Skip to contentSkip to conversation

Netfilter fixes for net

PATCH Applied Pablo Neira Ayuso 2 participants 7 messages Aug 30, 2023 – Aug 31, 2023

Subsystems (4)
lore
Share & tools
lore

Pablo Neira Ayuso

author maintainer
#0 lore raw

cover Netfilter fixes for net

Hi,

The following patchset contains Netfilter fixes for net:

1) Fix mangling of TCP options with non-linear skbuff, from Xiao Liang.

2) OOB read in xt_sctp due to missing sanitization of array length field. From Wander Lairson Costa.

3) OOB read in xt_u32 due to missing sanitization of array length field. Also from Wander Lairson Costa.

All of them above, always broken for several releases.

4) Missing audit log for set element reset command, from Phil Sutter.

5) Missing audit log for rule reset command, also from Phil.

These audit log support are missing in 6.5.

Please, pull these changes from:

git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf.git nf-23-08-31

Thanks.

----------------------------------------------------------------

The following changes since commit bd6c11bc43c496cddfc6cf603b5d45365606dbd5:

Merge tag 'net-next-6.6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next (2023-08-29 11:33:01 -0700)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf.git tags/nf-23-08-31

for you to fetch changes up to ea078ae9108e25fc881c84369f7c03931d22e555:

netfilter: nf_tables: Audit log rule reset (2023-08-31 01:29:28 +0200)

----------------------------------------------------------------
netfilter pull request 23-08-31

----------------------------------------------------------------
Phil Sutter (2):
netfilter: nf_tables: Audit log setelem reset
netfilter: nf_tables: Audit log rule reset

Wander Lairson Costa (2):
netfilter: xt_sctp: validate the flag_info count
netfilter: xt_u32: validate user space input

Xiao Liang (1):
netfilter: nft_exthdr: Fix non-linear header modification

 include/linux/audit.h         |  2 ++
 kernel/auditsc.c              |  2 ++
 net/netfilter/nf_tables_api.c | 49 ++++++++++++++++++++++++++++++++++++++++---
 net/netfilter/nft_exthdr.c    | 20 +++++++-----------
 net/netfilter/xt_sctp.c       |  2 ++
 net/netfilter/xt_u32.c        | 21 +++++++++++++++++++
 6 files changed, 81 insertions(+), 15 deletions(-)
author maintainer
#1 lore raw

PATCH 1/5 netfilter: nft_exthdr: Fix non-linear header modification

From: Xiao Liang <[email protected]>

Fix skb_ensure_writable() size. Don't use nft_tcp_header_pointer() to make it explicit that pointers point to the packet (not local buffer).

Fixes 99d1712bc41c ("netfilter: exthdr: tcp option set support")
Fixes 7890cbea66e7 ("netfilter: exthdr: add support for tcp option removal")
---
 net/netfilter/nft_exthdr.c | 20 ++++++++------------
 1 file changed, 8 insertions(+), 12 deletions(-)
net/netfilter/nft_exthdr.c +8 -12
index 7f856ceb3a66..a9844eefedeb 100644
@@ -238,7 +238,12 @@ static void nft_exthdr_tcp_set_eval(const struct nft_expr *expr,
 	if (!tcph)
 		goto err;
 
+	if (skb_ensure_writable(pkt->skb, nft_thoff(pkt) + tcphdr_len))
+		goto err;
+
+	tcph = (struct tcphdr *)(pkt->skb->data + nft_thoff(pkt));
 	opt = (u8 *)tcph;
+
 	for (i = sizeof(*tcph); i < tcphdr_len - 1; i += optl) {
 		union {
 			__be16 v16;
@@ -253,15 +258,6 @@ static void nft_exthdr_tcp_set_eval(const struct nft_expr *expr,
 		if (i + optl > tcphdr_len || priv->len + priv->offset > optl)
 			goto err;
 
-		if (skb_ensure_writable(pkt->skb,
-					nft_thoff(pkt) + i + priv->len))
-			goto err;
-
-		tcph = nft_tcp_header_pointer(pkt, sizeof(buff), buff,
-					      &tcphdr_len);
-		if (!tcph)
-			goto err;
-
 		offset = i + priv->offset;
 
 		switch (priv->len) {
@@ -325,9 +321,9 @@ static void nft_exthdr_tcp_strip_eval(const struct nft_expr *expr,
 	if (skb_ensure_writable(pkt->skb, nft_thoff(pkt) + tcphdr_len))
 		goto drop;
 
-	opt = (u8 *)nft_tcp_header_pointer(pkt, sizeof(buff), buff, &tcphdr_len);
-	if (!opt)
-		goto err;
+	tcph = (struct tcphdr *)(pkt->skb->data + nft_thoff(pkt));
+	opt = (u8 *)tcph;
+
 	for (i = sizeof(*tcph); i < tcphdr_len - 1; i += optl) {
 		unsigned int j;
 
Signature · 1 line
2.30.2
#2 lore raw

Re: [PATCH net 1/5] netfilter: nft_exthdr: Fix non-linear header modification

This series was applied to netdev/net.git (main) by Pablo Neira Ayuso <[email protected]>:

Hello:

This series was applied to netdev/net.git (main)
by Pablo Neira Ayuso <[email protected]>:

Quoted 5 lines · On Thu, 31 Aug 2023 01:59:31 +0200 you wrote:

From: Xiao Liang <[email protected]>

Fix skb_ensure_writable() size. Don't use nft_tcp_header_pointer() to make it explicit that pointers point to the packet (not local buffer).

Quoted 7 lines
Fixes 99d1712bc41c ("netfilter: exthdr: tcp option set support")
Fixes 7890cbea66e7 ("netfilter: exthdr: add support for tcp option removal")

[...]

Here is the summary with links:
- [net,1/5] netfilter: nft_exthdr: Fix non-linear header modification

    https://git.kernel.org/netdev/net/c/28427f368f0e
  - [net,2/5] netfilter: xt_sctp: validate the flag_info count
    https://git.kernel.org/netdev/net/c/e99476497687
  - [net,3/5] netfilter: xt_u32: validate user space input
    https://git.kernel.org/netdev/net/c/69c5d284f670
  - [net,4/5] netfilter: nf_tables: Audit log setelem reset
    https://git.kernel.org/netdev/net/c/7e9be1124dbe
  - [net,5/5] netfilter: nf_tables: Audit log rule reset
    https://git.kernel.org/netdev/net/c/ea078ae9108e

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
author maintainer
#3 lore raw

PATCH 2/5 netfilter: xt_sctp: validate the flag_info count

In reply to Pablo Neira Ayuso

From: Wander Lairson Costa <[email protected]>

sctp_mt_check doesn't validate the flag_count field. An attacker can take advantage of that to trigger a OOB read and leak memory information.

Add the field validation in the checkentry function.

Fixes 2e4e6a17af35 ("[NETFILTER] x_tables: Abstraction layer for {ip,ip6,arp}_tables")
---
 net/netfilter/xt_sctp.c | 2 ++
 1 file changed, 2 insertions(+)
net/netfilter/xt_sctp.c +2 -0
index e8961094a282..b46a6a512058 100644
@@ -149,6 +149,8 @@ static int sctp_mt_check(const struct xt_mtchk_param *par)
 {
 	const struct xt_sctp_info *info = par->matchinfo;
 
+	if (info->flag_count > ARRAY_SIZE(info->flag_info))
+		return -EINVAL;
 	if (info->flags & ~XT_SCTP_VALID_FLAGS)
 		return -EINVAL;
 	if (info->invflags & ~XT_SCTP_VALID_FLAGS)
Signature · 1 line
2.30.2
author maintainer
#4 lore raw

PATCH 3/5 netfilter: xt_u32: validate user space input

In reply to Pablo Neira Ayuso

From: Wander Lairson Costa <[email protected]>

The xt_u32 module doesn't validate the fields in the xt_u32 structure. An attacker may take advantage of this to trigger an OOB read by setting the size fields with a value beyond the arrays boundaries.

Add a checkentry function to validate the structure.

This was originally reported by the ZDI project (ZDI-CAN-18408).

Fixes 1b50b8a371e9 ("[NETFILTER]: Add u32 match")
---
 net/netfilter/xt_u32.c | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)
net/netfilter/xt_u32.c +21 -0
index 177b40d08098..117d4615d668 100644
@@ -96,11 +96,32 @@ static bool u32_mt(const struct sk_buff *skb, struct xt_action_param *par)
 	return ret ^ data->invert;
 }
 
+static int u32_mt_checkentry(const struct xt_mtchk_param *par)
+{
+	const struct xt_u32 *data = par->matchinfo;
+	const struct xt_u32_test *ct;
+	unsigned int i;
+
+	if (data->ntests > ARRAY_SIZE(data->tests))
+		return -EINVAL;
+
+	for (i = 0; i < data->ntests; ++i) {
+		ct = &data->tests[i];
+
+		if (ct->nnums > ARRAY_SIZE(ct->location) ||
+		    ct->nvalues > ARRAY_SIZE(ct->value))
+			return -EINVAL;
+	}
+
+	return 0;
+}
+
 static struct xt_match xt_u32_mt_reg __read_mostly = {
 	.name       = "u32",
 	.revision   = 0,
 	.family     = NFPROTO_UNSPEC,
 	.match      = u32_mt,
+	.checkentry = u32_mt_checkentry,
 	.matchsize  = sizeof(struct xt_u32),
 	.me         = THIS_MODULE,
 };
Signature · 1 line
2.30.2
author maintainer
#5 lore raw

PATCH 4/5 netfilter: nf_tables: Audit log setelem reset

In reply to Pablo Neira Ayuso

From: Phil Sutter <[email protected]>

Since set element reset is not integrated into nf_tables' transaction logic, an explicit log call is needed, similar to NFT_MSG_GETOBJ_RESET handling.

For the sake of simplicity, catchall element reset will always generate a dedicated log entry. This relieves nf_tables_dump_set() from having to adjust the logged element count depending on whether a catchall element was found or not.

Fixes 079cd633219d7 ("netfilter: nf_tables: Introduce NFT_MSG_GETSETELEM_RESET")
---
 include/linux/audit.h         |  1 +
 kernel/auditsc.c              |  1 +
 net/netfilter/nf_tables_api.c | 31 ++++++++++++++++++++++++++++---
 3 files changed, 30 insertions(+), 3 deletions(-)
include/linux/audit.h +1 -0
index 6a3a9e122bb5..192bf03aacc5 100644
@@ -117,6 +117,7 @@ enum audit_nfcfgop {
 	AUDIT_NFT_OP_OBJ_RESET,
 	AUDIT_NFT_OP_FLOWTABLE_REGISTER,
 	AUDIT_NFT_OP_FLOWTABLE_UNREGISTER,
+	AUDIT_NFT_OP_SETELEM_RESET,
 	AUDIT_NFT_OP_INVALID,
 };
 
kernel/auditsc.c +1 -0
index addeed3df15d..38481e318197 100644
@@ -143,6 +143,7 @@ static const struct audit_nfcfgop_tab audit_nfcfgs[] = {
 	{ AUDIT_NFT_OP_OBJ_RESET,		"nft_reset_obj"		   },
 	{ AUDIT_NFT_OP_FLOWTABLE_REGISTER,	"nft_register_flowtable"   },
 	{ AUDIT_NFT_OP_FLOWTABLE_UNREGISTER,	"nft_unregister_flowtable" },
+	{ AUDIT_NFT_OP_SETELEM_RESET,		"nft_reset_setelem"        },
 	{ AUDIT_NFT_OP_INVALID,			"nft_invalid"		   },
 };
 
net/netfilter/nf_tables_api.c +28 -3
index 41b826dff6f5..361e98e71692 100644
@@ -102,6 +102,7 @@ static const u8 nft2audit_op[NFT_MSG_MAX] = { // enum nf_tables_msg_types
 	[NFT_MSG_NEWFLOWTABLE]	= AUDIT_NFT_OP_FLOWTABLE_REGISTER,
 	[NFT_MSG_GETFLOWTABLE]	= AUDIT_NFT_OP_INVALID,
 	[NFT_MSG_DELFLOWTABLE]	= AUDIT_NFT_OP_FLOWTABLE_UNREGISTER,
+	[NFT_MSG_GETSETELEM_RESET] = AUDIT_NFT_OP_SETELEM_RESET,
 };
 
 static void nft_validate_state_update(struct nft_table *table, u8 new_validate_state)
@@ -5624,13 +5625,25 @@ static int nf_tables_dump_setelem(const struct nft_ctx *ctx,
 	return nf_tables_fill_setelem(args->skb, set, elem, args->reset);
 }
 
+static void audit_log_nft_set_reset(const struct nft_table *table,
+				    unsigned int base_seq,
+				    unsigned int nentries)
+{
+	char *buf = kasprintf(GFP_ATOMIC, "%s:%u", table->name, base_seq);
+
+	audit_log_nfcfg(buf, table->family, nentries,
+			AUDIT_NFT_OP_SETELEM_RESET, GFP_ATOMIC);
+	kfree(buf);
+}
+
 struct nft_set_dump_ctx {
 	const struct nft_set	*set;
 	struct nft_ctx		ctx;
 };
 
 static int nft_set_catchall_dump(struct net *net, struct sk_buff *skb,
-				 const struct nft_set *set, bool reset)
+				 const struct nft_set *set, bool reset,
+				 unsigned int base_seq)
 {
 	struct nft_set_elem_catchall *catchall;
 	u8 genmask = nft_genmask_cur(net);
@@ -5646,6 +5659,8 @@ static int nft_set_catchall_dump(struct net *net, struct sk_buff *skb,
 
 		elem.priv = catchall->elem;
 		ret = nf_tables_fill_setelem(skb, set, &elem, reset);
+		if (reset && !ret)
+			audit_log_nft_set_reset(set->table, base_seq, 1);
 		break;
 	}
 
@@ -5725,12 +5740,17 @@ static int nf_tables_dump_set(struct sk_buff *skb, struct netlink_callback *cb)
 	set->ops->walk(&dump_ctx->ctx, set, &args.iter);
 
 	if (!args.iter.err && args.iter.count == cb->args[0])
-		args.iter.err = nft_set_catchall_dump(net, skb, set, reset);
+		args.iter.err = nft_set_catchall_dump(net, skb, set,
+						      reset, cb->seq);
 	rcu_read_unlock();
 
 	nla_nest_end(skb, nest);
 	nlmsg_end(skb, nlh);
 
+	if (reset && args.iter.count > args.iter.skip)
+		audit_log_nft_set_reset(table, cb->seq,
+					args.iter.count - args.iter.skip);
+
 	if (args.iter.err && args.iter.err != -EMSGSIZE)
 		return args.iter.err;
 	if (args.iter.count == cb->args[0])
@@ -5955,13 +5975,13 @@ static int nf_tables_getsetelem(struct sk_buff *skb,
 	struct netlink_ext_ack *extack = info->extack;
 	u8 genmask = nft_genmask_cur(info->net);
 	u8 family = info->nfmsg->nfgen_family;
+	int rem, err = 0, nelems = 0;
 	struct net *net = info->net;
 	struct nft_table *table;
 	struct nft_set *set;
 	struct nlattr *attr;
 	struct nft_ctx ctx;
 	bool reset = false;
-	int rem, err = 0;
 
 	table = nft_table_lookup(net, nla[NFTA_SET_ELEM_LIST_TABLE], family,
 				 genmask, 0);
@@ -6004,8 +6024,13 @@ static int nf_tables_getsetelem(struct sk_buff *skb,
 			NL_SET_BAD_ATTR(extack, attr);
 			break;
 		}
+		nelems++;
 	}
 
+	if (reset)
+		audit_log_nft_set_reset(table, nft_pernet(net)->base_seq,
+					nelems);
+
 	return err;
 }
 
Signature · 1 line
2.30.2
author maintainer
#6 lore raw

PATCH 5/5 netfilter: nf_tables: Audit log rule reset

In reply to Pablo Neira Ayuso

From: Phil Sutter <[email protected]>

Resetting rules' stateful data happens outside of the transaction logic, so 'get' and 'dump' handlers have to emit audit log entries themselves.

Fixes 8daa8fde3fc3f ("netfilter: nf_tables: Introduce NFT_MSG_GETRULE_RESET")
---
 include/linux/audit.h         |  1 +
 kernel/auditsc.c              |  1 +
 net/netfilter/nf_tables_api.c | 18 ++++++++++++++++++
 3 files changed, 20 insertions(+)
include/linux/audit.h +1 -0
index 192bf03aacc5..51b1b7054a23 100644
@@ -118,6 +118,7 @@ enum audit_nfcfgop {
 	AUDIT_NFT_OP_FLOWTABLE_REGISTER,
 	AUDIT_NFT_OP_FLOWTABLE_UNREGISTER,
 	AUDIT_NFT_OP_SETELEM_RESET,
+	AUDIT_NFT_OP_RULE_RESET,
 	AUDIT_NFT_OP_INVALID,
 };
 
kernel/auditsc.c +1 -0
index 38481e318197..fc0c7c03eeab 100644
@@ -144,6 +144,7 @@ static const struct audit_nfcfgop_tab audit_nfcfgs[] = {
 	{ AUDIT_NFT_OP_FLOWTABLE_REGISTER,	"nft_register_flowtable"   },
 	{ AUDIT_NFT_OP_FLOWTABLE_UNREGISTER,	"nft_unregister_flowtable" },
 	{ AUDIT_NFT_OP_SETELEM_RESET,		"nft_reset_setelem"        },
+	{ AUDIT_NFT_OP_RULE_RESET,		"nft_reset_rule"           },
 	{ AUDIT_NFT_OP_INVALID,			"nft_invalid"		   },
 };
 
net/netfilter/nf_tables_api.c +18 -0
index 361e98e71692..2c81cee858d6 100644
@@ -3422,6 +3422,18 @@ static void nf_tables_rule_notify(const struct nft_ctx *ctx,
 	nfnetlink_set_err(ctx->net, ctx->portid, NFNLGRP_NFTABLES, -ENOBUFS);
 }
 
+static void audit_log_rule_reset(const struct nft_table *table,
+				 unsigned int base_seq,
+				 unsigned int nentries)
+{
+	char *buf = kasprintf(GFP_ATOMIC, "%s:%u",
+			      table->name, base_seq);
+
+	audit_log_nfcfg(buf, table->family, nentries,
+			AUDIT_NFT_OP_RULE_RESET, GFP_ATOMIC);
+	kfree(buf);
+}
+
 struct nft_rule_dump_ctx {
 	char *table;
 	char *chain;
@@ -3528,6 +3540,9 @@ static int nf_tables_dump_rules(struct sk_buff *skb,
 done:
 	rcu_read_unlock();
 
+	if (reset && idx > cb->args[0])
+		audit_log_rule_reset(table, cb->seq, idx - cb->args[0]);
+
 	cb->args[0] = idx;
 	return skb->len;
 }
@@ -3635,6 +3650,9 @@ static int nf_tables_getrule(struct sk_buff *skb, const struct nfnl_info *info,
 	if (err < 0)
 		goto err_fill_rule_info;
 
+	if (reset)
+		audit_log_rule_reset(table, nft_pernet(net)->base_seq, 1);
+
 	return nfnetlink_unicast(skb2, net, NETLINK_CB(skb).portid);
 
 err_fill_rule_info:
Signature · 1 line
2.30.2