* [PATCH nf-next 01/19] netfilter: nf_tables: make struct nft_trans first member of derived subtypes
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
@ 2024-06-27 11:26 ` Pablo Neira Ayuso
2024-06-27 11:26 ` [PATCH nf-next 02/19] netfilter: nf_tables: move bind list_head into relevant subtypes Pablo Neira Ayuso
` (18 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:26 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
There is 'struct nft_trans', the basic structure for all transactional
objects, and the the various different transactional objects, such as
nft_trans_table, chain, set, set_elem and so on.
Right now 'struct nft_trans' uses a flexible member at the tail
(data[]), and casting is needed to access the actual type-specific
members.
Change this to make the hierarchy visible in source code, i.e. make
struct nft_trans the first member of all derived subtypes.
This has several advantages:
1. pahole output reflects the real size needed by the particular subtype
2. allows to use container_of() to convert the base type to the actual
object type instead of casting ->data to the overlay structure.
3. It makes it easy to add intermediate types.
'struct nft_trans' contains a 'binding_list' that is only needed
by two subtypes, so it should be part of the two subtypes, not in
the base structure.
But that makes it hard to interate over the binding_list, because
there is no common base structure.
A follow patch moves the bind list to a new struct:
struct nft_trans_binding {
struct nft_trans nft_trans;
struct list_head binding_list;
};
... and makes that structure the new 'first member' for both
nft_trans_chain and nft_trans_set.
No functional change intended in this patch.
Some numbers:
struct nft_trans { /* size: 88, cachelines: 2, members: 5 */
struct nft_trans_chain { /* size: 152, cachelines: 3, members: 10 */
struct nft_trans_elem { /* size: 112, cachelines: 2, members: 4 */
struct nft_trans_flowtable { /* size: 128, cachelines: 2, members: 5 */
struct nft_trans_obj { /* size: 112, cachelines: 2, members: 4 */
struct nft_trans_rule { /* size: 112, cachelines: 2, members: 5 */
struct nft_trans_set { /* size: 120, cachelines: 2, members: 8 */
struct nft_trans_table { /* size: 96, cachelines: 2, members: 2 */
Of particular interest is nft_trans_elem, which needs to be allocated
once for each pending (to be added or removed) set element.
Add BUILD_BUG_ON to check struct nft_trans is placed at the top of
the container structure.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
include/net/netfilter/nf_tables.h | 162 +++++++++++++++++-------------
net/netfilter/nf_tables_api.c | 18 +++-
2 files changed, 105 insertions(+), 75 deletions(-)
diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index 2796153b03da..b25df037fceb 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -1608,14 +1608,16 @@ static inline int nft_set_elem_is_dead(const struct nft_set_ext *ext)
}
/**
- * struct nft_trans - nf_tables object update in transaction
+ * struct nft_trans - nf_tables object update in transaction
*
- * @list: used internally
- * @binding_list: list of objects with possible bindings
- * @msg_type: message type
- * @put_net: ctx->net needs to be put
- * @ctx: transaction context
- * @data: internal information related to the transaction
+ * @list: used internally
+ * @binding_list: list of objects with possible bindings
+ * @msg_type: message type
+ * @put_net: ctx->net needs to be put
+ * @ctx: transaction context
+ *
+ * This is the information common to all objects in the transaction,
+ * this must always be the first member of derived sub-types.
*/
struct nft_trans {
struct list_head list;
@@ -1623,26 +1625,29 @@ struct nft_trans {
int msg_type;
bool put_net;
struct nft_ctx ctx;
- char data[];
};
struct nft_trans_rule {
+ struct nft_trans nft_trans;
struct nft_rule *rule;
struct nft_flow_rule *flow;
u32 rule_id;
bool bound;
};
-#define nft_trans_rule(trans) \
- (((struct nft_trans_rule *)trans->data)->rule)
-#define nft_trans_flow_rule(trans) \
- (((struct nft_trans_rule *)trans->data)->flow)
-#define nft_trans_rule_id(trans) \
- (((struct nft_trans_rule *)trans->data)->rule_id)
-#define nft_trans_rule_bound(trans) \
- (((struct nft_trans_rule *)trans->data)->bound)
+#define nft_trans_container_rule(trans) \
+ container_of(trans, struct nft_trans_rule, nft_trans)
+#define nft_trans_rule(trans) \
+ nft_trans_container_rule(trans)->rule
+#define nft_trans_flow_rule(trans) \
+ nft_trans_container_rule(trans)->flow
+#define nft_trans_rule_id(trans) \
+ nft_trans_container_rule(trans)->rule_id
+#define nft_trans_rule_bound(trans) \
+ nft_trans_container_rule(trans)->bound
struct nft_trans_set {
+ struct nft_trans nft_trans;
struct nft_set *set;
u32 set_id;
u32 gc_int;
@@ -1652,22 +1657,25 @@ struct nft_trans_set {
u32 size;
};
-#define nft_trans_set(trans) \
- (((struct nft_trans_set *)trans->data)->set)
-#define nft_trans_set_id(trans) \
- (((struct nft_trans_set *)trans->data)->set_id)
-#define nft_trans_set_bound(trans) \
- (((struct nft_trans_set *)trans->data)->bound)
-#define nft_trans_set_update(trans) \
- (((struct nft_trans_set *)trans->data)->update)
-#define nft_trans_set_timeout(trans) \
- (((struct nft_trans_set *)trans->data)->timeout)
-#define nft_trans_set_gc_int(trans) \
- (((struct nft_trans_set *)trans->data)->gc_int)
-#define nft_trans_set_size(trans) \
- (((struct nft_trans_set *)trans->data)->size)
+#define nft_trans_container_set(trans) \
+ container_of(trans, struct nft_trans_set, nft_trans)
+#define nft_trans_set(trans) \
+ nft_trans_container_set(trans)->set
+#define nft_trans_set_id(trans) \
+ nft_trans_container_set(trans)->set_id
+#define nft_trans_set_bound(trans) \
+ nft_trans_container_set(trans)->bound
+#define nft_trans_set_update(trans) \
+ nft_trans_container_set(trans)->update
+#define nft_trans_set_timeout(trans) \
+ nft_trans_container_set(trans)->timeout
+#define nft_trans_set_gc_int(trans) \
+ nft_trans_container_set(trans)->gc_int
+#define nft_trans_set_size(trans) \
+ nft_trans_container_set(trans)->size
struct nft_trans_chain {
+ struct nft_trans nft_trans;
struct nft_chain *chain;
bool update;
char *name;
@@ -1679,73 +1687,87 @@ struct nft_trans_chain {
struct list_head hook_list;
};
-#define nft_trans_chain(trans) \
- (((struct nft_trans_chain *)trans->data)->chain)
-#define nft_trans_chain_update(trans) \
- (((struct nft_trans_chain *)trans->data)->update)
-#define nft_trans_chain_name(trans) \
- (((struct nft_trans_chain *)trans->data)->name)
-#define nft_trans_chain_stats(trans) \
- (((struct nft_trans_chain *)trans->data)->stats)
-#define nft_trans_chain_policy(trans) \
- (((struct nft_trans_chain *)trans->data)->policy)
-#define nft_trans_chain_bound(trans) \
- (((struct nft_trans_chain *)trans->data)->bound)
-#define nft_trans_chain_id(trans) \
- (((struct nft_trans_chain *)trans->data)->chain_id)
-#define nft_trans_basechain(trans) \
- (((struct nft_trans_chain *)trans->data)->basechain)
-#define nft_trans_chain_hooks(trans) \
- (((struct nft_trans_chain *)trans->data)->hook_list)
+#define nft_trans_container_chain(trans) \
+ container_of(trans, struct nft_trans_chain, nft_trans)
+#define nft_trans_chain(trans) \
+ nft_trans_container_chain(trans)->chain
+#define nft_trans_chain_update(trans) \
+ nft_trans_container_chain(trans)->update
+#define nft_trans_chain_name(trans) \
+ nft_trans_container_chain(trans)->name
+#define nft_trans_chain_stats(trans) \
+ nft_trans_container_chain(trans)->stats
+#define nft_trans_chain_policy(trans) \
+ nft_trans_container_chain(trans)->policy
+#define nft_trans_chain_bound(trans) \
+ nft_trans_container_chain(trans)->bound
+#define nft_trans_chain_id(trans) \
+ nft_trans_container_chain(trans)->chain_id
+#define nft_trans_basechain(trans) \
+ nft_trans_container_chain(trans)->basechain
+#define nft_trans_chain_hooks(trans) \
+ nft_trans_container_chain(trans)->hook_list
struct nft_trans_table {
+ struct nft_trans nft_trans;
bool update;
};
-#define nft_trans_table_update(trans) \
- (((struct nft_trans_table *)trans->data)->update)
+#define nft_trans_container_table(trans) \
+ container_of(trans, struct nft_trans_table, nft_trans)
+#define nft_trans_table_update(trans) \
+ nft_trans_container_table(trans)->update
struct nft_trans_elem {
+ struct nft_trans nft_trans;
struct nft_set *set;
struct nft_elem_priv *elem_priv;
bool bound;
};
-#define nft_trans_elem_set(trans) \
- (((struct nft_trans_elem *)trans->data)->set)
-#define nft_trans_elem_priv(trans) \
- (((struct nft_trans_elem *)trans->data)->elem_priv)
-#define nft_trans_elem_set_bound(trans) \
- (((struct nft_trans_elem *)trans->data)->bound)
+#define nft_trans_container_elem(t) \
+ container_of(t, struct nft_trans_elem, nft_trans)
+#define nft_trans_elem_set(trans) \
+ nft_trans_container_elem(trans)->set
+#define nft_trans_elem_priv(trans) \
+ nft_trans_container_elem(trans)->elem_priv
+#define nft_trans_elem_set_bound(trans) \
+ nft_trans_container_elem(trans)->bound
struct nft_trans_obj {
+ struct nft_trans nft_trans;
struct nft_object *obj;
struct nft_object *newobj;
bool update;
};
-#define nft_trans_obj(trans) \
- (((struct nft_trans_obj *)trans->data)->obj)
-#define nft_trans_obj_newobj(trans) \
- (((struct nft_trans_obj *)trans->data)->newobj)
-#define nft_trans_obj_update(trans) \
- (((struct nft_trans_obj *)trans->data)->update)
+#define nft_trans_container_obj(t) \
+ container_of(t, struct nft_trans_obj, nft_trans)
+#define nft_trans_obj(trans) \
+ nft_trans_container_obj(trans)->obj
+#define nft_trans_obj_newobj(trans) \
+ nft_trans_container_obj(trans)->newobj
+#define nft_trans_obj_update(trans) \
+ nft_trans_container_obj(trans)->update
struct nft_trans_flowtable {
+ struct nft_trans nft_trans;
struct nft_flowtable *flowtable;
bool update;
struct list_head hook_list;
u32 flags;
};
-#define nft_trans_flowtable(trans) \
- (((struct nft_trans_flowtable *)trans->data)->flowtable)
-#define nft_trans_flowtable_update(trans) \
- (((struct nft_trans_flowtable *)trans->data)->update)
-#define nft_trans_flowtable_hooks(trans) \
- (((struct nft_trans_flowtable *)trans->data)->hook_list)
-#define nft_trans_flowtable_flags(trans) \
- (((struct nft_trans_flowtable *)trans->data)->flags)
+#define nft_trans_container_flowtable(t) \
+ container_of(t, struct nft_trans_flowtable, nft_trans)
+#define nft_trans_flowtable(trans) \
+ nft_trans_container_flowtable(trans)->flowtable
+#define nft_trans_flowtable_update(trans) \
+ nft_trans_container_flowtable(trans)->update
+#define nft_trans_flowtable_hooks(trans) \
+ nft_trans_container_flowtable(trans)->hook_list
+#define nft_trans_flowtable_flags(trans) \
+ nft_trans_container_flowtable(trans)->flags
#define NFT_TRANS_GC_BATCHCOUNT 256
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index be3b4c90d2ed..19edd1bcecef 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -153,7 +153,7 @@ static struct nft_trans *nft_trans_alloc_gfp(const struct nft_ctx *ctx,
{
struct nft_trans *trans;
- trans = kzalloc(sizeof(struct nft_trans) + size, gfp);
+ trans = kzalloc(size, gfp);
if (trans == NULL)
return NULL;
@@ -10348,7 +10348,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
trans->msg_type, GFP_KERNEL);
break;
case NFT_MSG_NEWSETELEM:
- te = (struct nft_trans_elem *)trans->data;
+ te = nft_trans_container_elem(trans);
nft_setelem_activate(net, te->set, te->elem_priv);
nf_tables_setelem_notify(&trans->ctx, te->set,
@@ -10363,7 +10363,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
break;
case NFT_MSG_DELSETELEM:
case NFT_MSG_DESTROYSETELEM:
- te = (struct nft_trans_elem *)trans->data;
+ te = nft_trans_container_elem(trans);
nf_tables_setelem_notify(&trans->ctx, te->set,
te->elem_priv,
@@ -10643,7 +10643,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
nft_trans_destroy(trans);
break;
}
- te = (struct nft_trans_elem *)trans->data;
+ te = nft_trans_container_elem(trans);
nft_setelem_remove(net, te->set, te->elem_priv);
if (!nft_setelem_is_catchall(te->set, te->elem_priv))
atomic_dec(&te->set->nelems);
@@ -10656,7 +10656,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
break;
case NFT_MSG_DELSETELEM:
case NFT_MSG_DESTROYSETELEM:
- te = (struct nft_trans_elem *)trans->data;
+ te = nft_trans_container_elem(trans);
if (!nft_setelem_active_next(net, te->set, te->elem_priv)) {
nft_setelem_data_activate(net, te->set, te->elem_priv);
@@ -11588,6 +11588,14 @@ static int __init nf_tables_module_init(void)
{
int err;
+ BUILD_BUG_ON(offsetof(struct nft_trans_table, nft_trans) != 0);
+ BUILD_BUG_ON(offsetof(struct nft_trans_chain, nft_trans) != 0);
+ BUILD_BUG_ON(offsetof(struct nft_trans_rule, nft_trans) != 0);
+ BUILD_BUG_ON(offsetof(struct nft_trans_set, nft_trans) != 0);
+ BUILD_BUG_ON(offsetof(struct nft_trans_elem, nft_trans) != 0);
+ BUILD_BUG_ON(offsetof(struct nft_trans_obj, nft_trans) != 0);
+ BUILD_BUG_ON(offsetof(struct nft_trans_flowtable, nft_trans) != 0);
+
err = register_pernet_subsys(&nf_tables_net_ops);
if (err < 0)
return err;
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 02/19] netfilter: nf_tables: move bind list_head into relevant subtypes
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
2024-06-27 11:26 ` [PATCH nf-next 01/19] netfilter: nf_tables: make struct nft_trans first member of derived subtypes Pablo Neira Ayuso
@ 2024-06-27 11:26 ` Pablo Neira Ayuso
2024-06-27 11:26 ` [PATCH nf-next 03/19] netfilter: nf_tables: compact chain+ft transaction objects Pablo Neira Ayuso
` (17 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:26 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
Only nft_trans_chain and nft_trans_set subtypes use the
trans->binding_list member.
Add a new common binding subtype and move the member there.
This reduces size of all other subtypes by 16 bytes on 64bit platforms.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
include/net/netfilter/nf_tables.h | 26 +++++++----
net/netfilter/nf_tables_api.c | 71 +++++++++++++++++++++++++------
2 files changed, 75 insertions(+), 22 deletions(-)
diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index b25df037fceb..f72448095833 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -1611,7 +1611,6 @@ static inline int nft_set_elem_is_dead(const struct nft_set_ext *ext)
* struct nft_trans - nf_tables object update in transaction
*
* @list: used internally
- * @binding_list: list of objects with possible bindings
* @msg_type: message type
* @put_net: ctx->net needs to be put
* @ctx: transaction context
@@ -1621,12 +1620,23 @@ static inline int nft_set_elem_is_dead(const struct nft_set_ext *ext)
*/
struct nft_trans {
struct list_head list;
- struct list_head binding_list;
int msg_type;
bool put_net;
struct nft_ctx ctx;
};
+/**
+ * struct nft_trans_binding - nf_tables object with binding support in transaction
+ * @nft_trans: base structure, MUST be first member
+ * @binding_list: list of objects with possible bindings
+ *
+ * This is the base type used by objects that can be bound to a chain.
+ */
+struct nft_trans_binding {
+ struct nft_trans nft_trans;
+ struct list_head binding_list;
+};
+
struct nft_trans_rule {
struct nft_trans nft_trans;
struct nft_rule *rule;
@@ -1647,7 +1657,7 @@ struct nft_trans_rule {
nft_trans_container_rule(trans)->bound
struct nft_trans_set {
- struct nft_trans nft_trans;
+ struct nft_trans_binding nft_trans_binding;
struct nft_set *set;
u32 set_id;
u32 gc_int;
@@ -1657,8 +1667,8 @@ struct nft_trans_set {
u32 size;
};
-#define nft_trans_container_set(trans) \
- container_of(trans, struct nft_trans_set, nft_trans)
+#define nft_trans_container_set(t) \
+ container_of(t, struct nft_trans_set, nft_trans_binding.nft_trans)
#define nft_trans_set(trans) \
nft_trans_container_set(trans)->set
#define nft_trans_set_id(trans) \
@@ -1675,7 +1685,7 @@ struct nft_trans_set {
nft_trans_container_set(trans)->size
struct nft_trans_chain {
- struct nft_trans nft_trans;
+ struct nft_trans_binding nft_trans_binding;
struct nft_chain *chain;
bool update;
char *name;
@@ -1687,8 +1697,8 @@ struct nft_trans_chain {
struct list_head hook_list;
};
-#define nft_trans_container_chain(trans) \
- container_of(trans, struct nft_trans_chain, nft_trans)
+#define nft_trans_container_chain(t) \
+ container_of(t, struct nft_trans_chain, nft_trans_binding.nft_trans)
#define nft_trans_chain(trans) \
nft_trans_container_chain(trans)->chain
#define nft_trans_chain_update(trans) \
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 19edd1bcecef..c950938ef612 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -158,7 +158,6 @@ static struct nft_trans *nft_trans_alloc_gfp(const struct nft_ctx *ctx,
return NULL;
INIT_LIST_HEAD(&trans->list);
- INIT_LIST_HEAD(&trans->binding_list);
trans->msg_type = msg_type;
trans->ctx = *ctx;
@@ -171,10 +170,26 @@ static struct nft_trans *nft_trans_alloc(const struct nft_ctx *ctx,
return nft_trans_alloc_gfp(ctx, msg_type, size, GFP_KERNEL);
}
+static struct nft_trans_binding *nft_trans_get_binding(struct nft_trans *trans)
+{
+ switch (trans->msg_type) {
+ case NFT_MSG_NEWCHAIN:
+ case NFT_MSG_NEWSET:
+ return container_of(trans, struct nft_trans_binding, nft_trans);
+ }
+
+ return NULL;
+}
+
static void nft_trans_list_del(struct nft_trans *trans)
{
+ struct nft_trans_binding *trans_binding;
+
list_del(&trans->list);
- list_del(&trans->binding_list);
+
+ trans_binding = nft_trans_get_binding(trans);
+ if (trans_binding)
+ list_del(&trans_binding->binding_list);
}
static void nft_trans_destroy(struct nft_trans *trans)
@@ -372,21 +387,26 @@ static void nf_tables_unregister_hook(struct net *net,
static void nft_trans_commit_list_add_tail(struct net *net, struct nft_trans *trans)
{
struct nftables_pernet *nft_net = nft_pernet(net);
+ struct nft_trans_binding *binding;
+
+ list_add_tail(&trans->list, &nft_net->commit_list);
+
+ binding = nft_trans_get_binding(trans);
+ if (!binding)
+ return;
switch (trans->msg_type) {
case NFT_MSG_NEWSET:
if (!nft_trans_set_update(trans) &&
nft_set_is_anonymous(nft_trans_set(trans)))
- list_add_tail(&trans->binding_list, &nft_net->binding_list);
+ list_add_tail(&binding->binding_list, &nft_net->binding_list);
break;
case NFT_MSG_NEWCHAIN:
if (!nft_trans_chain_update(trans) &&
nft_chain_binding(nft_trans_chain(trans)))
- list_add_tail(&trans->binding_list, &nft_net->binding_list);
+ list_add_tail(&binding->binding_list, &nft_net->binding_list);
break;
}
-
- list_add_tail(&trans->list, &nft_net->commit_list);
}
static int nft_trans_table_add(struct nft_ctx *ctx, int msg_type)
@@ -416,11 +436,27 @@ static int nft_deltable(struct nft_ctx *ctx)
return err;
}
-static struct nft_trans *nft_trans_chain_add(struct nft_ctx *ctx, int msg_type)
+static struct nft_trans *
+nft_trans_alloc_chain(const struct nft_ctx *ctx, int msg_type)
{
+ struct nft_trans_chain *trans_chain;
struct nft_trans *trans;
trans = nft_trans_alloc(ctx, msg_type, sizeof(struct nft_trans_chain));
+ if (!trans)
+ return NULL;
+
+ trans_chain = nft_trans_container_chain(trans);
+ INIT_LIST_HEAD(&trans_chain->nft_trans_binding.binding_list);
+
+ return trans;
+}
+
+static struct nft_trans *nft_trans_chain_add(struct nft_ctx *ctx, int msg_type)
+{
+ struct nft_trans *trans;
+
+ trans = nft_trans_alloc_chain(ctx, msg_type);
if (trans == NULL)
return ERR_PTR(-ENOMEM);
@@ -560,12 +596,16 @@ static int __nft_trans_set_add(const struct nft_ctx *ctx, int msg_type,
struct nft_set *set,
const struct nft_set_desc *desc)
{
+ struct nft_trans_set *trans_set;
struct nft_trans *trans;
trans = nft_trans_alloc(ctx, msg_type, sizeof(struct nft_trans_set));
if (trans == NULL)
return -ENOMEM;
+ trans_set = nft_trans_container_set(trans);
+ INIT_LIST_HEAD(&trans_set->nft_trans_binding.binding_list);
+
if (msg_type == NFT_MSG_NEWSET && ctx->nla[NFTA_SET_ID] && !desc) {
nft_trans_set_id(trans) =
ntohl(nla_get_be32(ctx->nla[NFTA_SET_ID]));
@@ -2698,8 +2738,7 @@ static int nf_tables_updchain(struct nft_ctx *ctx, u8 genmask, u8 policy,
}
err = -ENOMEM;
- trans = nft_trans_alloc(ctx, NFT_MSG_NEWCHAIN,
- sizeof(struct nft_trans_chain));
+ trans = nft_trans_alloc_chain(ctx, NFT_MSG_NEWCHAIN);
if (trans == NULL)
goto err_trans;
@@ -2915,8 +2954,7 @@ static int nft_delchain_hook(struct nft_ctx *ctx,
list_move(&hook->list, &chain_del_list);
}
- trans = nft_trans_alloc(ctx, NFT_MSG_DELCHAIN,
- sizeof(struct nft_trans_chain));
+ trans = nft_trans_alloc_chain(ctx, NFT_MSG_DELCHAIN);
if (!trans) {
err = -ENOMEM;
goto err_chain_del_hook;
@@ -10147,6 +10185,7 @@ static void nft_gc_seq_end(struct nftables_pernet *nft_net, unsigned int gc_seq)
static int nf_tables_commit(struct net *net, struct sk_buff *skb)
{
struct nftables_pernet *nft_net = nft_pernet(net);
+ struct nft_trans_binding *trans_binding;
struct nft_trans *trans, *next;
unsigned int base_seq, gc_seq;
LIST_HEAD(set_update_list);
@@ -10161,7 +10200,8 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
return 0;
}
- list_for_each_entry(trans, &nft_net->binding_list, binding_list) {
+ list_for_each_entry(trans_binding, &nft_net->binding_list, binding_list) {
+ trans = &trans_binding->nft_trans;
switch (trans->msg_type) {
case NFT_MSG_NEWSET:
if (!nft_trans_set_update(trans) &&
@@ -10179,6 +10219,9 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
return -EINVAL;
}
break;
+ default:
+ WARN_ONCE(1, "Unhandled bind type %d", trans->msg_type);
+ break;
}
}
@@ -11589,9 +11632,9 @@ static int __init nf_tables_module_init(void)
int err;
BUILD_BUG_ON(offsetof(struct nft_trans_table, nft_trans) != 0);
- BUILD_BUG_ON(offsetof(struct nft_trans_chain, nft_trans) != 0);
+ BUILD_BUG_ON(offsetof(struct nft_trans_chain, nft_trans_binding.nft_trans) != 0);
BUILD_BUG_ON(offsetof(struct nft_trans_rule, nft_trans) != 0);
- BUILD_BUG_ON(offsetof(struct nft_trans_set, nft_trans) != 0);
+ BUILD_BUG_ON(offsetof(struct nft_trans_set, nft_trans_binding.nft_trans) != 0);
BUILD_BUG_ON(offsetof(struct nft_trans_elem, nft_trans) != 0);
BUILD_BUG_ON(offsetof(struct nft_trans_obj, nft_trans) != 0);
BUILD_BUG_ON(offsetof(struct nft_trans_flowtable, nft_trans) != 0);
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 03/19] netfilter: nf_tables: compact chain+ft transaction objects
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
2024-06-27 11:26 ` [PATCH nf-next 01/19] netfilter: nf_tables: make struct nft_trans first member of derived subtypes Pablo Neira Ayuso
2024-06-27 11:26 ` [PATCH nf-next 02/19] netfilter: nf_tables: move bind list_head into relevant subtypes Pablo Neira Ayuso
@ 2024-06-27 11:26 ` Pablo Neira Ayuso
2024-06-27 11:26 ` [PATCH nf-next 04/19] netfilter: nf_tables: reduce trans->ctx.table references Pablo Neira Ayuso
` (16 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:26 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
Cover holes to reduce both structures by 8 byte.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
include/net/netfilter/nf_tables.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index f72448095833..1f0607b671ac 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -1687,10 +1687,10 @@ struct nft_trans_set {
struct nft_trans_chain {
struct nft_trans_binding nft_trans_binding;
struct nft_chain *chain;
- bool update;
char *name;
struct nft_stats __percpu *stats;
u8 policy;
+ bool update;
bool bound;
u32 chain_id;
struct nft_base_chain *basechain;
@@ -1763,9 +1763,9 @@ struct nft_trans_obj {
struct nft_trans_flowtable {
struct nft_trans nft_trans;
struct nft_flowtable *flowtable;
- bool update;
struct list_head hook_list;
u32 flags;
+ bool update;
};
#define nft_trans_container_flowtable(t) \
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 04/19] netfilter: nf_tables: reduce trans->ctx.table references
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (2 preceding siblings ...)
2024-06-27 11:26 ` [PATCH nf-next 03/19] netfilter: nf_tables: compact chain+ft transaction objects Pablo Neira Ayuso
@ 2024-06-27 11:26 ` Pablo Neira Ayuso
2024-06-27 11:26 ` [PATCH nf-next 05/19] netfilter: nf_tables: pass nft_chain to destroy function, not nft_ctx Pablo Neira Ayuso
` (15 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:26 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
nft_ctx is huge, it should not be stored in nft_trans at all,
most information is not needed.
Preparation patch to remove trans->ctx, no change in behaviour intended.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nf_tables_api.c | 79 ++++++++++++++++++-----------------
1 file changed, 41 insertions(+), 38 deletions(-)
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index c950938ef612..60c435774db8 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -9472,14 +9472,15 @@ static void nft_chain_commit_drop_policy(struct nft_trans *trans)
static void nft_chain_commit_update(struct nft_trans *trans)
{
+ struct nft_table *table = trans->ctx.table;
struct nft_base_chain *basechain;
if (nft_trans_chain_name(trans)) {
- rhltable_remove(&trans->ctx.table->chains_ht,
+ rhltable_remove(&table->chains_ht,
&trans->ctx.chain->rhlhead,
nft_chain_ht_params);
swap(trans->ctx.chain->name, nft_trans_chain_name(trans));
- rhltable_insert_key(&trans->ctx.table->chains_ht,
+ rhltable_insert_key(&table->chains_ht,
trans->ctx.chain->name,
&trans->ctx.chain->rhlhead,
nft_chain_ht_params);
@@ -10237,9 +10238,10 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
/* 1. Allocate space for next generation rules_gen_X[] */
list_for_each_entry_safe(trans, next, &nft_net->commit_list, list) {
+ struct nft_table *table = trans->ctx.table;
int ret;
- ret = nf_tables_commit_audit_alloc(&adl, trans->ctx.table);
+ ret = nf_tables_commit_audit_alloc(&adl, table);
if (ret) {
nf_tables_commit_chain_prepare_cancel(net);
nf_tables_commit_audit_free(&adl);
@@ -10280,28 +10282,29 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
net->nft.gencursor = nft_gencursor_next(net);
list_for_each_entry_safe(trans, next, &nft_net->commit_list, list) {
- nf_tables_commit_audit_collect(&adl, trans->ctx.table,
- trans->msg_type);
+ struct nft_table *table = trans->ctx.table;
+
+ nf_tables_commit_audit_collect(&adl, table, trans->msg_type);
switch (trans->msg_type) {
case NFT_MSG_NEWTABLE:
if (nft_trans_table_update(trans)) {
- if (!(trans->ctx.table->flags & __NFT_TABLE_F_UPDATE)) {
+ if (!(table->flags & __NFT_TABLE_F_UPDATE)) {
nft_trans_destroy(trans);
break;
}
- if (trans->ctx.table->flags & NFT_TABLE_F_DORMANT)
- nf_tables_table_disable(net, trans->ctx.table);
+ if (table->flags & NFT_TABLE_F_DORMANT)
+ nf_tables_table_disable(net, table);
- trans->ctx.table->flags &= ~__NFT_TABLE_F_UPDATE;
+ table->flags &= ~__NFT_TABLE_F_UPDATE;
} else {
- nft_clear(net, trans->ctx.table);
+ nft_clear(net, table);
}
nf_tables_table_notify(&trans->ctx, NFT_MSG_NEWTABLE);
nft_trans_destroy(trans);
break;
case NFT_MSG_DELTABLE:
case NFT_MSG_DESTROYTABLE:
- list_del_rcu(&trans->ctx.table->list);
+ list_del_rcu(&table->list);
nf_tables_table_notify(&trans->ctx, trans->msg_type);
break;
case NFT_MSG_NEWCHAIN:
@@ -10324,7 +10327,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
if (nft_trans_chain_update(trans)) {
nf_tables_chain_notify(&trans->ctx, NFT_MSG_DELCHAIN,
&nft_trans_chain_hooks(trans));
- if (!(trans->ctx.table->flags & NFT_TABLE_F_DORMANT)) {
+ if (!(table->flags & NFT_TABLE_F_DORMANT)) {
nft_netdev_unregister_hooks(net,
&nft_trans_chain_hooks(trans),
true);
@@ -10333,8 +10336,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
nft_chain_del(trans->ctx.chain);
nf_tables_chain_notify(&trans->ctx, NFT_MSG_DELCHAIN,
NULL);
- nf_tables_unregister_hook(trans->ctx.net,
- trans->ctx.table,
+ nf_tables_unregister_hook(trans->ctx.net, table,
trans->ctx.chain);
}
break;
@@ -10377,7 +10379,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
*/
if (nft_set_is_anonymous(nft_trans_set(trans)) &&
!list_empty(&nft_trans_set(trans)->bindings))
- nft_use_dec(&trans->ctx.table->use);
+ nft_use_dec(&table->use);
}
nf_tables_set_notify(&trans->ctx, nft_trans_set(trans),
NFT_MSG_NEWSET, GFP_KERNEL);
@@ -10575,37 +10577,39 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
list_for_each_entry_safe_reverse(trans, next, &nft_net->commit_list,
list) {
+ struct nft_table *table = trans->ctx.table;
+
switch (trans->msg_type) {
case NFT_MSG_NEWTABLE:
if (nft_trans_table_update(trans)) {
- if (!(trans->ctx.table->flags & __NFT_TABLE_F_UPDATE)) {
+ if (!(table->flags & __NFT_TABLE_F_UPDATE)) {
nft_trans_destroy(trans);
break;
}
- if (trans->ctx.table->flags & __NFT_TABLE_F_WAS_DORMANT) {
- nf_tables_table_disable(net, trans->ctx.table);
- trans->ctx.table->flags |= NFT_TABLE_F_DORMANT;
- } else if (trans->ctx.table->flags & __NFT_TABLE_F_WAS_AWAKEN) {
- trans->ctx.table->flags &= ~NFT_TABLE_F_DORMANT;
+ if (table->flags & __NFT_TABLE_F_WAS_DORMANT) {
+ nf_tables_table_disable(net, table);
+ table->flags |= NFT_TABLE_F_DORMANT;
+ } else if (table->flags & __NFT_TABLE_F_WAS_AWAKEN) {
+ table->flags &= ~NFT_TABLE_F_DORMANT;
}
- if (trans->ctx.table->flags & __NFT_TABLE_F_WAS_ORPHAN) {
- trans->ctx.table->flags &= ~NFT_TABLE_F_OWNER;
- trans->ctx.table->nlpid = 0;
+ if (table->flags & __NFT_TABLE_F_WAS_ORPHAN) {
+ table->flags &= ~NFT_TABLE_F_OWNER;
+ table->nlpid = 0;
}
- trans->ctx.table->flags &= ~__NFT_TABLE_F_UPDATE;
+ table->flags &= ~__NFT_TABLE_F_UPDATE;
nft_trans_destroy(trans);
} else {
- list_del_rcu(&trans->ctx.table->list);
+ list_del_rcu(&table->list);
}
break;
case NFT_MSG_DELTABLE:
case NFT_MSG_DESTROYTABLE:
- nft_clear(trans->ctx.net, trans->ctx.table);
+ nft_clear(trans->ctx.net, table);
nft_trans_destroy(trans);
break;
case NFT_MSG_NEWCHAIN:
if (nft_trans_chain_update(trans)) {
- if (!(trans->ctx.table->flags & NFT_TABLE_F_DORMANT)) {
+ if (!(table->flags & NFT_TABLE_F_DORMANT)) {
nft_netdev_unregister_hooks(net,
&nft_trans_chain_hooks(trans),
true);
@@ -10618,10 +10622,9 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
nft_trans_destroy(trans);
break;
}
- nft_use_dec_restore(&trans->ctx.table->use);
+ nft_use_dec_restore(&table->use);
nft_chain_del(trans->ctx.chain);
- nf_tables_unregister_hook(trans->ctx.net,
- trans->ctx.table,
+ nf_tables_unregister_hook(trans->ctx.net, table,
trans->ctx.chain);
}
break;
@@ -10631,7 +10634,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
list_splice(&nft_trans_chain_hooks(trans),
&nft_trans_basechain(trans)->hook_list);
} else {
- nft_use_inc_restore(&trans->ctx.table->use);
+ nft_use_inc_restore(&table->use);
nft_clear(trans->ctx.net, trans->ctx.chain);
}
nft_trans_destroy(trans);
@@ -10664,7 +10667,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
nft_trans_destroy(trans);
break;
}
- nft_use_dec_restore(&trans->ctx.table->use);
+ nft_use_dec_restore(&table->use);
if (nft_trans_set_bound(trans)) {
nft_trans_destroy(trans);
break;
@@ -10674,7 +10677,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
break;
case NFT_MSG_DELSET:
case NFT_MSG_DESTROYSET:
- nft_use_inc_restore(&trans->ctx.table->use);
+ nft_use_inc_restore(&table->use);
nft_clear(trans->ctx.net, nft_trans_set(trans));
if (nft_trans_set(trans)->flags & (NFT_SET_MAP | NFT_SET_OBJECT))
nft_map_activate(&trans->ctx, nft_trans_set(trans));
@@ -10720,13 +10723,13 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
nft_obj_destroy(&trans->ctx, nft_trans_obj_newobj(trans));
nft_trans_destroy(trans);
} else {
- nft_use_dec_restore(&trans->ctx.table->use);
+ nft_use_dec_restore(&table->use);
nft_obj_del(nft_trans_obj(trans));
}
break;
case NFT_MSG_DELOBJ:
case NFT_MSG_DESTROYOBJ:
- nft_use_inc_restore(&trans->ctx.table->use);
+ nft_use_inc_restore(&table->use);
nft_clear(trans->ctx.net, nft_trans_obj(trans));
nft_trans_destroy(trans);
break;
@@ -10735,7 +10738,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
nft_unregister_flowtable_net_hooks(net,
&nft_trans_flowtable_hooks(trans));
} else {
- nft_use_dec_restore(&trans->ctx.table->use);
+ nft_use_dec_restore(&table->use);
list_del_rcu(&nft_trans_flowtable(trans)->list);
nft_unregister_flowtable_net_hooks(net,
&nft_trans_flowtable(trans)->hook_list);
@@ -10747,7 +10750,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
list_splice(&nft_trans_flowtable_hooks(trans),
&nft_trans_flowtable(trans)->hook_list);
} else {
- nft_use_inc_restore(&trans->ctx.table->use);
+ nft_use_inc_restore(&table->use);
nft_clear(trans->ctx.net, nft_trans_flowtable(trans));
}
nft_trans_destroy(trans);
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 05/19] netfilter: nf_tables: pass nft_chain to destroy function, not nft_ctx
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (3 preceding siblings ...)
2024-06-27 11:26 ` [PATCH nf-next 04/19] netfilter: nf_tables: reduce trans->ctx.table references Pablo Neira Ayuso
@ 2024-06-27 11:26 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 06/19] netfilter: nf_tables: pass more specific nft_trans_chain where possible Pablo Neira Ayuso
` (14 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:26 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
It would be better to not store nft_ctx inside nft_trans object,
the netlink ctx strucutre is huge and most of its information is
never needed in places that use trans->ctx.
Avoid/reduce its usage if possible, no runtime behaviour change
intended.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
include/net/netfilter/nf_tables.h | 2 +-
net/netfilter/nf_tables_api.c | 17 ++++++++---------
net/netfilter/nft_immediate.c | 2 +-
3 files changed, 10 insertions(+), 11 deletions(-)
diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index 1f0607b671ac..328fdc140551 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -1171,7 +1171,7 @@ static inline bool nft_chain_is_bound(struct nft_chain *chain)
int nft_chain_add(struct nft_table *table, struct nft_chain *chain);
void nft_chain_del(struct nft_chain *chain);
-void nf_tables_chain_destroy(struct nft_ctx *ctx);
+void nf_tables_chain_destroy(struct nft_chain *chain);
struct nft_stats {
u64 bytes;
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 60c435774db8..bdc2d7f781ca 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -2118,9 +2118,9 @@ static void nf_tables_chain_free_chain_rules(struct nft_chain *chain)
kvfree(chain->blob_next);
}
-void nf_tables_chain_destroy(struct nft_ctx *ctx)
+void nf_tables_chain_destroy(struct nft_chain *chain)
{
- struct nft_chain *chain = ctx->chain;
+ const struct nft_table *table = chain->table;
struct nft_hook *hook, *next;
if (WARN_ON(chain->use > 0))
@@ -2132,7 +2132,7 @@ void nf_tables_chain_destroy(struct nft_ctx *ctx)
if (nft_is_base_chain(chain)) {
struct nft_base_chain *basechain = nft_base_chain(chain);
- if (nft_base_chain_netdev(ctx->family, basechain->ops.hooknum)) {
+ if (nft_base_chain_netdev(table->family, basechain->ops.hooknum)) {
list_for_each_entry_safe(hook, next,
&basechain->hook_list, list) {
list_del_rcu(&hook->list);
@@ -2621,7 +2621,7 @@ static int nf_tables_addchain(struct nft_ctx *ctx, u8 family, u8 genmask,
err_trans:
nft_use_dec_restore(&table->use);
err_destroy_chain:
- nf_tables_chain_destroy(ctx);
+ nf_tables_chain_destroy(chain);
return err;
}
@@ -9532,7 +9532,7 @@ static void nft_commit_release(struct nft_trans *trans)
if (nft_trans_chain_update(trans))
nft_hooks_destroy(&nft_trans_chain_hooks(trans));
else
- nf_tables_chain_destroy(&trans->ctx);
+ nf_tables_chain_destroy(nft_trans_chain(trans));
break;
case NFT_MSG_DELRULE:
case NFT_MSG_DESTROYRULE:
@@ -10524,7 +10524,7 @@ static void nf_tables_abort_release(struct nft_trans *trans)
if (nft_trans_chain_update(trans))
nft_hooks_destroy(&nft_trans_chain_hooks(trans));
else
- nf_tables_chain_destroy(&trans->ctx);
+ nf_tables_chain_destroy(nft_trans_chain(trans));
break;
case NFT_MSG_NEWRULE:
nf_tables_rule_destroy(&trans->ctx, nft_trans_rule(trans));
@@ -11411,7 +11411,7 @@ int __nft_release_basechain(struct nft_ctx *ctx)
}
nft_chain_del(ctx->chain);
nft_use_dec(&ctx->table->use);
- nf_tables_chain_destroy(ctx);
+ nf_tables_chain_destroy(ctx->chain);
return 0;
}
@@ -11486,10 +11486,9 @@ static void __nft_release_table(struct net *net, struct nft_table *table)
nft_obj_destroy(&ctx, obj);
}
list_for_each_entry_safe(chain, nc, &table->chains, list) {
- ctx.chain = chain;
nft_chain_del(chain);
nft_use_dec(&table->use);
- nf_tables_chain_destroy(&ctx);
+ nf_tables_chain_destroy(chain);
}
nf_tables_table_destroy(&ctx);
}
diff --git a/net/netfilter/nft_immediate.c b/net/netfilter/nft_immediate.c
index 6475c7abc1fe..ac2422c215e5 100644
--- a/net/netfilter/nft_immediate.c
+++ b/net/netfilter/nft_immediate.c
@@ -221,7 +221,7 @@ static void nft_immediate_destroy(const struct nft_ctx *ctx,
list_del(&rule->list);
nf_tables_rule_destroy(&chain_ctx, rule);
}
- nf_tables_chain_destroy(&chain_ctx);
+ nf_tables_chain_destroy(chain);
break;
default:
break;
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 06/19] netfilter: nf_tables: pass more specific nft_trans_chain where possible
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (4 preceding siblings ...)
2024-06-27 11:26 ` [PATCH nf-next 05/19] netfilter: nf_tables: pass nft_chain to destroy function, not nft_ctx Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 07/19] netfilter: nf_tables: avoid usage of embedded nft_ctx Pablo Neira Ayuso
` (13 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
These functions pass a pointer to the base object type, use the
more specific one. No functional change intended.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nf_tables_api.c | 51 ++++++++++++++++++-----------------
1 file changed, 26 insertions(+), 25 deletions(-)
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index bdc2d7f781ca..62a4da955574 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -448,6 +448,7 @@ nft_trans_alloc_chain(const struct nft_ctx *ctx, int msg_type)
trans_chain = nft_trans_container_chain(trans);
INIT_LIST_HEAD(&trans_chain->nft_trans_binding.binding_list);
+ trans_chain->chain = ctx->chain;
return trans;
}
@@ -468,7 +469,6 @@ static struct nft_trans *nft_trans_chain_add(struct nft_ctx *ctx, int msg_type)
ntohl(nla_get_be32(ctx->nla[NFTA_CHAIN_ID]));
}
}
- nft_trans_chain(trans) = ctx->chain;
nft_trans_commit_list_add_tail(ctx->net, trans);
return trans;
@@ -2089,18 +2089,19 @@ static struct nft_stats __percpu *nft_stats_alloc(const struct nlattr *attr)
return newstats;
}
-static void nft_chain_stats_replace(struct nft_trans *trans)
+static void nft_chain_stats_replace(struct nft_trans_chain *trans)
{
- struct nft_base_chain *chain = nft_base_chain(trans->ctx.chain);
+ const struct nft_trans *t = &trans->nft_trans_binding.nft_trans;
+ struct nft_base_chain *chain = nft_base_chain(trans->chain);
- if (!nft_trans_chain_stats(trans))
+ if (!trans->stats)
return;
- nft_trans_chain_stats(trans) =
- rcu_replace_pointer(chain->stats, nft_trans_chain_stats(trans),
- lockdep_commit_lock_is_held(trans->ctx.net));
+ trans->stats =
+ rcu_replace_pointer(chain->stats, trans->stats,
+ lockdep_commit_lock_is_held(t->ctx.net));
- if (!nft_trans_chain_stats(trans))
+ if (!trans->stats)
static_branch_inc(&nft_counters_enabled);
}
@@ -9456,47 +9457,47 @@ static int nf_tables_validate(struct net *net)
*
* We defer the drop policy until the transaction has been finalized.
*/
-static void nft_chain_commit_drop_policy(struct nft_trans *trans)
+static void nft_chain_commit_drop_policy(struct nft_trans_chain *trans)
{
struct nft_base_chain *basechain;
- if (nft_trans_chain_policy(trans) != NF_DROP)
+ if (trans->policy != NF_DROP)
return;
- if (!nft_is_base_chain(trans->ctx.chain))
+ if (!nft_is_base_chain(trans->chain))
return;
- basechain = nft_base_chain(trans->ctx.chain);
+ basechain = nft_base_chain(trans->chain);
basechain->policy = NF_DROP;
}
-static void nft_chain_commit_update(struct nft_trans *trans)
+static void nft_chain_commit_update(struct nft_trans_chain *trans)
{
- struct nft_table *table = trans->ctx.table;
+ struct nft_table *table = trans->nft_trans_binding.nft_trans.ctx.table;
struct nft_base_chain *basechain;
- if (nft_trans_chain_name(trans)) {
+ if (trans->name) {
rhltable_remove(&table->chains_ht,
- &trans->ctx.chain->rhlhead,
+ &trans->chain->rhlhead,
nft_chain_ht_params);
- swap(trans->ctx.chain->name, nft_trans_chain_name(trans));
+ swap(trans->chain->name, trans->name);
rhltable_insert_key(&table->chains_ht,
- trans->ctx.chain->name,
- &trans->ctx.chain->rhlhead,
+ trans->chain->name,
+ &trans->chain->rhlhead,
nft_chain_ht_params);
}
- if (!nft_is_base_chain(trans->ctx.chain))
+ if (!nft_is_base_chain(trans->chain))
return;
nft_chain_stats_replace(trans);
- basechain = nft_base_chain(trans->ctx.chain);
+ basechain = nft_base_chain(trans->chain);
- switch (nft_trans_chain_policy(trans)) {
+ switch (trans->policy) {
case NF_DROP:
case NF_ACCEPT:
- basechain->policy = nft_trans_chain_policy(trans);
+ basechain->policy = trans->policy;
break;
}
}
@@ -10309,14 +10310,14 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
break;
case NFT_MSG_NEWCHAIN:
if (nft_trans_chain_update(trans)) {
- nft_chain_commit_update(trans);
+ nft_chain_commit_update(nft_trans_container_chain(trans));
nf_tables_chain_notify(&trans->ctx, NFT_MSG_NEWCHAIN,
&nft_trans_chain_hooks(trans));
list_splice(&nft_trans_chain_hooks(trans),
&nft_trans_basechain(trans)->hook_list);
/* trans destroyed after rcu grace period */
} else {
- nft_chain_commit_drop_policy(trans);
+ nft_chain_commit_drop_policy(nft_trans_container_chain(trans));
nft_clear(net, trans->ctx.chain);
nf_tables_chain_notify(&trans->ctx, NFT_MSG_NEWCHAIN, NULL);
nft_trans_destroy(trans);
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 07/19] netfilter: nf_tables: avoid usage of embedded nft_ctx
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (5 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 06/19] netfilter: nf_tables: pass more specific nft_trans_chain where possible Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 08/19] netfilter: nf_tables: store chain pointer in rule transaction Pablo Neira Ayuso
` (12 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
nft_ctx is stored in nft_trans object, but nft_ctx is large
(48 bytes on 64-bit platforms), it should not be embedded in
the transaction structures.
Reduce its usage so we can remove it eventually.
This replaces trans->ctx.chain with the chain pointer
already available in nft_trans_chain structure.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nf_tables_api.c | 10 +++++-----
net/netfilter/nf_tables_offload.c | 16 ++++++++--------
2 files changed, 13 insertions(+), 13 deletions(-)
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 62a4da955574..f4e39816104f 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -9682,10 +9682,10 @@ static void nf_tables_commit_chain_prepare_cancel(struct net *net)
struct nft_trans *trans, *next;
list_for_each_entry_safe(trans, next, &nft_net->commit_list, list) {
- struct nft_chain *chain = trans->ctx.chain;
-
if (trans->msg_type == NFT_MSG_NEWRULE ||
trans->msg_type == NFT_MSG_DELRULE) {
+ struct nft_chain *chain = trans->ctx.chain;
+
kvfree(chain->blob_next);
chain->blob_next = NULL;
}
@@ -10318,7 +10318,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
/* trans destroyed after rcu grace period */
} else {
nft_chain_commit_drop_policy(nft_trans_container_chain(trans));
- nft_clear(net, trans->ctx.chain);
+ nft_clear(net, nft_trans_chain(trans));
nf_tables_chain_notify(&trans->ctx, NFT_MSG_NEWCHAIN, NULL);
nft_trans_destroy(trans);
}
@@ -10334,11 +10334,11 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
true);
}
} else {
- nft_chain_del(trans->ctx.chain);
+ nft_chain_del(nft_trans_chain(trans));
nf_tables_chain_notify(&trans->ctx, NFT_MSG_DELCHAIN,
NULL);
nf_tables_unregister_hook(trans->ctx.net, table,
- trans->ctx.chain);
+ nft_trans_chain(trans));
}
break;
case NFT_MSG_NEWRULE:
diff --git a/net/netfilter/nf_tables_offload.c b/net/netfilter/nf_tables_offload.c
index 12ab78fa5d84..8d892a0d2438 100644
--- a/net/netfilter/nf_tables_offload.c
+++ b/net/netfilter/nf_tables_offload.c
@@ -518,18 +518,18 @@ static void nft_flow_rule_offload_abort(struct net *net,
switch (trans->msg_type) {
case NFT_MSG_NEWCHAIN:
- if (!(trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD) ||
+ if (!(nft_trans_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD) ||
nft_trans_chain_update(trans))
continue;
- err = nft_flow_offload_chain(trans->ctx.chain, NULL,
+ err = nft_flow_offload_chain(nft_trans_chain(trans), NULL,
FLOW_BLOCK_UNBIND);
break;
case NFT_MSG_DELCHAIN:
- if (!(trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD))
+ if (!(nft_trans_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD))
continue;
- err = nft_flow_offload_chain(trans->ctx.chain, NULL,
+ err = nft_flow_offload_chain(nft_trans_chain(trans), NULL,
FLOW_BLOCK_BIND);
break;
case NFT_MSG_NEWRULE:
@@ -569,20 +569,20 @@ int nft_flow_rule_offload_commit(struct net *net)
switch (trans->msg_type) {
case NFT_MSG_NEWCHAIN:
- if (!(trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD) ||
+ if (!(nft_trans_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD) ||
nft_trans_chain_update(trans))
continue;
policy = nft_trans_chain_policy(trans);
- err = nft_flow_offload_chain(trans->ctx.chain, &policy,
+ err = nft_flow_offload_chain(nft_trans_chain(trans), &policy,
FLOW_BLOCK_BIND);
break;
case NFT_MSG_DELCHAIN:
- if (!(trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD))
+ if (!(nft_trans_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD))
continue;
policy = nft_trans_chain_policy(trans);
- err = nft_flow_offload_chain(trans->ctx.chain, &policy,
+ err = nft_flow_offload_chain(nft_trans_chain(trans), &policy,
FLOW_BLOCK_UNBIND);
break;
case NFT_MSG_NEWRULE:
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 08/19] netfilter: nf_tables: store chain pointer in rule transaction
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (6 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 07/19] netfilter: nf_tables: avoid usage of embedded nft_ctx Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 09/19] netfilter: nf_tables: reduce trans->ctx.chain references Pablo Neira Ayuso
` (11 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
Currently the chain can be derived from trans->ctx.chain, but
the ctx will go away soon.
Thus add the chain pointer to nft_trans_rule structure itself.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
include/net/netfilter/nf_tables.h | 3 +++
net/netfilter/nf_tables_api.c | 21 +++++++++++----------
net/netfilter/nf_tables_offload.c | 16 ++++++++--------
3 files changed, 22 insertions(+), 18 deletions(-)
diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index 328fdc140551..86e6bd63a205 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -1640,6 +1640,7 @@ struct nft_trans_binding {
struct nft_trans_rule {
struct nft_trans nft_trans;
struct nft_rule *rule;
+ struct nft_chain *chain;
struct nft_flow_rule *flow;
u32 rule_id;
bool bound;
@@ -1655,6 +1656,8 @@ struct nft_trans_rule {
nft_trans_container_rule(trans)->rule_id
#define nft_trans_rule_bound(trans) \
nft_trans_container_rule(trans)->bound
+#define nft_trans_rule_chain(trans) \
+ nft_trans_container_rule(trans)->chain
struct nft_trans_set {
struct nft_trans_binding nft_trans_binding;
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index f4e39816104f..3e5980f0bf71 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -251,7 +251,7 @@ static void __nft_chain_trans_bind(const struct nft_ctx *ctx,
nft_trans_chain_bound(trans) = bind;
break;
case NFT_MSG_NEWRULE:
- if (trans->ctx.chain == chain)
+ if (nft_trans_rule_chain(trans) == chain)
nft_trans_rule_bound(trans) = bind;
break;
}
@@ -541,6 +541,7 @@ static struct nft_trans *nft_trans_rule_add(struct nft_ctx *ctx, int msg_type,
ntohl(nla_get_be32(ctx->nla[NFTA_RULE_ID]));
}
nft_trans_rule(trans) = rule;
+ nft_trans_rule_chain(trans) = ctx->chain;
nft_trans_commit_list_add_tail(ctx->net, trans);
return trans;
@@ -4227,7 +4228,7 @@ static struct nft_rule *nft_rule_lookup_byid(const struct net *net,
list_for_each_entry(trans, &nft_net->commit_list, list) {
if (trans->msg_type == NFT_MSG_NEWRULE &&
- trans->ctx.chain == chain &&
+ nft_trans_rule_chain(trans) == chain &&
id == nft_trans_rule_id(trans))
return nft_trans_rule(trans);
}
@@ -9684,7 +9685,7 @@ static void nf_tables_commit_chain_prepare_cancel(struct net *net)
list_for_each_entry_safe(trans, next, &nft_net->commit_list, list) {
if (trans->msg_type == NFT_MSG_NEWRULE ||
trans->msg_type == NFT_MSG_DELRULE) {
- struct nft_chain *chain = trans->ctx.chain;
+ struct nft_chain *chain = nft_trans_rule_chain(trans);
kvfree(chain->blob_next);
chain->blob_next = NULL;
@@ -10250,7 +10251,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
}
if (trans->msg_type == NFT_MSG_NEWRULE ||
trans->msg_type == NFT_MSG_DELRULE) {
- chain = trans->ctx.chain;
+ chain = nft_trans_rule_chain(trans);
ret = nf_tables_commit_chain_prepare(net, chain);
if (ret < 0) {
@@ -10346,7 +10347,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
nf_tables_rule_notify(&trans->ctx,
nft_trans_rule(trans),
NFT_MSG_NEWRULE);
- if (trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD)
+ if (nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD)
nft_flow_rule_destroy(nft_trans_flow_rule(trans));
nft_trans_destroy(trans);
@@ -10361,7 +10362,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
nft_trans_rule(trans),
NFT_TRANS_COMMIT);
- if (trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD)
+ if (nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD)
nft_flow_rule_destroy(nft_trans_flow_rule(trans));
break;
case NFT_MSG_NEWSET:
@@ -10645,20 +10646,20 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
nft_trans_destroy(trans);
break;
}
- nft_use_dec_restore(&trans->ctx.chain->use);
+ nft_use_dec_restore(&nft_trans_rule_chain(trans)->use);
list_del_rcu(&nft_trans_rule(trans)->list);
nft_rule_expr_deactivate(&trans->ctx,
nft_trans_rule(trans),
NFT_TRANS_ABORT);
- if (trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD)
+ if (nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD)
nft_flow_rule_destroy(nft_trans_flow_rule(trans));
break;
case NFT_MSG_DELRULE:
case NFT_MSG_DESTROYRULE:
- nft_use_inc_restore(&trans->ctx.chain->use);
+ nft_use_inc_restore(&nft_trans_rule_chain(trans)->use);
nft_clear(trans->ctx.net, nft_trans_rule(trans));
nft_rule_expr_activate(&trans->ctx, nft_trans_rule(trans));
- if (trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD)
+ if (nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD)
nft_flow_rule_destroy(nft_trans_flow_rule(trans));
nft_trans_destroy(trans);
diff --git a/net/netfilter/nf_tables_offload.c b/net/netfilter/nf_tables_offload.c
index 8d892a0d2438..0619feb10abb 100644
--- a/net/netfilter/nf_tables_offload.c
+++ b/net/netfilter/nf_tables_offload.c
@@ -533,18 +533,18 @@ static void nft_flow_rule_offload_abort(struct net *net,
FLOW_BLOCK_BIND);
break;
case NFT_MSG_NEWRULE:
- if (!(trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD))
+ if (!(nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD))
continue;
- err = nft_flow_offload_rule(trans->ctx.chain,
+ err = nft_flow_offload_rule(nft_trans_rule_chain(trans),
nft_trans_rule(trans),
NULL, FLOW_CLS_DESTROY);
break;
case NFT_MSG_DELRULE:
- if (!(trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD))
+ if (!(nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD))
continue;
- err = nft_flow_offload_rule(trans->ctx.chain,
+ err = nft_flow_offload_rule(nft_trans_rule_chain(trans),
nft_trans_rule(trans),
nft_trans_flow_rule(trans),
FLOW_CLS_REPLACE);
@@ -586,7 +586,7 @@ int nft_flow_rule_offload_commit(struct net *net)
FLOW_BLOCK_UNBIND);
break;
case NFT_MSG_NEWRULE:
- if (!(trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD))
+ if (!(nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD))
continue;
if (trans->ctx.flags & NLM_F_REPLACE ||
@@ -594,16 +594,16 @@ int nft_flow_rule_offload_commit(struct net *net)
err = -EOPNOTSUPP;
break;
}
- err = nft_flow_offload_rule(trans->ctx.chain,
+ err = nft_flow_offload_rule(nft_trans_rule_chain(trans),
nft_trans_rule(trans),
nft_trans_flow_rule(trans),
FLOW_CLS_REPLACE);
break;
case NFT_MSG_DELRULE:
- if (!(trans->ctx.chain->flags & NFT_CHAIN_HW_OFFLOAD))
+ if (!(nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD))
continue;
- err = nft_flow_offload_rule(trans->ctx.chain,
+ err = nft_flow_offload_rule(nft_trans_rule_chain(trans),
nft_trans_rule(trans),
NULL, FLOW_CLS_DESTROY);
break;
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 09/19] netfilter: nf_tables: reduce trans->ctx.chain references
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (7 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 08/19] netfilter: nf_tables: store chain pointer in rule transaction Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 10/19] netfilter: nf_tables: pass nft_table to destroy function Pablo Neira Ayuso
` (10 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
These objects are the trans_chain subtype, so use the helper instead
of referencing trans->ctx, which will be removed soon.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nf_tables_api.c | 16 +++++++---------
1 file changed, 7 insertions(+), 9 deletions(-)
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 3e5980f0bf71..bd311b37fc61 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -1262,7 +1262,7 @@ static bool nft_table_pending_update(const struct nft_ctx *ctx)
((trans->msg_type == NFT_MSG_NEWCHAIN &&
nft_trans_chain_update(trans)) ||
(trans->msg_type == NFT_MSG_DELCHAIN &&
- nft_is_base_chain(trans->ctx.chain))))
+ nft_is_base_chain(nft_trans_chain(trans)))))
return true;
}
@@ -2815,13 +2815,11 @@ static struct nft_chain *nft_chain_lookup_byid(const struct net *net,
struct nft_trans *trans;
list_for_each_entry(trans, &nft_net->commit_list, list) {
- struct nft_chain *chain = trans->ctx.chain;
-
if (trans->msg_type == NFT_MSG_NEWCHAIN &&
- chain->table == table &&
+ nft_trans_chain(trans)->table == table &&
id == nft_trans_chain_id(trans) &&
- nft_active_genmask(chain, genmask))
- return chain;
+ nft_active_genmask(nft_trans_chain(trans), genmask))
+ return nft_trans_chain(trans);
}
return ERR_PTR(-ENOENT);
}
@@ -10625,9 +10623,9 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
break;
}
nft_use_dec_restore(&table->use);
- nft_chain_del(trans->ctx.chain);
+ nft_chain_del(nft_trans_chain(trans));
nf_tables_unregister_hook(trans->ctx.net, table,
- trans->ctx.chain);
+ nft_trans_chain(trans));
}
break;
case NFT_MSG_DELCHAIN:
@@ -10637,7 +10635,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
&nft_trans_basechain(trans)->hook_list);
} else {
nft_use_inc_restore(&table->use);
- nft_clear(trans->ctx.net, trans->ctx.chain);
+ nft_clear(trans->ctx.net, nft_trans_chain(trans));
}
nft_trans_destroy(trans);
break;
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 10/19] netfilter: nf_tables: pass nft_table to destroy function
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (8 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 09/19] netfilter: nf_tables: reduce trans->ctx.chain references Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 11/19] netfilter: nf_tables: do not store nft_ctx in transaction objects Pablo Neira Ayuso
` (9 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
No functional change intended.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nf_tables_api.c | 18 +++++++++---------
1 file changed, 9 insertions(+), 9 deletions(-)
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index bd311b37fc61..6958f922f95a 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -1656,15 +1656,15 @@ static int nf_tables_deltable(struct sk_buff *skb, const struct nfnl_info *info,
return nft_flush_table(&ctx);
}
-static void nf_tables_table_destroy(struct nft_ctx *ctx)
+static void nf_tables_table_destroy(struct nft_table *table)
{
- if (WARN_ON(ctx->table->use > 0))
+ if (WARN_ON(table->use > 0))
return;
- rhltable_destroy(&ctx->table->chains_ht);
- kfree(ctx->table->name);
- kfree(ctx->table->udata);
- kfree(ctx->table);
+ rhltable_destroy(&table->chains_ht);
+ kfree(table->name);
+ kfree(table->udata);
+ kfree(table);
}
void nft_register_chain_type(const struct nft_chain_type *ctype)
@@ -9521,7 +9521,7 @@ static void nft_commit_release(struct nft_trans *trans)
switch (trans->msg_type) {
case NFT_MSG_DELTABLE:
case NFT_MSG_DESTROYTABLE:
- nf_tables_table_destroy(&trans->ctx);
+ nf_tables_table_destroy(trans->ctx.table);
break;
case NFT_MSG_NEWCHAIN:
free_percpu(nft_trans_chain_stats(trans));
@@ -10518,7 +10518,7 @@ static void nf_tables_abort_release(struct nft_trans *trans)
{
switch (trans->msg_type) {
case NFT_MSG_NEWTABLE:
- nf_tables_table_destroy(&trans->ctx);
+ nf_tables_table_destroy(trans->ctx.table);
break;
case NFT_MSG_NEWCHAIN:
if (nft_trans_chain_update(trans))
@@ -11490,7 +11490,7 @@ static void __nft_release_table(struct net *net, struct nft_table *table)
nft_use_dec(&table->use);
nf_tables_chain_destroy(chain);
}
- nf_tables_table_destroy(&ctx);
+ nf_tables_table_destroy(table);
}
static void __nft_release_tables(struct net *net)
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 11/19] netfilter: nf_tables: do not store nft_ctx in transaction objects
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (9 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 10/19] netfilter: nf_tables: pass nft_table to destroy function Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 12/19] ipvs: Avoid unnecessary calls to skb_is_gso_sctp Pablo Neira Ayuso
` (8 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
nft_ctx is huge and most of the information stored within isn't used
at all.
Remove nft_ctx member from the base transaction structure and store
only what is needed.
After this change, relevant struct sizes are:
struct nft_trans_chain { /* size: 120 (-32), cachelines: 2, members: 10 */
struct nft_trans_elem { /* size: 72 (-40), cachelines: 2, members: 4 */
struct nft_trans_flowtable { /* size: 80 (-48), cachelines: 2, members: 5 */
struct nft_trans_obj { /* size: 72 (-40), cachelines: 2, members: 4 */
struct nft_trans_rule { /* size: 80 (-32), cachelines: 2, members: 6 */
struct nft_trans_set { /* size: 96 (-24), cachelines: 2, members: 8 */
struct nft_trans_table { /* size: 56 (-40), cachelines: 1, members: 2 */
struct nft_trans_elem can now be allocated from kmalloc-96 instead of
kmalloc-128 slab.
A further reduction by 8 bytes would even allow for kmalloc-64.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
include/net/netfilter/nf_tables.h | 43 ++++++++-
net/netfilter/nf_tables_api.c | 140 +++++++++++++++++-------------
net/netfilter/nf_tables_offload.c | 8 +-
3 files changed, 125 insertions(+), 66 deletions(-)
diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index 86e6bd63a205..1e8da1b882ac 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -1611,18 +1611,26 @@ static inline int nft_set_elem_is_dead(const struct nft_set_ext *ext)
* struct nft_trans - nf_tables object update in transaction
*
* @list: used internally
+ * @net: struct net
+ * @table: struct nft_table the object resides in
* @msg_type: message type
- * @put_net: ctx->net needs to be put
- * @ctx: transaction context
+ * @seq: netlink sequence number
+ * @flags: modifiers to new request
+ * @report: notify via unicast netlink message
+ * @put_net: net needs to be put
*
* This is the information common to all objects in the transaction,
* this must always be the first member of derived sub-types.
*/
struct nft_trans {
struct list_head list;
+ struct net *net;
+ struct nft_table *table;
int msg_type;
- bool put_net;
- struct nft_ctx ctx;
+ u32 seq;
+ u16 flags;
+ u8 report:1;
+ u8 put_net:1;
};
/**
@@ -1794,6 +1802,33 @@ struct nft_trans_gc {
struct rcu_head rcu;
};
+static inline void nft_ctx_update(struct nft_ctx *ctx,
+ const struct nft_trans *trans)
+{
+ switch (trans->msg_type) {
+ case NFT_MSG_NEWRULE:
+ case NFT_MSG_DELRULE:
+ case NFT_MSG_DESTROYRULE:
+ ctx->chain = nft_trans_rule_chain(trans);
+ break;
+ case NFT_MSG_NEWCHAIN:
+ case NFT_MSG_DELCHAIN:
+ case NFT_MSG_DESTROYCHAIN:
+ ctx->chain = nft_trans_chain(trans);
+ break;
+ default:
+ ctx->chain = NULL;
+ break;
+ }
+
+ ctx->net = trans->net;
+ ctx->table = trans->table;
+ ctx->family = trans->table->family;
+ ctx->report = trans->report;
+ ctx->flags = trans->flags;
+ ctx->seq = trans->seq;
+}
+
struct nft_trans_gc *nft_trans_gc_alloc(struct nft_set *set,
unsigned int gc_seq, gfp_t gfp);
void nft_trans_gc_destroy(struct nft_trans_gc *trans);
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 6958f922f95a..02d75aefaa8e 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -159,7 +159,12 @@ static struct nft_trans *nft_trans_alloc_gfp(const struct nft_ctx *ctx,
INIT_LIST_HEAD(&trans->list);
trans->msg_type = msg_type;
- trans->ctx = *ctx;
+
+ trans->net = ctx->net;
+ trans->table = ctx->table;
+ trans->seq = ctx->seq;
+ trans->flags = ctx->flags;
+ trans->report = ctx->report;
return trans;
}
@@ -1258,7 +1263,7 @@ static bool nft_table_pending_update(const struct nft_ctx *ctx)
return true;
list_for_each_entry(trans, &nft_net->commit_list, list) {
- if (trans->ctx.table == ctx->table &&
+ if (trans->table == ctx->table &&
((trans->msg_type == NFT_MSG_NEWCHAIN &&
nft_trans_chain_update(trans)) ||
(trans->msg_type == NFT_MSG_DELCHAIN &&
@@ -2100,7 +2105,7 @@ static void nft_chain_stats_replace(struct nft_trans_chain *trans)
trans->stats =
rcu_replace_pointer(chain->stats, trans->stats,
- lockdep_commit_lock_is_held(t->ctx.net));
+ lockdep_commit_lock_is_held(t->net));
if (!trans->stats)
static_branch_inc(&nft_counters_enabled);
@@ -2766,7 +2771,7 @@ static int nf_tables_updchain(struct nft_ctx *ctx, u8 genmask, u8 policy,
err = -EEXIST;
list_for_each_entry(tmp, &nft_net->commit_list, list) {
if (tmp->msg_type == NFT_MSG_NEWCHAIN &&
- tmp->ctx.table == table &&
+ tmp->table == table &&
nft_trans_chain_update(tmp) &&
nft_trans_chain_name(tmp) &&
strcmp(name, nft_trans_chain_name(tmp)) == 0) {
@@ -9472,7 +9477,7 @@ static void nft_chain_commit_drop_policy(struct nft_trans_chain *trans)
static void nft_chain_commit_update(struct nft_trans_chain *trans)
{
- struct nft_table *table = trans->nft_trans_binding.nft_trans.ctx.table;
+ struct nft_table *table = trans->nft_trans_binding.nft_trans.table;
struct nft_base_chain *basechain;
if (trans->name) {
@@ -9501,7 +9506,8 @@ static void nft_chain_commit_update(struct nft_trans_chain *trans)
}
}
-static void nft_obj_commit_update(struct nft_trans *trans)
+static void nft_obj_commit_update(const struct nft_ctx *ctx,
+ struct nft_trans *trans)
{
struct nft_object *newobj;
struct nft_object *obj;
@@ -9513,15 +9519,21 @@ static void nft_obj_commit_update(struct nft_trans *trans)
return;
obj->ops->update(obj, newobj);
- nft_obj_destroy(&trans->ctx, newobj);
+ nft_obj_destroy(ctx, newobj);
}
static void nft_commit_release(struct nft_trans *trans)
{
+ struct nft_ctx ctx = {
+ .net = trans->net,
+ };
+
+ nft_ctx_update(&ctx, trans);
+
switch (trans->msg_type) {
case NFT_MSG_DELTABLE:
case NFT_MSG_DESTROYTABLE:
- nf_tables_table_destroy(trans->ctx.table);
+ nf_tables_table_destroy(trans->table);
break;
case NFT_MSG_NEWCHAIN:
free_percpu(nft_trans_chain_stats(trans));
@@ -9536,21 +9548,21 @@ static void nft_commit_release(struct nft_trans *trans)
break;
case NFT_MSG_DELRULE:
case NFT_MSG_DESTROYRULE:
- nf_tables_rule_destroy(&trans->ctx, nft_trans_rule(trans));
+ nf_tables_rule_destroy(&ctx, nft_trans_rule(trans));
break;
case NFT_MSG_DELSET:
case NFT_MSG_DESTROYSET:
- nft_set_destroy(&trans->ctx, nft_trans_set(trans));
+ nft_set_destroy(&ctx, nft_trans_set(trans));
break;
case NFT_MSG_DELSETELEM:
case NFT_MSG_DESTROYSETELEM:
- nf_tables_set_elem_destroy(&trans->ctx,
+ nf_tables_set_elem_destroy(&ctx,
nft_trans_elem_set(trans),
nft_trans_elem_priv(trans));
break;
case NFT_MSG_DELOBJ:
case NFT_MSG_DESTROYOBJ:
- nft_obj_destroy(&trans->ctx, nft_trans_obj(trans));
+ nft_obj_destroy(&ctx, nft_trans_obj(trans));
break;
case NFT_MSG_DELFLOWTABLE:
case NFT_MSG_DESTROYFLOWTABLE:
@@ -9562,7 +9574,7 @@ static void nft_commit_release(struct nft_trans *trans)
}
if (trans->put_net)
- put_net(trans->ctx.net);
+ put_net(trans->net);
kfree(trans);
}
@@ -10042,7 +10054,7 @@ static void nf_tables_commit_release(struct net *net)
trans = list_last_entry(&nft_net->commit_list,
struct nft_trans, list);
- get_net(trans->ctx.net);
+ get_net(trans->net);
WARN_ON_ONCE(trans->put_net);
trans->put_net = true;
@@ -10186,6 +10198,7 @@ static void nft_gc_seq_end(struct nftables_pernet *nft_net, unsigned int gc_seq)
static int nf_tables_commit(struct net *net, struct sk_buff *skb)
{
struct nftables_pernet *nft_net = nft_pernet(net);
+ const struct nlmsghdr *nlh = nlmsg_hdr(skb);
struct nft_trans_binding *trans_binding;
struct nft_trans *trans, *next;
unsigned int base_seq, gc_seq;
@@ -10193,6 +10206,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
struct nft_trans_elem *te;
struct nft_chain *chain;
struct nft_table *table;
+ struct nft_ctx ctx;
LIST_HEAD(adl);
int err;
@@ -10201,6 +10215,8 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
return 0;
}
+ nft_ctx_init(&ctx, net, skb, nlh, NFPROTO_UNSPEC, NULL, NULL, NULL);
+
list_for_each_entry(trans_binding, &nft_net->binding_list, binding_list) {
trans = &trans_binding->nft_trans;
switch (trans->msg_type) {
@@ -10238,7 +10254,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
/* 1. Allocate space for next generation rules_gen_X[] */
list_for_each_entry_safe(trans, next, &nft_net->commit_list, list) {
- struct nft_table *table = trans->ctx.table;
+ struct nft_table *table = trans->table;
int ret;
ret = nf_tables_commit_audit_alloc(&adl, table);
@@ -10282,7 +10298,9 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
net->nft.gencursor = nft_gencursor_next(net);
list_for_each_entry_safe(trans, next, &nft_net->commit_list, list) {
- struct nft_table *table = trans->ctx.table;
+ struct nft_table *table = trans->table;
+
+ nft_ctx_update(&ctx, trans);
nf_tables_commit_audit_collect(&adl, table, trans->msg_type);
switch (trans->msg_type) {
@@ -10299,18 +10317,18 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
} else {
nft_clear(net, table);
}
- nf_tables_table_notify(&trans->ctx, NFT_MSG_NEWTABLE);
+ nf_tables_table_notify(&ctx, NFT_MSG_NEWTABLE);
nft_trans_destroy(trans);
break;
case NFT_MSG_DELTABLE:
case NFT_MSG_DESTROYTABLE:
list_del_rcu(&table->list);
- nf_tables_table_notify(&trans->ctx, trans->msg_type);
+ nf_tables_table_notify(&ctx, trans->msg_type);
break;
case NFT_MSG_NEWCHAIN:
if (nft_trans_chain_update(trans)) {
nft_chain_commit_update(nft_trans_container_chain(trans));
- nf_tables_chain_notify(&trans->ctx, NFT_MSG_NEWCHAIN,
+ nf_tables_chain_notify(&ctx, NFT_MSG_NEWCHAIN,
&nft_trans_chain_hooks(trans));
list_splice(&nft_trans_chain_hooks(trans),
&nft_trans_basechain(trans)->hook_list);
@@ -10318,14 +10336,14 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
} else {
nft_chain_commit_drop_policy(nft_trans_container_chain(trans));
nft_clear(net, nft_trans_chain(trans));
- nf_tables_chain_notify(&trans->ctx, NFT_MSG_NEWCHAIN, NULL);
+ nf_tables_chain_notify(&ctx, NFT_MSG_NEWCHAIN, NULL);
nft_trans_destroy(trans);
}
break;
case NFT_MSG_DELCHAIN:
case NFT_MSG_DESTROYCHAIN:
if (nft_trans_chain_update(trans)) {
- nf_tables_chain_notify(&trans->ctx, NFT_MSG_DELCHAIN,
+ nf_tables_chain_notify(&ctx, NFT_MSG_DELCHAIN,
&nft_trans_chain_hooks(trans));
if (!(table->flags & NFT_TABLE_F_DORMANT)) {
nft_netdev_unregister_hooks(net,
@@ -10334,16 +10352,15 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
}
} else {
nft_chain_del(nft_trans_chain(trans));
- nf_tables_chain_notify(&trans->ctx, NFT_MSG_DELCHAIN,
+ nf_tables_chain_notify(&ctx, NFT_MSG_DELCHAIN,
NULL);
- nf_tables_unregister_hook(trans->ctx.net, table,
+ nf_tables_unregister_hook(ctx.net, ctx.table,
nft_trans_chain(trans));
}
break;
case NFT_MSG_NEWRULE:
- nft_clear(trans->ctx.net, nft_trans_rule(trans));
- nf_tables_rule_notify(&trans->ctx,
- nft_trans_rule(trans),
+ nft_clear(net, nft_trans_rule(trans));
+ nf_tables_rule_notify(&ctx, nft_trans_rule(trans),
NFT_MSG_NEWRULE);
if (nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD)
nft_flow_rule_destroy(nft_trans_flow_rule(trans));
@@ -10353,11 +10370,9 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
case NFT_MSG_DELRULE:
case NFT_MSG_DESTROYRULE:
list_del_rcu(&nft_trans_rule(trans)->list);
- nf_tables_rule_notify(&trans->ctx,
- nft_trans_rule(trans),
+ nf_tables_rule_notify(&ctx, nft_trans_rule(trans),
trans->msg_type);
- nft_rule_expr_deactivate(&trans->ctx,
- nft_trans_rule(trans),
+ nft_rule_expr_deactivate(&ctx, nft_trans_rule(trans),
NFT_TRANS_COMMIT);
if (nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD)
@@ -10381,7 +10396,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
!list_empty(&nft_trans_set(trans)->bindings))
nft_use_dec(&table->use);
}
- nf_tables_set_notify(&trans->ctx, nft_trans_set(trans),
+ nf_tables_set_notify(&ctx, nft_trans_set(trans),
NFT_MSG_NEWSET, GFP_KERNEL);
nft_trans_destroy(trans);
break;
@@ -10389,14 +10404,14 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
case NFT_MSG_DESTROYSET:
nft_trans_set(trans)->dead = 1;
list_del_rcu(&nft_trans_set(trans)->list);
- nf_tables_set_notify(&trans->ctx, nft_trans_set(trans),
+ nf_tables_set_notify(&ctx, nft_trans_set(trans),
trans->msg_type, GFP_KERNEL);
break;
case NFT_MSG_NEWSETELEM:
te = nft_trans_container_elem(trans);
nft_setelem_activate(net, te->set, te->elem_priv);
- nf_tables_setelem_notify(&trans->ctx, te->set,
+ nf_tables_setelem_notify(&ctx, te->set,
te->elem_priv,
NFT_MSG_NEWSETELEM);
if (te->set->ops->commit &&
@@ -10410,7 +10425,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
case NFT_MSG_DESTROYSETELEM:
te = nft_trans_container_elem(trans);
- nf_tables_setelem_notify(&trans->ctx, te->set,
+ nf_tables_setelem_notify(&ctx, te->set,
te->elem_priv,
trans->msg_type);
nft_setelem_remove(net, te->set, te->elem_priv);
@@ -10426,13 +10441,13 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
break;
case NFT_MSG_NEWOBJ:
if (nft_trans_obj_update(trans)) {
- nft_obj_commit_update(trans);
- nf_tables_obj_notify(&trans->ctx,
+ nft_obj_commit_update(&ctx, trans);
+ nf_tables_obj_notify(&ctx,
nft_trans_obj(trans),
NFT_MSG_NEWOBJ);
} else {
nft_clear(net, nft_trans_obj(trans));
- nf_tables_obj_notify(&trans->ctx,
+ nf_tables_obj_notify(&ctx,
nft_trans_obj(trans),
NFT_MSG_NEWOBJ);
nft_trans_destroy(trans);
@@ -10441,14 +10456,14 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
case NFT_MSG_DELOBJ:
case NFT_MSG_DESTROYOBJ:
nft_obj_del(nft_trans_obj(trans));
- nf_tables_obj_notify(&trans->ctx, nft_trans_obj(trans),
+ nf_tables_obj_notify(&ctx, nft_trans_obj(trans),
trans->msg_type);
break;
case NFT_MSG_NEWFLOWTABLE:
if (nft_trans_flowtable_update(trans)) {
nft_trans_flowtable(trans)->data.flags =
nft_trans_flowtable_flags(trans);
- nf_tables_flowtable_notify(&trans->ctx,
+ nf_tables_flowtable_notify(&ctx,
nft_trans_flowtable(trans),
&nft_trans_flowtable_hooks(trans),
NFT_MSG_NEWFLOWTABLE);
@@ -10456,7 +10471,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
&nft_trans_flowtable(trans)->hook_list);
} else {
nft_clear(net, nft_trans_flowtable(trans));
- nf_tables_flowtable_notify(&trans->ctx,
+ nf_tables_flowtable_notify(&ctx,
nft_trans_flowtable(trans),
NULL,
NFT_MSG_NEWFLOWTABLE);
@@ -10466,7 +10481,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
case NFT_MSG_DELFLOWTABLE:
case NFT_MSG_DESTROYFLOWTABLE:
if (nft_trans_flowtable_update(trans)) {
- nf_tables_flowtable_notify(&trans->ctx,
+ nf_tables_flowtable_notify(&ctx,
nft_trans_flowtable(trans),
&nft_trans_flowtable_hooks(trans),
trans->msg_type);
@@ -10474,7 +10489,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
&nft_trans_flowtable_hooks(trans));
} else {
list_del_rcu(&nft_trans_flowtable(trans)->list);
- nf_tables_flowtable_notify(&trans->ctx,
+ nf_tables_flowtable_notify(&ctx,
nft_trans_flowtable(trans),
NULL,
trans->msg_type);
@@ -10516,9 +10531,13 @@ static void nf_tables_module_autoload(struct net *net)
static void nf_tables_abort_release(struct nft_trans *trans)
{
+ struct nft_ctx ctx = { };
+
+ nft_ctx_update(&ctx, trans);
+
switch (trans->msg_type) {
case NFT_MSG_NEWTABLE:
- nf_tables_table_destroy(trans->ctx.table);
+ nf_tables_table_destroy(trans->table);
break;
case NFT_MSG_NEWCHAIN:
if (nft_trans_chain_update(trans))
@@ -10527,17 +10546,17 @@ static void nf_tables_abort_release(struct nft_trans *trans)
nf_tables_chain_destroy(nft_trans_chain(trans));
break;
case NFT_MSG_NEWRULE:
- nf_tables_rule_destroy(&trans->ctx, nft_trans_rule(trans));
+ nf_tables_rule_destroy(&ctx, nft_trans_rule(trans));
break;
case NFT_MSG_NEWSET:
- nft_set_destroy(&trans->ctx, nft_trans_set(trans));
+ nft_set_destroy(&ctx, nft_trans_set(trans));
break;
case NFT_MSG_NEWSETELEM:
nft_set_elem_destroy(nft_trans_elem_set(trans),
nft_trans_elem_priv(trans), true);
break;
case NFT_MSG_NEWOBJ:
- nft_obj_destroy(&trans->ctx, nft_trans_obj(trans));
+ nft_obj_destroy(&ctx, nft_trans_obj(trans));
break;
case NFT_MSG_NEWFLOWTABLE:
if (nft_trans_flowtable_update(trans))
@@ -10569,6 +10588,9 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
struct nft_trans *trans, *next;
LIST_HEAD(set_update_list);
struct nft_trans_elem *te;
+ struct nft_ctx ctx = {
+ .net = net,
+ };
int err = 0;
if (action == NFNL_ABORT_VALIDATE &&
@@ -10577,7 +10599,9 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
list_for_each_entry_safe_reverse(trans, next, &nft_net->commit_list,
list) {
- struct nft_table *table = trans->ctx.table;
+ struct nft_table *table = trans->table;
+
+ nft_ctx_update(&ctx, trans);
switch (trans->msg_type) {
case NFT_MSG_NEWTABLE:
@@ -10604,7 +10628,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
break;
case NFT_MSG_DELTABLE:
case NFT_MSG_DESTROYTABLE:
- nft_clear(trans->ctx.net, table);
+ nft_clear(trans->net, table);
nft_trans_destroy(trans);
break;
case NFT_MSG_NEWCHAIN:
@@ -10624,7 +10648,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
}
nft_use_dec_restore(&table->use);
nft_chain_del(nft_trans_chain(trans));
- nf_tables_unregister_hook(trans->ctx.net, table,
+ nf_tables_unregister_hook(trans->net, table,
nft_trans_chain(trans));
}
break;
@@ -10635,7 +10659,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
&nft_trans_basechain(trans)->hook_list);
} else {
nft_use_inc_restore(&table->use);
- nft_clear(trans->ctx.net, nft_trans_chain(trans));
+ nft_clear(trans->net, nft_trans_chain(trans));
}
nft_trans_destroy(trans);
break;
@@ -10646,7 +10670,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
}
nft_use_dec_restore(&nft_trans_rule_chain(trans)->use);
list_del_rcu(&nft_trans_rule(trans)->list);
- nft_rule_expr_deactivate(&trans->ctx,
+ nft_rule_expr_deactivate(&ctx,
nft_trans_rule(trans),
NFT_TRANS_ABORT);
if (nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD)
@@ -10655,8 +10679,8 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
case NFT_MSG_DELRULE:
case NFT_MSG_DESTROYRULE:
nft_use_inc_restore(&nft_trans_rule_chain(trans)->use);
- nft_clear(trans->ctx.net, nft_trans_rule(trans));
- nft_rule_expr_activate(&trans->ctx, nft_trans_rule(trans));
+ nft_clear(trans->net, nft_trans_rule(trans));
+ nft_rule_expr_activate(&ctx, nft_trans_rule(trans));
if (nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD)
nft_flow_rule_destroy(nft_trans_flow_rule(trans));
@@ -10678,9 +10702,9 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
case NFT_MSG_DELSET:
case NFT_MSG_DESTROYSET:
nft_use_inc_restore(&table->use);
- nft_clear(trans->ctx.net, nft_trans_set(trans));
+ nft_clear(trans->net, nft_trans_set(trans));
if (nft_trans_set(trans)->flags & (NFT_SET_MAP | NFT_SET_OBJECT))
- nft_map_activate(&trans->ctx, nft_trans_set(trans));
+ nft_map_activate(&ctx, nft_trans_set(trans));
nft_trans_destroy(trans);
break;
@@ -10720,7 +10744,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
break;
case NFT_MSG_NEWOBJ:
if (nft_trans_obj_update(trans)) {
- nft_obj_destroy(&trans->ctx, nft_trans_obj_newobj(trans));
+ nft_obj_destroy(&ctx, nft_trans_obj_newobj(trans));
nft_trans_destroy(trans);
} else {
nft_use_dec_restore(&table->use);
@@ -10730,7 +10754,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
case NFT_MSG_DELOBJ:
case NFT_MSG_DESTROYOBJ:
nft_use_inc_restore(&table->use);
- nft_clear(trans->ctx.net, nft_trans_obj(trans));
+ nft_clear(trans->net, nft_trans_obj(trans));
nft_trans_destroy(trans);
break;
case NFT_MSG_NEWFLOWTABLE:
@@ -10751,7 +10775,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
&nft_trans_flowtable(trans)->hook_list);
} else {
nft_use_inc_restore(&table->use);
- nft_clear(trans->ctx.net, nft_trans_flowtable(trans));
+ nft_clear(trans->net, nft_trans_flowtable(trans));
}
nft_trans_destroy(trans);
break;
diff --git a/net/netfilter/nf_tables_offload.c b/net/netfilter/nf_tables_offload.c
index 0619feb10abb..64675f1c7f29 100644
--- a/net/netfilter/nf_tables_offload.c
+++ b/net/netfilter/nf_tables_offload.c
@@ -513,7 +513,7 @@ static void nft_flow_rule_offload_abort(struct net *net,
int err = 0;
list_for_each_entry_continue_reverse(trans, &nft_net->commit_list, list) {
- if (trans->ctx.family != NFPROTO_NETDEV)
+ if (trans->table->family != NFPROTO_NETDEV)
continue;
switch (trans->msg_type) {
@@ -564,7 +564,7 @@ int nft_flow_rule_offload_commit(struct net *net)
u8 policy;
list_for_each_entry(trans, &nft_net->commit_list, list) {
- if (trans->ctx.family != NFPROTO_NETDEV)
+ if (trans->table->family != NFPROTO_NETDEV)
continue;
switch (trans->msg_type) {
@@ -589,8 +589,8 @@ int nft_flow_rule_offload_commit(struct net *net)
if (!(nft_trans_rule_chain(trans)->flags & NFT_CHAIN_HW_OFFLOAD))
continue;
- if (trans->ctx.flags & NLM_F_REPLACE ||
- !(trans->ctx.flags & NLM_F_APPEND)) {
+ if (trans->flags & NLM_F_REPLACE ||
+ !(trans->flags & NLM_F_APPEND)) {
err = -EOPNOTSUPP;
break;
}
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 12/19] ipvs: Avoid unnecessary calls to skb_is_gso_sctp
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (10 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 11/19] netfilter: nf_tables: do not store nft_ctx in transaction objects Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 13/19] netfilter: nf_conncount: fix wrong variable type Pablo Neira Ayuso
` (7 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Ismael Luceno <iluceno@suse.de>
In the context of the SCTP SNAT/DNAT handler, these calls can only
return true.
Fixes: e10d3ba4d434 ("ipvs: Fix checksumming on GSO of SCTP packets")
Signed-off-by: Ismael Luceno <iluceno@suse.de>
Acked-by: Julian Anastasov <ja@ssi.bg>
Acked-by: Simon Horman <horms@kernel.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/ipvs/ip_vs_proto_sctp.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/netfilter/ipvs/ip_vs_proto_sctp.c b/net/netfilter/ipvs/ip_vs_proto_sctp.c
index 1e689c714127..83e452916403 100644
--- a/net/netfilter/ipvs/ip_vs_proto_sctp.c
+++ b/net/netfilter/ipvs/ip_vs_proto_sctp.c
@@ -126,7 +126,7 @@ sctp_snat_handler(struct sk_buff *skb, struct ip_vs_protocol *pp,
if (sctph->source != cp->vport || payload_csum ||
skb->ip_summed == CHECKSUM_PARTIAL) {
sctph->source = cp->vport;
- if (!skb_is_gso(skb) || !skb_is_gso_sctp(skb))
+ if (!skb_is_gso(skb))
sctp_nat_csum(skb, sctph, sctphoff);
} else {
skb->ip_summed = CHECKSUM_UNNECESSARY;
@@ -175,7 +175,7 @@ sctp_dnat_handler(struct sk_buff *skb, struct ip_vs_protocol *pp,
(skb->ip_summed == CHECKSUM_PARTIAL &&
!(skb_dst(skb)->dev->features & NETIF_F_SCTP_CRC))) {
sctph->dest = cp->dport;
- if (!skb_is_gso(skb) || !skb_is_gso_sctp(skb))
+ if (!skb_is_gso(skb))
sctp_nat_csum(skb, sctph, sctphoff);
} else if (skb->ip_summed != CHECKSUM_PARTIAL) {
skb->ip_summed = CHECKSUM_UNNECESSARY;
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 13/19] netfilter: nf_conncount: fix wrong variable type
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (11 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 12/19] ipvs: Avoid unnecessary calls to skb_is_gso_sctp Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 14/19] netfilter: cttimeout: remove 'l3num' attr check Pablo Neira Ayuso
` (6 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Yunjian Wang <wangyunjian@huawei.com>
Now there is a issue is that code checks reports a warning: implicit
narrowing conversion from type 'unsigned int' to small type 'u8' (the
'keylen' variable). Fix it by removing the 'keylen' variable.
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nf_conncount.c | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/net/netfilter/nf_conncount.c b/net/netfilter/nf_conncount.c
index 8715617b02fe..34ba14e59e95 100644
--- a/net/netfilter/nf_conncount.c
+++ b/net/netfilter/nf_conncount.c
@@ -321,7 +321,6 @@ insert_tree(struct net *net,
struct nf_conncount_rb *rbconn;
struct nf_conncount_tuple *conn;
unsigned int count = 0, gc_count = 0;
- u8 keylen = data->keylen;
bool do_gc = true;
spin_lock_bh(&nf_conncount_locks[hash]);
@@ -333,7 +332,7 @@ insert_tree(struct net *net,
rbconn = rb_entry(*rbnode, struct nf_conncount_rb, node);
parent = *rbnode;
- diff = key_diff(key, rbconn->key, keylen);
+ diff = key_diff(key, rbconn->key, data->keylen);
if (diff < 0) {
rbnode = &((*rbnode)->rb_left);
} else if (diff > 0) {
@@ -378,7 +377,7 @@ insert_tree(struct net *net,
conn->tuple = *tuple;
conn->zone = *zone;
- memcpy(rbconn->key, key, sizeof(u32) * keylen);
+ memcpy(rbconn->key, key, sizeof(u32) * data->keylen);
nf_conncount_list_init(&rbconn->list);
list_add(&conn->node, &rbconn->list.head);
@@ -403,7 +402,6 @@ count_tree(struct net *net,
struct rb_node *parent;
struct nf_conncount_rb *rbconn;
unsigned int hash;
- u8 keylen = data->keylen;
hash = jhash2(key, data->keylen, conncount_rnd) % CONNCOUNT_SLOTS;
root = &data->root[hash];
@@ -414,7 +412,7 @@ count_tree(struct net *net,
rbconn = rb_entry(parent, struct nf_conncount_rb, node);
- diff = key_diff(key, rbconn->key, keylen);
+ diff = key_diff(key, rbconn->key, data->keylen);
if (diff < 0) {
parent = rcu_dereference_raw(parent->rb_left);
} else if (diff > 0) {
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 14/19] netfilter: cttimeout: remove 'l3num' attr check
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (12 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 13/19] netfilter: nf_conncount: fix wrong variable type Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 15/19] netfilter: nf_tables: rise cap on SELinux secmark context Pablo Neira Ayuso
` (5 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Lin Ma <linma@zju.edu.cn>
After commit dd2934a95701 ("netfilter: conntrack: remove l3->l4 mapping
information"), the attribute of type `CTA_TIMEOUT_L3PROTO` is not used
any more in function cttimeout_default_set.
However, the previous commit ea9cf2a55a7b ("netfilter: cttimeout: remove
set but not used variable 'l3num'") forgot to remove the attribute
present check when removing the related variable.
This commit removes that check to ensure consistency.
Signed-off-by: Lin Ma <linma@zju.edu.cn>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nfnetlink_cttimeout.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/net/netfilter/nfnetlink_cttimeout.c b/net/netfilter/nfnetlink_cttimeout.c
index f466af4f8531..eab4f476b47f 100644
--- a/net/netfilter/nfnetlink_cttimeout.c
+++ b/net/netfilter/nfnetlink_cttimeout.c
@@ -366,8 +366,7 @@ static int cttimeout_default_set(struct sk_buff *skb,
__u8 l4num;
int ret;
- if (!cda[CTA_TIMEOUT_L3PROTO] ||
- !cda[CTA_TIMEOUT_L4PROTO] ||
+ if (!cda[CTA_TIMEOUT_L4PROTO] ||
!cda[CTA_TIMEOUT_DATA])
return -EINVAL;
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 15/19] netfilter: nf_tables: rise cap on SELinux secmark context
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (13 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 14/19] netfilter: cttimeout: remove 'l3num' attr check Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 16/19] netfilter: nfnetlink_queue: unbreak SCTP traffic Pablo Neira Ayuso
` (4 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
secmark context is artificially limited 256 bytes, rise it to 4Kbytes.
Fixes: fb961945457f ("netfilter: nf_tables: add SECMARK support")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
include/uapi/linux/netfilter/nf_tables.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/uapi/linux/netfilter/nf_tables.h b/include/uapi/linux/netfilter/nf_tables.h
index aa4094ca2444..639894ed1b97 100644
--- a/include/uapi/linux/netfilter/nf_tables.h
+++ b/include/uapi/linux/netfilter/nf_tables.h
@@ -1376,7 +1376,7 @@ enum nft_secmark_attributes {
#define NFTA_SECMARK_MAX (__NFTA_SECMARK_MAX - 1)
/* Max security context length */
-#define NFT_SECMARK_CTX_MAXLEN 256
+#define NFT_SECMARK_CTX_MAXLEN 4096
/**
* enum nft_reject_types - nf_tables reject expression reject types
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 16/19] netfilter: nfnetlink_queue: unbreak SCTP traffic
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (14 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 15/19] netfilter: nf_tables: rise cap on SELinux secmark context Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 17/19] selftests: netfilter: nft_queue.sh: sctp coverage Pablo Neira Ayuso
` (3 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Antonio Ojea <aojea@google.com>
when packet is enqueued with nfqueue and GSO is enabled, checksum
calculation has to take into account the protocol, as SCTP uses a
32 bits CRC checksum.
Enter skb_gso_segment() path in case of SCTP GSO packets because
skb_zerocopy() does not support for GSO_BY_FRAGS.
Joint work with Pablo.
Signed-off-by: Antonio Ojea <aojea@google.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/core/dev.c | 1 +
net/netfilter/nfnetlink_queue.c | 12 ++++++++++--
2 files changed, 11 insertions(+), 2 deletions(-)
diff --git a/net/core/dev.c b/net/core/dev.c
index b94fb4e63a28..6c13a3072f04 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -3386,6 +3386,7 @@ int skb_crc32c_csum_help(struct sk_buff *skb)
out:
return ret;
}
+EXPORT_SYMBOL(skb_crc32c_csum_help);
__be16 skb_network_protocol(struct sk_buff *skb, int *depth)
{
diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
index f1c31757e496..fe550cebae1e 100644
--- a/net/netfilter/nfnetlink_queue.c
+++ b/net/netfilter/nfnetlink_queue.c
@@ -540,6 +540,14 @@ static int nfqnl_put_bridge(struct nf_queue_entry *entry, struct sk_buff *skb)
return -1;
}
+static int nf_queue_checksum_help(struct sk_buff *entskb)
+{
+ if (skb_csum_is_sctp(entskb))
+ return skb_crc32c_csum_help(entskb);
+
+ return skb_checksum_help(entskb);
+}
+
static struct sk_buff *
nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue,
struct nf_queue_entry *entry,
@@ -602,7 +610,7 @@ nfqnl_build_packet_message(struct net *net, struct nfqnl_instance *queue,
case NFQNL_COPY_PACKET:
if (!(queue->flags & NFQA_CFG_F_GSO) &&
entskb->ip_summed == CHECKSUM_PARTIAL &&
- skb_checksum_help(entskb))
+ nf_queue_checksum_help(entskb))
return NULL;
data_len = READ_ONCE(queue->copy_range);
@@ -983,7 +991,7 @@ nfqnl_enqueue_packet(struct nf_queue_entry *entry, unsigned int queuenum)
break;
}
- if ((queue->flags & NFQA_CFG_F_GSO) || !skb_is_gso(skb))
+ if (!skb_is_gso(skb) || ((queue->flags & NFQA_CFG_F_GSO) && !skb_is_gso_sctp(skb)))
return __nfqnl_enqueue_packet(net, queue, entry);
nf_bridge_adjust_skb_data(skb);
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 17/19] selftests: netfilter: nft_queue.sh: sctp coverage
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (15 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 16/19] netfilter: nfnetlink_queue: unbreak SCTP traffic Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 18/19] selftests: netfilter: nft_queue.sh: add test for disappearing listener Pablo Neira Ayuso
` (2 subsequent siblings)
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Antonio Ojea <aojea@google.com>
Test that nfqueue with and without GSO process SCTP packets correctly.
Joint work with Pablo.
Signed-off-by: Antonio Ojea <aojea@google.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
.../selftests/net/netfilter/nft_queue.sh | 76 +++++++++++++++++++
1 file changed, 76 insertions(+)
diff --git a/tools/testing/selftests/net/netfilter/nft_queue.sh b/tools/testing/selftests/net/netfilter/nft_queue.sh
index 8538f08c64c2..288b3cc55ed7 100755
--- a/tools/testing/selftests/net/netfilter/nft_queue.sh
+++ b/tools/testing/selftests/net/netfilter/nft_queue.sh
@@ -25,6 +25,9 @@ cleanup()
}
checktool "nft --version" "test without nft tool"
+checktool "socat -h" "run test without socat"
+
+modprobe -q sctp
trap cleanup EXIT
@@ -375,6 +378,77 @@ EOF
wait 2>/dev/null
}
+sctp_listener_ready()
+{
+ ss -S -N "$1" -lnt -o "sport = :12345" | grep -q 12345
+}
+
+test_sctp_forward()
+{
+ ip netns exec "$nsrouter" nft -f /dev/stdin <<EOF
+table inet sctpq {
+ chain forward {
+ type filter hook forward priority 0; policy accept;
+ sctp dport 12345 queue num 10
+ }
+}
+EOF
+ ip netns exec "$nsrouter" ./nf_queue -q 10 -G -t "$timeout" &
+ local nfqpid=$!
+
+ timeout 60 ip netns exec "$ns2" socat -u SCTP-LISTEN:12345 STDOUT > "$TMPFILE1" &
+ local rpid=$!
+
+ busywait "$BUSYWAIT_TIMEOUT" sctp_listener_ready "$ns2"
+
+ ip netns exec "$ns1" socat -u STDIN SCTP:10.0.2.99:12345 <"$TMPINPUT" >/dev/null
+
+ if ! ip netns exec "$nsrouter" nft delete table inet sctpq; then
+ echo "FAIL: Could not delete sctpq table"
+ exit 1
+ fi
+
+ if ! diff -u "$TMPINPUT" "$TMPFILE1" ; then
+ echo "FAIL: lost packets?!" 1>&2
+ return
+ fi
+
+ wait "$rpid" && echo "PASS: sctp and nfqueue in forward chain"
+}
+
+test_sctp_output()
+{
+ ip netns exec "$ns1" nft -f /dev/stdin <<EOF
+table inet sctpq {
+ chain output {
+ type filter hook output priority 0; policy accept;
+ sctp dport 12345 queue num 11
+ }
+}
+EOF
+ ip netns exec "$ns1" ./nf_queue -q 11 -t "$timeout" &
+ local nfqpid=$!
+
+ timeout 60 ip netns exec "$ns2" socat -u SCTP-LISTEN:12345 STDOUT > "$TMPFILE1" &
+ local rpid=$!
+
+ busywait "$BUSYWAIT_TIMEOUT" sctp_listener_ready "$ns2"
+
+ ip netns exec "$ns1" socat -u STDIN SCTP:10.0.2.99:12345 <"$TMPINPUT" >/dev/null
+
+ if ! ip netns exec "$ns1" nft delete table inet sctpq; then
+ echo "FAIL: Could not delete sctpq table"
+ exit 1
+ fi
+
+ if ! diff -u "$TMPINPUT" "$TMPFILE1" ; then
+ echo "FAIL: lost packets?!" 1>&2
+ return
+ fi
+
+ wait "$rpid" && echo "PASS: sctp and nfqueue in output chain with GSO"
+}
+
ip netns exec "$nsrouter" sysctl net.ipv6.conf.all.forwarding=1 > /dev/null
ip netns exec "$nsrouter" sysctl net.ipv4.conf.veth0.forwarding=1 > /dev/null
ip netns exec "$nsrouter" sysctl net.ipv4.conf.veth1.forwarding=1 > /dev/null
@@ -413,5 +487,7 @@ test_tcp_localhost
test_tcp_localhost_connectclose
test_tcp_localhost_requeue
test_icmp_vrf
+test_sctp_forward
+test_sctp_output
exit $ret
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 18/19] selftests: netfilter: nft_queue.sh: add test for disappearing listener
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (16 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 17/19] selftests: netfilter: nft_queue.sh: sctp coverage Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:27 ` [PATCH nf-next 19/19] netfilter: xt_recent: Lift restrictions on max hitcount value Pablo Neira Ayuso
2024-06-27 11:28 ` [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Florian Westphal <fw@strlen.de>
If userspace program exits while the queue its subscribed to has packets
those need to be discarded.
commit dc21c6cc3d69 ("netfilter: nfnetlink_queue: acquire rcu_read_lock()
in instance_destroy_rcu()") fixed a (harmless) rcu splat that could be
triggered in this case.
Add a test case to cover this.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
.../selftests/net/netfilter/nft_queue.sh | 37 +++++++++++++++++++
1 file changed, 37 insertions(+)
diff --git a/tools/testing/selftests/net/netfilter/nft_queue.sh b/tools/testing/selftests/net/netfilter/nft_queue.sh
index 288b3cc55ed7..7e210d75e738 100755
--- a/tools/testing/selftests/net/netfilter/nft_queue.sh
+++ b/tools/testing/selftests/net/netfilter/nft_queue.sh
@@ -449,6 +449,42 @@ EOF
wait "$rpid" && echo "PASS: sctp and nfqueue in output chain with GSO"
}
+test_queue_removal()
+{
+ read tainted_then < /proc/sys/kernel/tainted
+
+ ip netns exec "$ns1" nft -f - <<EOF
+flush ruleset
+table ip filter {
+ chain output {
+ type filter hook output priority 0; policy accept;
+ ip protocol icmp queue num 0
+ }
+}
+EOF
+ ip netns exec "$ns1" ./nf_queue -q 0 -d 30000 -t "$timeout" &
+ local nfqpid=$!
+
+ busywait "$BUSYWAIT_TIMEOUT" nf_queue_wait "$ns1" 0
+
+ ip netns exec "$ns1" ping -w 2 -f -c 10 127.0.0.1 -q >/dev/null
+ kill $nfqpid
+
+ ip netns exec "$ns1" nft flush ruleset
+
+ if [ "$tainted_then" -ne 0 ];then
+ return
+ fi
+
+ read tainted_now < /proc/sys/kernel/tainted
+ if [ "$tainted_now" -eq 0 ];then
+ echo "PASS: queue program exiting while packets queued"
+ else
+ echo "TAINT: queue program exiting while packets queued"
+ ret=1
+ fi
+}
+
ip netns exec "$nsrouter" sysctl net.ipv6.conf.all.forwarding=1 > /dev/null
ip netns exec "$nsrouter" sysctl net.ipv4.conf.veth0.forwarding=1 > /dev/null
ip netns exec "$nsrouter" sysctl net.ipv4.conf.veth1.forwarding=1 > /dev/null
@@ -489,5 +525,6 @@ test_tcp_localhost_requeue
test_icmp_vrf
test_sctp_forward
test_sctp_output
+test_queue_removal
exit $ret
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* [PATCH nf-next 19/19] netfilter: xt_recent: Lift restrictions on max hitcount value
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (17 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 18/19] selftests: netfilter: nft_queue.sh: add test for disappearing listener Pablo Neira Ayuso
@ 2024-06-27 11:27 ` Pablo Neira Ayuso
2024-06-27 11:28 ` [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
19 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:27 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
From: Phil Sutter <phil@nwl.cc>
Support tracking of up to 65535 packets per table entry instead of just
255 to better facilitate longer term tracking or higher throughput
scenarios.
Note how this aligns sizes of struct recent_entry's 'nstamps' and
'index' fields when 'nstamps' was larger before. This is unnecessary as
the value of 'nstamps' grows along with that of 'index' after being
initialized to 1 (see recent_entry_update()). Its value will thus never
exceed that of 'index' and therefore does not need to provide space for
larger values.
Requested-by: Fabio <pedretti.fabio@gmail.com>
Link: https://bugzilla.netfilter.org/show_bug.cgi?id=1745
Signed-off-by: Phil Sutter <phil@nwl.cc>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/xt_recent.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/net/netfilter/xt_recent.c b/net/netfilter/xt_recent.c
index ef93e0d3bee0..588a5e6ad899 100644
--- a/net/netfilter/xt_recent.c
+++ b/net/netfilter/xt_recent.c
@@ -59,9 +59,9 @@ MODULE_PARM_DESC(ip_list_gid, "default owning group of /proc/net/xt_recent/* fil
/* retained for backwards compatibility */
static unsigned int ip_pkt_list_tot __read_mostly;
module_param(ip_pkt_list_tot, uint, 0400);
-MODULE_PARM_DESC(ip_pkt_list_tot, "number of packets per IP address to remember (max. 255)");
+MODULE_PARM_DESC(ip_pkt_list_tot, "number of packets per IP address to remember (max. 65535)");
-#define XT_RECENT_MAX_NSTAMPS 256
+#define XT_RECENT_MAX_NSTAMPS 65536
struct recent_entry {
struct list_head list;
@@ -69,7 +69,7 @@ struct recent_entry {
union nf_inet_addr addr;
u_int16_t family;
u_int8_t ttl;
- u_int8_t index;
+ u_int16_t index;
u_int16_t nstamps;
unsigned long stamps[];
};
@@ -80,7 +80,7 @@ struct recent_table {
union nf_inet_addr mask;
unsigned int refcnt;
unsigned int entries;
- u8 nstamps_max_mask;
+ u_int16_t nstamps_max_mask;
struct list_head lru_list;
struct list_head iphash[];
};
--
2.30.2
^ permalink raw reply related [flat|nested] 25+ messages in thread* Re: [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next
2024-06-27 11:26 [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
` (18 preceding siblings ...)
2024-06-27 11:27 ` [PATCH nf-next 19/19] netfilter: xt_recent: Lift restrictions on max hitcount value Pablo Neira Ayuso
@ 2024-06-27 11:28 ` Pablo Neira Ayuso
2024-06-27 18:32 ` Jakub Kicinski
19 siblings, 1 reply; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-27 11:28 UTC (permalink / raw)
To: netfilter-devel; +Cc: davem, netdev, kuba, pabeni, edumazet, fw
Note for netdev maintainer: This PR is actually targeted at *net-next*.
Please, let me know if you prefer I resubmit.
On Thu, Jun 27, 2024 at 01:26:54PM +0200, Pablo Neira Ayuso wrote:
> Hi,
>
> The following patchset contains Netfilter/IPVS updates for net-next:
>
> Patch #1 to #11 to shrink memory consumption for transaction objects:
>
> struct nft_trans_chain { /* size: 120 (-32), cachelines: 2, members: 10 */
> struct nft_trans_elem { /* size: 72 (-40), cachelines: 2, members: 4 */
> struct nft_trans_flowtable { /* size: 80 (-48), cachelines: 2, members: 5 */
> struct nft_trans_obj { /* size: 72 (-40), cachelines: 2, members: 4 */
> struct nft_trans_rule { /* size: 80 (-32), cachelines: 2, members: 6 */
> struct nft_trans_set { /* size: 96 (-24), cachelines: 2, members: 8 */
> struct nft_trans_table { /* size: 56 (-40), cachelines: 1, members: 2 */
>
> struct nft_trans_elem can now be allocated from kmalloc-96 instead of
> kmalloc-128 slab.
>
> Series from Florian Westphal. For the record, I have mangled patch #1
> to add nft_trans_container_*() and use if for every transaction object.
> I have also added BUILD_BUG_ON to ensure struct nft_trans always comes
> at the beginning of the container transaction object. And few minor
> cleanups, any new bugs are of my own.
>
> Patch #12 simplify check for SCTP GSO in IPVS, from Ismael Luceno.
>
> Patch #13 nf_conncount key length remains in the u32 bound, from Yunjian Wang.
>
> Patch #14 removes unnecessary check for CTA_TIMEOUT_L3PROTO when setting
> default conntrack timeouts via nfnetlink_cttimeout API, from
> Lin Ma.
>
> Patch #15 updates NFT_SECMARK_CTX_MAXLEN to 4096, SELinux could use
> larger secctx names than the existing 256 bytes length.
>
> Patch #16 fixes nfnetlink_queue with SCTP traffic, from Antonio Ojea.
>
> Patch #17 adds a selftest for SCTP traffic under nfnetlink_queue,
> also from Antonio.
>
> Patch #18 adds a selftest to exercise nfnetlink_queue listeners leaving
> nfnetlink_queue, from Florian Westphal.
>
> Patch #19 increases hitcount from 255 to 65535 in xt_recent, from Phil Sutter.
>
> Please, pull these changes from:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git nf-next-24-06-27
>
> Thanks.
>
> ----------------------------------------------------------------
>
> The following changes since commit c4532232fa2a4f8d9b9a88135a666545157f3d13:
>
> selftests: net: remove unneeded IP_GRE config (2024-06-25 08:37:55 -0700)
>
> are available in the Git repository at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git nf-next-24-06-27
>
> for you to fetch changes up to 8871d1e4dceb6692ea8217b1fc835c4bf2e93d97:
>
> netfilter: xt_recent: Lift restrictions on max hitcount value (2024-06-27 01:55:57 +0200)
>
> ----------------------------------------------------------------
> netfilter pull request 24-06-27
>
> ----------------------------------------------------------------
> Antonio Ojea (2):
> netfilter: nfnetlink_queue: unbreak SCTP traffic
> selftests: netfilter: nft_queue.sh: sctp coverage
>
> Florian Westphal (12):
> netfilter: nf_tables: make struct nft_trans first member of derived subtypes
> netfilter: nf_tables: move bind list_head into relevant subtypes
> netfilter: nf_tables: compact chain+ft transaction objects
> netfilter: nf_tables: reduce trans->ctx.table references
> netfilter: nf_tables: pass nft_chain to destroy function, not nft_ctx
> netfilter: nf_tables: pass more specific nft_trans_chain where possible
> netfilter: nf_tables: avoid usage of embedded nft_ctx
> netfilter: nf_tables: store chain pointer in rule transaction
> netfilter: nf_tables: reduce trans->ctx.chain references
> netfilter: nf_tables: pass nft_table to destroy function
> netfilter: nf_tables: do not store nft_ctx in transaction objects
> selftests: netfilter: nft_queue.sh: add test for disappearing listener
>
> Ismael Luceno (1):
> ipvs: Avoid unnecessary calls to skb_is_gso_sctp
>
> Lin Ma (1):
> netfilter: cttimeout: remove 'l3num' attr check
>
> Pablo Neira Ayuso (1):
> netfilter: nf_tables: rise cap on SELinux secmark context
>
> Phil Sutter (1):
> netfilter: xt_recent: Lift restrictions on max hitcount value
>
> Yunjian Wang (1):
> netfilter: nf_conncount: fix wrong variable type
>
> include/net/netfilter/nf_tables.h | 222 +++++++----
> include/uapi/linux/netfilter/nf_tables.h | 2 +-
> net/core/dev.c | 1 +
> net/netfilter/ipvs/ip_vs_proto_sctp.c | 4 +-
> net/netfilter/nf_conncount.c | 8 +-
> net/netfilter/nf_tables_api.c | 411 ++++++++++++---------
> net/netfilter/nf_tables_offload.c | 40 +-
> net/netfilter/nfnetlink_cttimeout.c | 3 +-
> net/netfilter/nfnetlink_queue.c | 12 +-
> net/netfilter/nft_immediate.c | 2 +-
> net/netfilter/xt_recent.c | 8 +-
> tools/testing/selftests/net/netfilter/nft_queue.sh | 113 ++++++
> 12 files changed, 546 insertions(+), 280 deletions(-)
>
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next
2024-06-27 11:28 ` [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next Pablo Neira Ayuso
@ 2024-06-27 18:32 ` Jakub Kicinski
2024-06-27 20:00 ` Florian Westphal
` (2 more replies)
0 siblings, 3 replies; 25+ messages in thread
From: Jakub Kicinski @ 2024-06-27 18:32 UTC (permalink / raw)
To: Pablo Neira Ayuso; +Cc: netfilter-devel, davem, netdev, pabeni, edumazet, fw
On Thu, 27 Jun 2024 13:28:51 +0200 Pablo Neira Ayuso wrote:
> Note for netdev maintainer: This PR is actually targeted at *net-next*.
>
> Please, let me know if you prefer I resubmit.
Not a big deal, but since you offered I have another ask - looks like
this series makes the nf_queue test time out in our infra.
https://netdev.bots.linux.dev/contest.html?test=nft-queue-sh
Could you take a look before you respin? It used to take 24 sec,
not it times out after 224 sec..
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next
2024-06-27 18:32 ` Jakub Kicinski
@ 2024-06-27 20:00 ` Florian Westphal
2024-06-28 13:36 ` Pablo Neira Ayuso
2024-06-28 15:17 ` Florian Westphal
2 siblings, 0 replies; 25+ messages in thread
From: Florian Westphal @ 2024-06-27 20:00 UTC (permalink / raw)
To: Jakub Kicinski
Cc: Pablo Neira Ayuso, netfilter-devel, davem, netdev, pabeni,
edumazet, fw
Jakub Kicinski <kuba@kernel.org> wrote:
> On Thu, 27 Jun 2024 13:28:51 +0200 Pablo Neira Ayuso wrote:
> > Note for netdev maintainer: This PR is actually targeted at *net-next*.
> >
> > Please, let me know if you prefer I resubmit.
>
> Not a big deal, but since you offered I have another ask - looks like
> this series makes the nf_queue test time out in our infra.
>
> https://netdev.bots.linux.dev/contest.html?test=nft-queue-sh
>
> Could you take a look before you respin? It used to take 24 sec,
> not it times out after 224 sec..
Looks like its the sctp selftests that got added, I can have a look
tomorrow.
^ permalink raw reply [flat|nested] 25+ messages in thread
* Re: [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next
2024-06-27 18:32 ` Jakub Kicinski
2024-06-27 20:00 ` Florian Westphal
@ 2024-06-28 13:36 ` Pablo Neira Ayuso
2024-06-28 15:17 ` Florian Westphal
2 siblings, 0 replies; 25+ messages in thread
From: Pablo Neira Ayuso @ 2024-06-28 13:36 UTC (permalink / raw)
To: Jakub Kicinski; +Cc: netfilter-devel, davem, netdev, pabeni, edumazet, fw
On Thu, Jun 27, 2024 at 11:32:02AM -0700, Jakub Kicinski wrote:
> On Thu, 27 Jun 2024 13:28:51 +0200 Pablo Neira Ayuso wrote:
> > Note for netdev maintainer: This PR is actually targeted at *net-next*.
> >
> > Please, let me know if you prefer I resubmit.
>
> Not a big deal, but since you offered I have another ask - looks like
> this series makes the nf_queue test time out in our infra.
>
> https://netdev.bots.linux.dev/contest.html?test=nft-queue-sh
>
> Could you take a look before you respin? It used to take 24 sec,
> not it times out after 224 sec..
Please, ditch this PR, SCTP support for nfqueue is not yet ready.
I will submit a new PR without these bits.
Apologies for the noise, thanks.
^ permalink raw reply [flat|nested] 25+ messages in thread
* Re: [PATCH nf-next 00/19] Netfilter/IPVS updates for net-next
2024-06-27 18:32 ` Jakub Kicinski
2024-06-27 20:00 ` Florian Westphal
2024-06-28 13:36 ` Pablo Neira Ayuso
@ 2024-06-28 15:17 ` Florian Westphal
2 siblings, 0 replies; 25+ messages in thread
From: Florian Westphal @ 2024-06-28 15:17 UTC (permalink / raw)
To: Jakub Kicinski
Cc: Pablo Neira Ayuso, netfilter-devel, davem, netdev, pabeni,
edumazet, fw
Jakub Kicinski <kuba@kernel.org> wrote:
> On Thu, 27 Jun 2024 13:28:51 +0200 Pablo Neira Ayuso wrote:
> > Note for netdev maintainer: This PR is actually targeted at *net-next*.
> >
> > Please, let me know if you prefer I resubmit.
>
> Not a big deal, but since you offered I have another ask - looks like
> this series makes the nf_queue test time out in our infra.
>
> https://netdev.bots.linux.dev/contest.html?test=nft-queue-sh
>
> Could you take a look before you respin? It used to take 24 sec,
> not it times out after 224 sec..
FTR, its missing update to config file:
diff --git a/tools/testing/selftests/net/netfilter/config b/tools/testing/selftests/net/netfilter/config
index 63ef80ef47a4..b2dd4db45215 100644
--- a/tools/testing/selftests/net/netfilter/config
+++ b/tools/testing/selftests/net/netfilter/config
@@ -87,3 +87,5 @@ CONFIG_XFRM_USER=m
CONFIG_XFRM_STATISTICS=y
CONFIG_NET_PKTGEN=m
CONFIG_TUN=m
+CONFIG_INET_DIAG=m
+CONFIG_SCTP_DIAG=m
so the 'wait for sctp listener to appear' takes 1m, after that the
nfqueue listener has timed out aeons ago and sctp connect hangs until
timeout.
But fixing the config shows the tests are very very flaky, this needs more
work, will look into it on monday.
Sorry for this.
^ permalink raw reply related [flat|nested] 25+ messages in thread