Skip to contentSkip to conversation

Add locking for NFT_MSG_GETOBJ_RESET requests

PATCH In review Phil Sutter 2 participants 5 messages Aug 9, 2024 – Aug 11, 2024

Subsystems (2)
lore
Share & tools
lore

Phil Sutter

author reviewer
#0 lore raw

v4 cover Add locking for NFT_MSG_GETOBJ_RESET requests

Analogous to getrule reset locking, prevent concurrent resetting of named objects' state.

Changes since v3:
- Applied patches onto current nf.git

Phil Sutter (3):
netfilter: nf_tables: Audit log dump reset after the fact
netfilter: nf_tables: Introduce nf_tables_getobj_single
netfilter: nf_tables: Add locking for NFT_MSG_GETOBJ_RESET requests

 net/netfilter/nf_tables_api.c | 147 +++++++++++++++++++++++-----------
 1 file changed, 102 insertions(+), 45 deletions(-)
Signature · 1 line
2.43.0
author reviewer
#1 lore raw

PATCH v4 1/3 netfilter: nf_tables: Audit log dump reset after the fact

In theory, dumpreset may fail and invalidate the preceeding log message. Fix this and use the occasion to prepare for object reset locking, which benefits from a few unrelated changes:

* Add an early call to nfnetlink_unicast if not resetting which effectively skips the audit logging but also unindents it.
* Extract the table's name from the netlink attribute (which is verified via earlier table lookup) to not rely upon validity of the looked up table pointer.
* Do not use local variable family, it will vanish.

---
 net/netfilter/nf_tables_api.c | 28 +++++++++++++---------------
 1 file changed, 13 insertions(+), 15 deletions(-)
net/netfilter/nf_tables_api.c +13 -15
index 481ee78e77bc..4fa132715fcc 100644
@@ -8055,6 +8055,7 @@ static int nf_tables_dump_obj_done(struct netlink_callback *cb)
 static int nf_tables_getobj(struct sk_buff *skb, const struct nfnl_info *info,
 			    const struct nlattr * const nla[])
 {
+	const struct nftables_pernet *nft_net = nft_pernet(info->net);
 	struct netlink_ext_ack *extack = info->extack;
 	u8 genmask = nft_genmask_cur(info->net);
 	u8 family = info->nfmsg->nfgen_family;
@@ -8064,6 +8065,7 @@ static int nf_tables_getobj(struct sk_buff *skb, const struct nfnl_info *info,
 	struct sk_buff *skb2;
 	bool reset = false;
 	u32 objtype;
+	char *buf;
 	int err;
 
 	if (info->nlh->nlmsg_flags & NLM_F_DUMP) {
@@ -8102,27 +8104,23 @@ static int nf_tables_getobj(struct sk_buff *skb, const struct nfnl_info *info,
 	if (NFNL_MSG_TYPE(info->nlh->nlmsg_type) == NFT_MSG_GETOBJ_RESET)
 		reset = true;
 
-	if (reset) {
-		const struct nftables_pernet *nft_net;
-		char *buf;
-
-		nft_net = nft_pernet(net);
-		buf = kasprintf(GFP_ATOMIC, "%s:%u", table->name, nft_net->base_seq);
-
-		audit_log_nfcfg(buf,
-				family,
-				1,
-				AUDIT_NFT_OP_OBJ_RESET,
-				GFP_ATOMIC);
-		kfree(buf);
-	}
-
 	err = nf_tables_fill_obj_info(skb2, net, NETLINK_CB(skb).portid,
 				      info->nlh->nlmsg_seq, NFT_MSG_NEWOBJ, 0,
 				      family, table, obj, reset);
 	if (err < 0)
 		goto err_fill_obj_info;
 
+	if (!reset)
+		return nfnetlink_unicast(skb2, net, NETLINK_CB(skb).portid);
+
+	buf = kasprintf(GFP_ATOMIC, "%.*s:%u",
+			nla_len(nla[NFTA_OBJ_TABLE]),
+			(char *)nla_data(nla[NFTA_OBJ_TABLE]),
+			nft_net->base_seq);
+	audit_log_nfcfg(buf, info->nfmsg->nfgen_family, 1,
+			AUDIT_NFT_OP_OBJ_RESET, GFP_ATOMIC);
+	kfree(buf);
+
 	return nfnetlink_unicast(skb2, net, NETLINK_CB(skb).portid);
 
 err_fill_obj_info:
Signature · 1 line
2.43.0
author reviewer
#2 lore raw

PATCH v4 2/3 netfilter: nf_tables: Introduce nf_tables_getobj_single

In reply to Phil Sutter

Outsource the reply skb preparation for non-dump getrule requests into a distinct function. Prep work for object reset locking.

---
 net/netfilter/nf_tables_api.c | 75 ++++++++++++++++++++---------------
 1 file changed, 44 insertions(+), 31 deletions(-)
net/netfilter/nf_tables_api.c +44 -31
index 4fa132715fcc..c12c9cae784d 100644
@@ -8052,10 +8052,10 @@ static int nf_tables_dump_obj_done(struct netlink_callback *cb)
 }
 
 /* called with rcu_read_lock held */
-static int nf_tables_getobj(struct sk_buff *skb, const struct nfnl_info *info,
-			    const struct nlattr * const nla[])
+static struct sk_buff *
+nf_tables_getobj_single(u32 portid, const struct nfnl_info *info,
+			const struct nlattr * const nla[], bool reset)
 {
-	const struct nftables_pernet *nft_net = nft_pernet(info->net);
 	struct netlink_ext_ack *extack = info->extack;
 	u8 genmask = nft_genmask_cur(info->net);
 	u8 family = info->nfmsg->nfgen_family;
@@ -8063,52 +8063,69 @@ static int nf_tables_getobj(struct sk_buff *skb, const struct nfnl_info *info,
 	struct net *net = info->net;
 	struct nft_object *obj;
 	struct sk_buff *skb2;
-	bool reset = false;
 	u32 objtype;
-	char *buf;
 	int err;
 
-	if (info->nlh->nlmsg_flags & NLM_F_DUMP) {
-		struct netlink_dump_control c = {
-			.start = nf_tables_dump_obj_start,
-			.dump = nf_tables_dump_obj,
-			.done = nf_tables_dump_obj_done,
-			.module = THIS_MODULE,
-			.data = (void *)nla,
-		};
-
-		return nft_netlink_dump_start_rcu(info->sk, skb, info->nlh, &c);
-	}
-
 	if (!nla[NFTA_OBJ_NAME] ||
 	    !nla[NFTA_OBJ_TYPE])
-		return -EINVAL;
+		return ERR_PTR(-EINVAL);
 
 	table = nft_table_lookup(net, nla[NFTA_OBJ_TABLE], family, genmask, 0);
 	if (IS_ERR(table)) {
 		NL_SET_BAD_ATTR(extack, nla[NFTA_OBJ_TABLE]);
-		return PTR_ERR(table);
+		return ERR_CAST(table);
 	}
 
 	objtype = ntohl(nla_get_be32(nla[NFTA_OBJ_TYPE]));
 	obj = nft_obj_lookup(net, table, nla[NFTA_OBJ_NAME], objtype, genmask);
 	if (IS_ERR(obj)) {
 		NL_SET_BAD_ATTR(extack, nla[NFTA_OBJ_NAME]);
-		return PTR_ERR(obj);
+		return ERR_CAST(obj);
 	}
 
 	skb2 = alloc_skb(NLMSG_GOODSIZE, GFP_ATOMIC);
 	if (!skb2)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
+
+	err = nf_tables_fill_obj_info(skb2, net, portid,
+				      info->nlh->nlmsg_seq, NFT_MSG_NEWOBJ, 0,
+				      family, table, obj, reset);
+	if (err < 0) {
+		kfree_skb(skb2);
+		return ERR_PTR(err);
+	}
+
+	return skb2;
+}
+
+static int nf_tables_getobj(struct sk_buff *skb, const struct nfnl_info *info,
+			    const struct nlattr * const nla[])
+{
+	struct nftables_pernet *nft_net = nft_pernet(info->net);
+	u32 portid = NETLINK_CB(skb).portid;
+	struct net *net = info->net;
+	struct sk_buff *skb2;
+	bool reset = false;
+	char *buf;
+
+	if (info->nlh->nlmsg_flags & NLM_F_DUMP) {
+		struct netlink_dump_control c = {
+			.start = nf_tables_dump_obj_start,
+			.dump = nf_tables_dump_obj,
+			.done = nf_tables_dump_obj_done,
+			.module = THIS_MODULE,
+			.data = (void *)nla,
+		};
+
+		return nft_netlink_dump_start_rcu(info->sk, skb, info->nlh, &c);
+	}
 
 	if (NFNL_MSG_TYPE(info->nlh->nlmsg_type) == NFT_MSG_GETOBJ_RESET)
 		reset = true;
 
-	err = nf_tables_fill_obj_info(skb2, net, NETLINK_CB(skb).portid,
-				      info->nlh->nlmsg_seq, NFT_MSG_NEWOBJ, 0,
-				      family, table, obj, reset);
-	if (err < 0)
-		goto err_fill_obj_info;
+	skb2 = nf_tables_getobj_single(portid, info, nla, reset);
+	if (IS_ERR(skb2))
+		return PTR_ERR(skb2);
 
 	if (!reset)
 		return nfnetlink_unicast(skb2, net, NETLINK_CB(skb).portid);
@@ -8121,11 +8138,7 @@ static int nf_tables_getobj(struct sk_buff *skb, const struct nfnl_info *info,
 			AUDIT_NFT_OP_OBJ_RESET, GFP_ATOMIC);
 	kfree(buf);
 
-	return nfnetlink_unicast(skb2, net, NETLINK_CB(skb).portid);
-
-err_fill_obj_info:
-	kfree_skb(skb2);
-	return err;
+	return nfnetlink_unicast(skb2, net, portid);
 }
 
 static void nft_obj_destroy(const struct nft_ctx *ctx, struct nft_object *obj)
Signature · 1 line
2.43.0
author reviewer
#3 lore raw

PATCH v4 3/3 netfilter: nf_tables: Add locking for NFT_MSG_GETOBJ_RESET requests

In reply to Phil Sutter

Objects' dump callbacks are not concurrency-safe per-se with reset bit set. If two CPUs perform a reset at the same time, at least counter and quota objects suffer from value underrun.

Prevent this by introducing dedicated locking callbacks for nfnetlink and the asynchronous dump handling to serialize access.

---
 net/netfilter/nf_tables_api.c | 72 ++++++++++++++++++++++++++++-------
 1 file changed, 59 insertions(+), 13 deletions(-)
net/netfilter/nf_tables_api.c +59 -13
index c12c9cae784d..0a2f79346958 100644
@@ -8020,6 +8020,19 @@ static int nf_tables_dump_obj(struct sk_buff *skb, struct netlink_callback *cb)
 	return skb->len;
 }
 
+static int nf_tables_dumpreset_obj(struct sk_buff *skb,
+				   struct netlink_callback *cb)
+{
+	struct nftables_pernet *nft_net = nft_pernet(sock_net(skb->sk));
+	int ret;
+
+	mutex_lock(&nft_net->commit_mutex);
+	ret = nf_tables_dump_obj(skb, cb);
+	mutex_unlock(&nft_net->commit_mutex);
+
+	return ret;
+}
+
 static int nf_tables_dump_obj_start(struct netlink_callback *cb)
 {
 	struct nft_obj_dump_ctx *ctx = (void *)cb->ctx;
@@ -8036,12 +8049,18 @@ static int nf_tables_dump_obj_start(struct netlink_callback *cb)
 	if (nla[NFTA_OBJ_TYPE])
 		ctx->type = ntohl(nla_get_be32(nla[NFTA_OBJ_TYPE]));
 
-	if (NFNL_MSG_TYPE(cb->nlh->nlmsg_type) == NFT_MSG_GETOBJ_RESET)
-		ctx->reset = true;
-
 	return 0;
 }
 
+static int nf_tables_dumpreset_obj_start(struct netlink_callback *cb)
+{
+	struct nft_obj_dump_ctx *ctx = (void *)cb->ctx;
+
+	ctx->reset = true;
+
+	return nf_tables_dump_obj_start(cb);
+}
+
 static int nf_tables_dump_obj_done(struct netlink_callback *cb)
 {
 	struct nft_obj_dump_ctx *ctx = (void *)cb->ctx;
@@ -8100,18 +8119,43 @@ nf_tables_getobj_single(u32 portid, const struct nfnl_info *info,
 
 static int nf_tables_getobj(struct sk_buff *skb, const struct nfnl_info *info,
 			    const struct nlattr * const nla[])
+{
+	u32 portid = NETLINK_CB(skb).portid;
+	struct sk_buff *skb2;
+
+	if (info->nlh->nlmsg_flags & NLM_F_DUMP) {
+		struct netlink_dump_control c = {
+			.start = nf_tables_dump_obj_start,
+			.dump = nf_tables_dump_obj,
+			.done = nf_tables_dump_obj_done,
+			.module = THIS_MODULE,
+			.data = (void *)nla,
+		};
+
+		return nft_netlink_dump_start_rcu(info->sk, skb, info->nlh, &c);
+	}
+
+	skb2 = nf_tables_getobj_single(portid, info, nla, false);
+	if (IS_ERR(skb2))
+		return PTR_ERR(skb2);
+
+	return nfnetlink_unicast(skb2, info->net, portid);
+}
+
+static int nf_tables_getobj_reset(struct sk_buff *skb,
+				  const struct nfnl_info *info,
+				  const struct nlattr * const nla[])
 {
 	struct nftables_pernet *nft_net = nft_pernet(info->net);
 	u32 portid = NETLINK_CB(skb).portid;
 	struct net *net = info->net;
 	struct sk_buff *skb2;
-	bool reset = false;
 	char *buf;
 
 	if (info->nlh->nlmsg_flags & NLM_F_DUMP) {
 		struct netlink_dump_control c = {
-			.start = nf_tables_dump_obj_start,
-			.dump = nf_tables_dump_obj,
+			.start = nf_tables_dumpreset_obj_start,
+			.dump = nf_tables_dumpreset_obj,
 			.done = nf_tables_dump_obj_done,
 			.module = THIS_MODULE,
 			.data = (void *)nla,
@@ -8120,16 +8164,18 @@ static int nf_tables_getobj(struct sk_buff *skb, const struct nfnl_info *info,
 		return nft_netlink_dump_start_rcu(info->sk, skb, info->nlh, &c);
 	}
 
-	if (NFNL_MSG_TYPE(info->nlh->nlmsg_type) == NFT_MSG_GETOBJ_RESET)
-		reset = true;
+	if (!try_module_get(THIS_MODULE))
+		return -EINVAL;
+	rcu_read_unlock();
+	mutex_lock(&nft_net->commit_mutex);
+	skb2 = nf_tables_getobj_single(portid, info, nla, true);
+	mutex_unlock(&nft_net->commit_mutex);
+	rcu_read_lock();
+	module_put(THIS_MODULE);
 
-	skb2 = nf_tables_getobj_single(portid, info, nla, reset);
 	if (IS_ERR(skb2))
 		return PTR_ERR(skb2);
 
-	if (!reset)
-		return nfnetlink_unicast(skb2, net, NETLINK_CB(skb).portid);
-
 	buf = kasprintf(GFP_ATOMIC, "%.*s:%u",
 			nla_len(nla[NFTA_OBJ_TABLE]),
 			(char *)nla_data(nla[NFTA_OBJ_TABLE]),
@@ -9421,7 +9467,7 @@ static const struct nfnl_callback nf_tables_cb[NFT_MSG_MAX] = {
 		.policy		= nft_obj_policy,
 	},
 	[NFT_MSG_GETOBJ_RESET] = {
-		.call		= nf_tables_getobj,
+		.call		= nf_tables_getobj_reset,
 		.type		= NFNL_CB_RCU,
 		.attr_count	= NFTA_OBJ_MAX,
 		.policy		= nft_obj_policy,
Signature · 1 line
2.43.0