55e4c1d1ba73 ("net/mlx5: enforce limitation on IPv6 next proto")
---
 drivers/net/mlx5/mlx5_flow.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)
drivers/net/mlx5/mlx5_flow.c +7 -6
index c56dac8..74af207 100644
@@ -1987,12 +1987,13 @@ struct mlx5_flow_tunnel_info {
 						  "multiple tunnel "
 						  "not supported");
 	}
-	if (next_proto == IPPROTO_HOPOPTS  ||
-	    next_proto == IPPROTO_ROUTING  ||
-	    next_proto == IPPROTO_FRAGMENT ||
-	    next_proto == IPPROTO_ESP	   ||
-	    next_proto == IPPROTO_AH	   ||
-	    next_proto == IPPROTO_DSTOPTS)
+	if (next_proto &&
+	    (next_proto == IPPROTO_HOPOPTS  ||
+	     next_proto == IPPROTO_ROUTING  ||
+	     next_proto == IPPROTO_FRAGMENT ||
+	     next_proto == IPPROTO_ESP      ||
+	     next_proto == IPPROTO_AH	    ||
+	     next_proto == IPPROTO_DSTOPTS))
 		return rte_flow_error_set(error, EINVAL,
 					  RTE_FLOW_ERROR_TYPE_ITEM, item,
 					  "IPv6 proto (next header) should "
Signature · 1 line
1.8.3.1
#1 lore raw

PATCH v2 1/1 net/mlx5: fix IPv6 next proto validation

Previous patch added validation of the IPv6 next proto field, in order to overcome a known limitation.
One of the values checked is IPPROTO_HOPOPTS, which is defined as 0. If proto field is not specified for matching, or mask=0, as in the following, a wrong validation takes place.
flow create 0 ingress pattern eth / ipv6 has_frag_ext is 0 / end actions drop / end
Fix the validation only on proto asked by the user.

Fixes 55e4c1d1ba73 ("net/mlx5: enforce limitation on IPv6 next proto")
---
 drivers/net/mlx5/mlx5_flow.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/mlx5/mlx5_flow.c +1 -1
index c56dac89f9..e0ca8e06b1 100644
@@ -1977,7 +1977,7 @@ mlx5_flow_validate_item_ipv6(const struct rte_flow_item *item,
 					  RTE_FLOW_ERROR_TYPE_ITEM, item,
 					  "IPv6 cannot follow L2/VLAN layer "
 					  "which ether type is not IPv6");
-	if (mask && spec)
+	if (mask && mask->hdr.proto && spec)
 		next_proto = mask->hdr.proto & spec->hdr.proto;
 	if (item_flags & MLX5_FLOW_LAYER_IPV6_ENCAP) {
 		if (next_proto == IPPROTO_IPIP || next_proto == IPPROTO_IPV6)
Signature · 1 line
2.28.0.546.g385c171
#2 lore raw

++

Quoted 30 lines · Eli Britstein

Previous patch added validation of the IPv6 next proto field, in order to overcome a known limitation.
One of the values checked is IPPROTO_HOPOPTS, which is defined as 0. If proto field is not specified for matching, or mask=0, as in the following, a wrong validation takes place.
flow create 0 ingress pattern eth / ipv6 has_frag_ext is 0 / end actions drop / end
Fix the validation only on proto asked by the user.

Fixes 55e4c1d1ba73 ("net/mlx5: enforce limitation on IPv6 next proto")
---
  drivers/net/mlx5/mlx5_flow.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/mlx5/mlx5_flow.c +1 -1
index c56dac89f9..e0ca8e06b1 100644
@@ -1977,7 +1977,7 @@ mlx5_flow_validate_item_ipv6(const struct rte_flow_item *item,
  					  RTE_FLOW_ERROR_TYPE_ITEM, item,
  					  "IPv6 cannot follow L2/VLAN layer "
  					  "which ether type is not IPv6");
-	if (mask && spec)
+	if (mask && mask->hdr.proto && spec)
  		next_proto = mask->hdr.proto & spec->hdr.proto;
  	if (item_flags & MLX5_FLOW_LAYER_IPV6_ENCAP) {
  		if (next_proto == IPPROTO_IPIP || next_proto == IPPROTO_IPV6)
#3 lore raw
In reply to Eli Britstein

From: Eli Britstein

Quoted 21 lines

Previous patch added validation of the IPv6 next proto field, in order to overcome a known limitation.
One of the values checked is IPPROTO_HOPOPTS, which is defined as 0. If proto field is not specified for matching, or mask=0, as in the following, a wrong validation takes place.
flow create 0 ingress pattern eth / ipv6 has_frag_ext is 0 / end actions drop / end Fix the validation only on proto asked by the user.

Fixes 55e4c1d1ba73 ("net/mlx5: enforce limitation on IPv6 next proto")
---
 drivers/net/mlx5/mlx5_flow.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/mlx5/mlx5_flow.c +0 -0
index c56dac89f9..e0ca8e06b1 100644

drivers/net/mlx5/mlx5_flow.c mlx5_flow_validate_item_ipv6(const struct · line 1982

Quoted 7 lines
drivers/net/mlx5/mlx5_flow.c +1 -1
@@ -1977,7 +1977,7 @@ mlx5_flow_validate_item_ipv6(const struct
 rte_flow_item *item,
 					  RTE_FLOW_ERROR_TYPE_ITEM,
 item,
 					  "IPv6 cannot follow L2/VLAN layer "
 					  "which ether type is not IPv6");
-	if (mask && spec)
+	if (mask && mask->hdr.proto && spec)

If you want to be sure os specific proto only you need to check mask->hdr.proto==0xFF, no?

Quoted 6 lines
drivers/net/mlx5/mlx5_flow.c +0 -0
 		next_proto = mask->hdr.proto & spec->hdr.proto;
drivers/net/mlx5/mlx5_flow.c +0 -0
 	if (item_flags & MLX5_FLOW_LAYER_IPV6_ENCAP) {
 		if (next_proto == IPPROTO_IPIP || next_proto ==

IPPROTO_IPV6)
--
2.28.0.546.g385c171

#4 lore raw

PATCH v3 1/1 net/mlx5: fix IPv6 next proto validation

Previous patch added validation of the IPv6 next proto field, in order to overcome a known limitation.
One of the values checked is IPPROTO_HOPOPTS, which is defined as 0. If proto field is not specified for matching, or mask=0, as in the following, a wrong validation takes place.
flow create 0 ingress pattern eth / ipv6 has_frag_ext is 0 / end actions drop / end
Fix the validation only on proto asked by the user.

Fixes 92be60e1b541 ("net/mlx5: enforce limitation on IPv6 next proto")
---
 drivers/net/mlx5/mlx5_flow.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/mlx5/mlx5_flow.c +2 -2
index 22fb4ee60a..d7243a878b 100644
@@ -1979,8 +1979,8 @@ mlx5_flow_validate_item_ipv6(const struct rte_flow_item *item,
 					  RTE_FLOW_ERROR_TYPE_ITEM, item,
 					  "IPv6 cannot follow L2/VLAN layer "
 					  "which ether type is not IPv6");
-	if (mask && spec)
-		next_proto = mask->hdr.proto & spec->hdr.proto;
+	if (mask && mask->hdr.proto == UINT8_MAX && spec)
+		next_proto = spec->hdr.proto;
 	if (item_flags & MLX5_FLOW_LAYER_IPV6_ENCAP) {
 		if (next_proto == IPPROTO_IPIP || next_proto == IPPROTO_IPV6)
 			return rte_flow_error_set(error, EINVAL,
Signature · 1 line
2.28.0.546.g385c171
#5 lore raw

Hi,

Quoted 43 lines

-----Original Message-----
From: Eli Britstein <[email protected]>
Sent: Thursday, October 22, 2020 11:26 AM
To: [email protected]
Cc: Dekel Peled <[email protected]>; Ori Kam <[email protected]>; Matan Azrad <[email protected]>; Raslan Darawsheh <[email protected]>; Slava Ovsiienko <[email protected]>; Shahaf Shuler <[email protected]>; Asaf Penso <[email protected]>; Eli Britstein <[email protected]>
Subject: [PATCH V3 1/1] net/mlx5: fix IPv6 next proto validation

Previous patch added validation of the IPv6 next proto field, in order to overcome a known limitation.
One of the values checked is IPPROTO_HOPOPTS, which is defined as 0. If proto field is not specified for matching, or mask=0, as in the following, a wrong validation takes place.
flow create 0 ingress pattern eth / ipv6 has_frag_ext is 0 / end actions drop / end
Fix the validation only on proto asked by the user.

Fixes 92be60e1b541 ("net/mlx5: enforce limitation on IPv6 next proto")
---
 drivers/net/mlx5/mlx5_flow.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/mlx5/mlx5_flow.c +2 -2
index 22fb4ee60a..d7243a878b 100644
@@ -1979,8 +1979,8 @@ mlx5_flow_validate_item_ipv6(const struct
 rte_flow_item *item,
 					  RTE_FLOW_ERROR_TYPE_ITEM,
 item,
 					  "IPv6 cannot follow L2/VLAN layer "
 					  "which ether type is not IPv6");
-	if (mask && spec)
-		next_proto = mask->hdr.proto & spec->hdr.proto;
+	if (mask && mask->hdr.proto == UINT8_MAX && spec)
+		next_proto = spec->hdr.proto;
 	if (item_flags & MLX5_FLOW_LAYER_IPV6_ENCAP) {
Quoted 5 lines

if (next_proto == IPPROTO_IPIP || next_proto ==
IPPROTO_IPV6)
return rte_flow_error_set(error, EINVAL,
--
2.28.0.546.g385c171

Patch applied to next-net-mlx,

Kindest regards,
Raslan Darawsheh

#6 lore raw
Quoted 21 lines · Raslan Darawsheh

Hi,

Quoted 19 lines

-----Original Message-----
From: Eli Britstein <[email protected]>
Sent: Thursday, October 22, 2020 11:26 AM
To: [email protected]
Cc: Dekel Peled <[email protected]>; Ori Kam <[email protected]>; Matan Azrad <[email protected]>; Raslan Darawsheh <[email protected]>; Slava Ovsiienko <[email protected]>; Shahaf Shuler <[email protected]>; Asaf Penso <[email protected]>; Eli Britstein <[email protected]>
Subject: [PATCH V3 1/1] net/mlx5: fix IPv6 next proto validation

Previous patch added validation of the IPv6 next proto field, in order to overcome a known limitation.
One of the values checked is IPPROTO_HOPOPTS, which is defined as 0. If proto field is not specified for matching, or mask=0, as in the following, a wrong validation takes place.
flow create 0 ingress pattern eth / ipv6 has_frag_ext is 0 / end actions drop / end
Fix the validation only on proto asked by the user.

Quoted 9 lines
Quoted 5 lines
Fixes 92be60e1b541 ("net/mlx5: enforce limitation on IPv6 next proto")

Patch applied to next-net-mlx,

Squashed into relevant commit in next-net, thanks.