[PATCH v2 net-next 0/7] netfilter: updates for net-next

[PATCH v2 net-next 0/7] netfilter: updates for net-next

[Florian Westphal]

v2: drop patch 5, to be routed via net tree. No other changes.

Hi,

The following patchset contains Netfilter fixes for *net-next*:

1) prefer vmalloc_array in ebtables, from  Qianfeng Rong.
2) Use csum_replace4 instead of open-coding it, from Christophe Leroy.
3+4) Get rid of GFP_ATOMIC in transaction object allocations, those
     cause silly failures with large sets under memory pressure, from
     myself.
5) Remove test for AVX cpu feature in nftables pipapo set type,
   testing for AVX2 feature is sufficient.
6) Unexport a few function in nf_reject infra: no external callers.
7) Extend payload offset to u16, this was restricted to values <=255
   so far, from Fernando Fernandez Mancera.

Please, pull these changes from:
The following changes since commit cd8a4cfa6bb43a441901e82f5c222dddc75a18a3:

  Merge branch 'e-switch-vport-sharing-delegation' (2025-09-02 15:18:19 +0200)

are available in the Git repository at:

  https://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf-next.git tags/nf-next-25-09-02

for you to fetch changes up to 077dc4a275790b09e8a2ce80822ba8970e9dfb99:

  netfilter: nft_payload: extend offset to 65535 bytes (2025-09-02 15:28:18 +0200)

----------------------------------------------------------------
netfilter pull request nf-next-25-09-02

----------------------------------------------------------------
Christophe Leroy (1):
      netfilter: nft_payload: Use csum_replace4() instead of opencoding

Fernando Fernandez Mancera (1):
      netfilter: nft_payload: extend offset to 65535 bytes

Florian Westphal (4):
      netfilter: nf_tables: allow iter callbacks to sleep
      netfilter: nf_tables: all transaction allocations can now sleep
      netfilter: nft_set_pipapo: remove redundant test for avx feature bit
      netfilter: nf_reject: remove unneeded exports

Qianfeng Rong (1):
      netfilter: ebtables: Use vmalloc_array() to improve code

 include/net/netfilter/ipv4/nf_reject.h |   8 ---
 include/net/netfilter/ipv6/nf_reject.h |  10 ----
 include/net/netfilter/nf_tables.h      |   2 +
 include/net/netfilter/nf_tables_core.h |   2 +-
 net/bridge/netfilter/ebtables.c        |  14 ++---
 net/ipv4/netfilter/nf_reject_ipv4.c    |  27 +++++----
 net/ipv6/netfilter/nf_reject_ipv6.c    |  37 ++++++++----
 net/netfilter/nf_tables_api.c          |  47 +++++++---------
 net/netfilter/nft_payload.c            |  20 ++++---
 net/netfilter/nft_set_hash.c           | 100 ++++++++++++++++++++++++++++++++-
 net/netfilter/nft_set_pipapo.c         |   3 +-
 net/netfilter/nft_set_pipapo_avx2.c    |   2 +-
 net/netfilter/nft_set_rbtree.c         |  35 +++++++++---
 13 files changed, 209 insertions(+), 98 deletions(-)

[PATCH v2 net-next 1/7] netfilter: ebtables: Use vmalloc_array() to improve code

[Florian Westphal]

From: Qianfeng Rong <rongqianfeng@vivo.com>

Remove array_size() calls and replace vmalloc() with vmalloc_array() to
simplify the code.  vmalloc_array() is also optimized better, uses fewer
instructions, and handles overflow more concisely[1].

[1]: https://lore.kernel.org/lkml/abc66ec5-85a4-47e1-9759-2f60ab111971@vivo.com/
Signed-off-by: Qianfeng Rong <rongqianfeng@vivo.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
---
 net/bridge/netfilter/ebtables.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/net/bridge/netfilter/ebtables.c b/net/bridge/netfilter/ebtables.c
index 3e67d4aff419..5697e3949a36 100644
--- a/net/bridge/netfilter/ebtables.c
+++ b/net/bridge/netfilter/ebtables.c
@@ -920,8 +920,8 @@ static int translate_table(struct net *net, const char *name,
 		 * if an error occurs
 		 */
 		newinfo->chainstack =
-			vmalloc(array_size(nr_cpu_ids,
-					   sizeof(*(newinfo->chainstack))));
+			vmalloc_array(nr_cpu_ids,
+				      sizeof(*(newinfo->chainstack)));
 		if (!newinfo->chainstack)
 			return -ENOMEM;
 		for_each_possible_cpu(i) {
@@ -938,7 +938,7 @@ static int translate_table(struct net *net, const char *name,
 			}
 		}
 
-		cl_s = vmalloc(array_size(udc_cnt, sizeof(*cl_s)));
+		cl_s = vmalloc_array(udc_cnt, sizeof(*cl_s));
 		if (!cl_s)
 			return -ENOMEM;
 		i = 0; /* the i'th udc */
@@ -1018,8 +1018,8 @@ static int do_replace_finish(struct net *net, struct ebt_replace *repl,
 	 * the check on the size is done later, when we have the lock
 	 */
 	if (repl->num_counters) {
-		unsigned long size = repl->num_counters * sizeof(*counterstmp);
-		counterstmp = vmalloc(size);
+		counterstmp = vmalloc_array(repl->num_counters,
+					    sizeof(*counterstmp));
 		if (!counterstmp)
 			return -ENOMEM;
 	}
@@ -1386,7 +1386,7 @@ static int do_update_counters(struct net *net, const char *name,
 	if (num_counters == 0)
 		return -EINVAL;
 
-	tmp = vmalloc(array_size(num_counters, sizeof(*tmp)));
+	tmp = vmalloc_array(num_counters, sizeof(*tmp));
 	if (!tmp)
 		return -ENOMEM;
 
@@ -1526,7 +1526,7 @@ static int copy_counters_to_user(struct ebt_table *t,
 	if (num_counters != nentries)
 		return -EINVAL;
 
-	counterstmp = vmalloc(array_size(nentries, sizeof(*counterstmp)));
+	counterstmp = vmalloc_array(nentries, sizeof(*counterstmp));
 	if (!counterstmp)
 		return -ENOMEM;
 
-- 
2.49.1

Re: [PATCH v2 net-next 1/7] netfilter: ebtables: Use vmalloc_array() to improve code

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (main)
by Florian Westphal <fw@strlen.de>:

On Tue,  2 Sep 2025 15:35:43 +0200 you wrote:
> From: Qianfeng Rong <rongqianfeng@vivo.com>
> 
> Remove array_size() calls and replace vmalloc() with vmalloc_array() to
> simplify the code.  vmalloc_array() is also optimized better, uses fewer
> instructions, and handles overflow more concisely[1].
> 
> [1]: https://lore.kernel.org/lkml/abc66ec5-85a4-47e1-9759-2f60ab111971@vivo.com/
> Signed-off-by: Qianfeng Rong <rongqianfeng@vivo.com>
> Signed-off-by: Florian Westphal <fw@strlen.de>
> 
> [...]

Here is the summary with links:
  - [v2,net-next,1/7] netfilter: ebtables: Use vmalloc_array() to improve code
    https://git.kernel.org/netdev/net-next/c/46015e6b3ea7
  - [v2,net-next,2/7] netfilter: nft_payload: Use csum_replace4() instead of opencoding
    https://git.kernel.org/netdev/net-next/c/c015e17ba111
  - [v2,net-next,3/7] netfilter: nf_tables: allow iter callbacks to sleep
    https://git.kernel.org/netdev/net-next/c/a60a5abe19d6
  - [v2,net-next,4/7] netfilter: nf_tables: all transaction allocations can now sleep
    https://git.kernel.org/netdev/net-next/c/3d95a2e016ab
  - [v2,net-next,5/7] netfilter: nft_set_pipapo: remove redundant test for avx feature bit
    https://git.kernel.org/netdev/net-next/c/8959f27d39d6
  - [v2,net-next,6/7] netfilter: nf_reject: remove unneeded exports
    https://git.kernel.org/netdev/net-next/c/f4f9e05904e1
  - [v2,net-next,7/7] netfilter: nft_payload: extend offset to 65535 bytes
    https://git.kernel.org/netdev/net-next/c/077dc4a27579

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

[PATCH v2 net-next 2/7] netfilter: nft_payload: Use csum_replace4() instead of opencoding

[Florian Westphal]

From: Christophe Leroy <christophe.leroy@csgroup.eu>

Open coded calculation can be avoided and replaced by the
equivalent csum_replace4() in nft_csum_replace().

Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
---
 net/netfilter/nft_payload.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/netfilter/nft_payload.c b/net/netfilter/nft_payload.c
index 7dfc5343dae4..059b28ffad0e 100644
--- a/net/netfilter/nft_payload.c
+++ b/net/netfilter/nft_payload.c
@@ -684,7 +684,7 @@ static const struct nft_expr_ops nft_payload_inner_ops = {
 
 static inline void nft_csum_replace(__sum16 *sum, __wsum fsum, __wsum tsum)
 {
-	*sum = csum_fold(csum_add(csum_sub(~csum_unfold(*sum), fsum), tsum));
+	csum_replace4(sum, (__force __be32)fsum, (__force __be32)tsum);
 	if (*sum == 0)
 		*sum = CSUM_MANGLED_0;
 }
-- 
2.49.1

[PATCH v2 net-next 3/7] netfilter: nf_tables: allow iter callbacks to sleep

[Florian Westphal]

Quoting Sven Auhagen:
  we do see on occasions that we get the following error message, more so on
  x86 systems than on arm64:

  Error: Could not process rule: Cannot allocate memory delete table inet filter

  It is not a consistent error and does not happen all the time.
  We are on Kernel 6.6.80, seems to me like we have something along the lines
  of the nf_tables: allow clone callbacks to sleep problem using GFP_ATOMIC.

As hinted at by Sven, this is because of GFP_ATOMIC allocations during
set flush.

When set is flushed, all elements are deactivated. This triggers a set
walk and each element gets added to the transaction list.

The rbtree and rhashtable sets don't allow the iter callback to sleep:
rbtree walk acquires read side of an rwlock with bh disabled, rhashtable
walk happens with rcu read lock held.

Rbtree is simple enough to resolve:
When the walk context is ITER_READ, no change is needed (the iter
callback must not deactivate elements; we're not in a transaction).

When the iter type is ITER_UPDATE, the rwlock isn't needed because the
caller holds the transaction mutex, this prevents any and all changes to
the ruleset, including add/remove of set elements.

Rhashtable is slightly more complex.
When the iter type is ITER_READ, no change is needed, like rbtree.

For ITER_UPDATE, we hold transaction mutex which prevents elements from
getting free'd, even outside of rcu read lock section.

So build a temporary list of all elements while doing the rcu iteration
and then call the iterator in a second pass.

The disadvantage is the need to iterate twice, but this cost comes with
the benefit to allow the iter callback to use GFP_KERNEL allocations in
a followup patch.

The new list based logic makes it necessary to catch recursive calls to
the same set earlier.

Such walk -> iter -> walk recursion for the same set can happen during
ruleset validation in case userspace gave us a bogus (cyclic) ruleset
where verdict map m jumps to chain that sooner or later also calls
"vmap @m".

Before the new ->in_update_walk test, the ruleset is rejected because the
infinite recursion causes ctx->level to exceed the allowed maximum.

But with the new logic added here, elements would get skipped:

nft_rhash_walk_update would see elements that are on the walk_list of
an older stack frame.

As all recursive calls into same map results in -EMLINK, we can avoid this
problem by using the new in_update_walk flag and reject immediately.

Next patch converts the problematic GFP_ATOMIC allocations.

Reported-by: Sven Auhagen <Sven.Auhagen@belden.com>
Closes: https://lore.kernel.org/netfilter-devel/BY1PR18MB5874110CAFF1ED098D0BC4E7E07BA@BY1PR18MB5874.namprd18.prod.outlook.com/
Signed-off-by: Florian Westphal <fw@strlen.de>
---
 include/net/netfilter/nf_tables.h |   2 +
 net/netfilter/nft_set_hash.c      | 100 +++++++++++++++++++++++++++++-
 net/netfilter/nft_set_rbtree.c    |  35 ++++++++---
 3 files changed, 126 insertions(+), 11 deletions(-)

diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h
index 891e43a01bdc..e2128663b160 100644
--- a/include/net/netfilter/nf_tables.h
+++ b/include/net/netfilter/nf_tables.h
@@ -556,6 +556,7 @@ struct nft_set_elem_expr {
  * 	@size: maximum set size
  *	@field_len: length of each field in concatenation, bytes
  *	@field_count: number of concatenated fields in element
+ *	@in_update_walk: true during ->walk() in transaction phase
  *	@use: number of rules references to this set
  * 	@nelems: number of elements
  * 	@ndeact: number of deactivated elements queued for removal
@@ -590,6 +591,7 @@ struct nft_set {
 	u32				size;
 	u8				field_len[NFT_REG32_COUNT];
 	u8				field_count;
+	bool				in_update_walk;
 	u32				use;
 	atomic_t			nelems;
 	u32				ndeact;
diff --git a/net/netfilter/nft_set_hash.c b/net/netfilter/nft_set_hash.c
index 266d0c637225..ba01ce75d6de 100644
--- a/net/netfilter/nft_set_hash.c
+++ b/net/netfilter/nft_set_hash.c
@@ -30,6 +30,7 @@ struct nft_rhash {
 struct nft_rhash_elem {
 	struct nft_elem_priv		priv;
 	struct rhash_head		node;
+	struct llist_node		walk_node;
 	u32				wq_gc_seq;
 	struct nft_set_ext		ext;
 };
@@ -144,6 +145,7 @@ nft_rhash_update(struct nft_set *set, const u32 *key,
 		goto err1;
 
 	he = nft_elem_priv_cast(elem_priv);
+	init_llist_node(&he->walk_node);
 	prev = rhashtable_lookup_get_insert_key(&priv->ht, &arg, &he->node,
 						nft_rhash_params);
 	if (IS_ERR(prev))
@@ -180,6 +182,7 @@ static int nft_rhash_insert(const struct net *net, const struct nft_set *set,
 	};
 	struct nft_rhash_elem *prev;
 
+	init_llist_node(&he->walk_node);
 	prev = rhashtable_lookup_get_insert_key(&priv->ht, &arg, &he->node,
 						nft_rhash_params);
 	if (IS_ERR(prev))
@@ -261,12 +264,12 @@ static bool nft_rhash_delete(const struct nft_set *set,
 	return true;
 }
 
-static void nft_rhash_walk(const struct nft_ctx *ctx, struct nft_set *set,
-			   struct nft_set_iter *iter)
+static void nft_rhash_walk_ro(const struct nft_ctx *ctx, struct nft_set *set,
+			      struct nft_set_iter *iter)
 {
 	struct nft_rhash *priv = nft_set_priv(set);
-	struct nft_rhash_elem *he;
 	struct rhashtable_iter hti;
+	struct nft_rhash_elem *he;
 
 	rhashtable_walk_enter(&priv->ht, &hti);
 	rhashtable_walk_start(&hti);
@@ -295,6 +298,97 @@ static void nft_rhash_walk(const struct nft_ctx *ctx, struct nft_set *set,
 	rhashtable_walk_exit(&hti);
 }
 
+static void nft_rhash_walk_update(const struct nft_ctx *ctx,
+				  struct nft_set *set,
+				  struct nft_set_iter *iter)
+{
+	struct nft_rhash *priv = nft_set_priv(set);
+	struct nft_rhash_elem *he, *tmp;
+	struct llist_node *first_node;
+	struct rhashtable_iter hti;
+	LLIST_HEAD(walk_list);
+
+	lockdep_assert_held(&nft_pernet(ctx->net)->commit_mutex);
+
+	if (set->in_update_walk) {
+		/* This can happen with bogus rulesets during ruleset validation
+		 * when a verdict map causes a jump back to the same map.
+		 *
+		 * Without this extra check the walk_next loop below will see
+		 * elems on the callers walk_list and skip (not validate) them.
+		 */
+		iter->err = -EMLINK;
+		return;
+	}
+
+	/* walk happens under RCU.
+	 *
+	 * We create a snapshot list so ->iter callback can sleep.
+	 * commit_mutex is held, elements can ...
+	 * .. be added in parallel from dataplane (dynset)
+	 * .. be marked as dead in parallel from dataplane (dynset).
+	 * .. be queued for removal in parallel (gc timeout).
+	 * .. not be freed: transaction mutex is held.
+	 */
+	rhashtable_walk_enter(&priv->ht, &hti);
+	rhashtable_walk_start(&hti);
+
+	while ((he = rhashtable_walk_next(&hti))) {
+		if (IS_ERR(he)) {
+			if (PTR_ERR(he) != -EAGAIN) {
+				iter->err = PTR_ERR(he);
+				break;
+			}
+
+			continue;
+		}
+
+		/* rhashtable resized during walk, skip */
+		if (llist_on_list(&he->walk_node))
+			continue;
+
+		llist_add(&he->walk_node, &walk_list);
+	}
+	rhashtable_walk_stop(&hti);
+	rhashtable_walk_exit(&hti);
+
+	first_node = __llist_del_all(&walk_list);
+	set->in_update_walk = true;
+	llist_for_each_entry_safe(he, tmp, first_node, walk_node) {
+		if (iter->err == 0) {
+			iter->err = iter->fn(ctx, set, iter, &he->priv);
+			if (iter->err == 0)
+				iter->count++;
+		}
+
+		/* all entries must be cleared again, else next ->walk iteration
+		 * will skip entries.
+		 */
+		init_llist_node(&he->walk_node);
+	}
+	set->in_update_walk = false;
+}
+
+static void nft_rhash_walk(const struct nft_ctx *ctx, struct nft_set *set,
+			   struct nft_set_iter *iter)
+{
+	switch (iter->type) {
+	case NFT_ITER_UPDATE:
+		/* only relevant for netlink dumps which use READ type */
+		WARN_ON_ONCE(iter->skip != 0);
+
+		nft_rhash_walk_update(ctx, set, iter);
+		break;
+	case NFT_ITER_READ:
+		nft_rhash_walk_ro(ctx, set, iter);
+		break;
+	default:
+		iter->err = -EINVAL;
+		WARN_ON_ONCE(1);
+		break;
+	}
+}
+
 static bool nft_rhash_expr_needs_gc_run(const struct nft_set *set,
 					struct nft_set_ext *ext)
 {
diff --git a/net/netfilter/nft_set_rbtree.c b/net/netfilter/nft_set_rbtree.c
index 938a257c069e..b311b66df3e9 100644
--- a/net/netfilter/nft_set_rbtree.c
+++ b/net/netfilter/nft_set_rbtree.c
@@ -584,15 +584,14 @@ nft_rbtree_deactivate(const struct net *net, const struct nft_set *set,
 	return NULL;
 }
 
-static void nft_rbtree_walk(const struct nft_ctx *ctx,
-			    struct nft_set *set,
-			    struct nft_set_iter *iter)
+static void nft_rbtree_do_walk(const struct nft_ctx *ctx,
+			       struct nft_set *set,
+			       struct nft_set_iter *iter)
 {
 	struct nft_rbtree *priv = nft_set_priv(set);
 	struct nft_rbtree_elem *rbe;
 	struct rb_node *node;
 
-	read_lock_bh(&priv->lock);
 	for (node = rb_first(&priv->root); node != NULL; node = rb_next(node)) {
 		rbe = rb_entry(node, struct nft_rbtree_elem, node);
 
@@ -600,14 +599,34 @@ static void nft_rbtree_walk(const struct nft_ctx *ctx,
 			goto cont;
 
 		iter->err = iter->fn(ctx, set, iter, &rbe->priv);
-		if (iter->err < 0) {
-			read_unlock_bh(&priv->lock);
+		if (iter->err < 0)
 			return;
-		}
 cont:
 		iter->count++;
 	}
-	read_unlock_bh(&priv->lock);
+}
+
+static void nft_rbtree_walk(const struct nft_ctx *ctx,
+			    struct nft_set *set,
+			    struct nft_set_iter *iter)
+{
+	struct nft_rbtree *priv = nft_set_priv(set);
+
+	switch (iter->type) {
+	case NFT_ITER_UPDATE:
+		lockdep_assert_held(&nft_pernet(ctx->net)->commit_mutex);
+		nft_rbtree_do_walk(ctx, set, iter);
+		break;
+	case NFT_ITER_READ:
+		read_lock_bh(&priv->lock);
+		nft_rbtree_do_walk(ctx, set, iter);
+		read_unlock_bh(&priv->lock);
+		break;
+	default:
+		iter->err = -EINVAL;
+		WARN_ON_ONCE(1);
+		break;
+	}
 }
 
 static void nft_rbtree_gc_remove(struct net *net, struct nft_set *set,
-- 
2.49.1

[PATCH v2 net-next 4/7] netfilter: nf_tables: all transaction allocations can now sleep

[Florian Westphal]

Now that nft_setelem_flush is not called with rcu read lock held or
disabled softinterrupts anymore this can now use GFP_KERNEL too.

This is the last atomic allocation of transaction elements, so remove
all gfp_t arguments and the wrapper function.

This makes attempts to delete large sets much more reliable, before
this was prone to transient memory allocation failures.

Signed-off-by: Florian Westphal <fw@strlen.de>
---
 net/netfilter/nf_tables_api.c | 47 ++++++++++++++---------------------
 1 file changed, 19 insertions(+), 28 deletions(-)

diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 58c5425d61c2..54519b3d2868 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -151,12 +151,12 @@ static void nft_ctx_init(struct nft_ctx *ctx,
 	bitmap_zero(ctx->reg_inited, NFT_REG32_NUM);
 }
 
-static struct nft_trans *nft_trans_alloc_gfp(const struct nft_ctx *ctx,
-					     int msg_type, u32 size, gfp_t gfp)
+static struct nft_trans *nft_trans_alloc(const struct nft_ctx *ctx,
+					 int msg_type, u32 size)
 {
 	struct nft_trans *trans;
 
-	trans = kzalloc(size, gfp);
+	trans = kzalloc(size, GFP_KERNEL);
 	if (trans == NULL)
 		return NULL;
 
@@ -172,12 +172,6 @@ static struct nft_trans *nft_trans_alloc_gfp(const struct nft_ctx *ctx,
 	return trans;
 }
 
-static struct nft_trans *nft_trans_alloc(const struct nft_ctx *ctx,
-					 int msg_type, u32 size)
-{
-	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) {
@@ -442,8 +436,7 @@ static bool nft_trans_collapse_set_elem_allowed(const struct nft_trans_elem *a,
 
 static bool nft_trans_collapse_set_elem(struct nftables_pernet *nft_net,
 					struct nft_trans_elem *tail,
-					struct nft_trans_elem *trans,
-					gfp_t gfp)
+					struct nft_trans_elem *trans)
 {
 	unsigned int nelems, old_nelems = tail->nelems;
 	struct nft_trans_elem *new_trans;
@@ -466,9 +459,11 @@ static bool nft_trans_collapse_set_elem(struct nftables_pernet *nft_net,
 	/* krealloc might free tail which invalidates list pointers */
 	list_del_init(&tail->nft_trans.list);
 
-	new_trans = krealloc(tail, struct_size(tail, elems, nelems), gfp);
+	new_trans = krealloc(tail, struct_size(tail, elems, nelems),
+			     GFP_KERNEL);
 	if (!new_trans) {
-		list_add_tail(&tail->nft_trans.list, &nft_net->commit_list);
+		list_add_tail(&tail->nft_trans.list,
+			      &nft_net->commit_list);
 		return false;
 	}
 
@@ -484,7 +479,7 @@ static bool nft_trans_collapse_set_elem(struct nftables_pernet *nft_net,
 }
 
 static bool nft_trans_try_collapse(struct nftables_pernet *nft_net,
-				   struct nft_trans *trans, gfp_t gfp)
+				   struct nft_trans *trans)
 {
 	struct nft_trans *tail;
 
@@ -501,7 +496,7 @@ static bool nft_trans_try_collapse(struct nftables_pernet *nft_net,
 	case NFT_MSG_DELSETELEM:
 		return nft_trans_collapse_set_elem(nft_net,
 						   nft_trans_container_elem(tail),
-						   nft_trans_container_elem(trans), gfp);
+						   nft_trans_container_elem(trans));
 	}
 
 	return false;
@@ -537,17 +532,14 @@ static void nft_trans_commit_list_add_tail(struct net *net, struct nft_trans *tr
 	}
 }
 
-static void nft_trans_commit_list_add_elem(struct net *net, struct nft_trans *trans,
-					   gfp_t gfp)
+static void nft_trans_commit_list_add_elem(struct net *net, struct nft_trans *trans)
 {
 	struct nftables_pernet *nft_net = nft_pernet(net);
 
 	WARN_ON_ONCE(trans->msg_type != NFT_MSG_NEWSETELEM &&
 		     trans->msg_type != NFT_MSG_DELSETELEM);
 
-	might_alloc(gfp);
-
-	if (nft_trans_try_collapse(nft_net, trans, gfp)) {
+	if (nft_trans_try_collapse(nft_net, trans)) {
 		kfree(trans);
 		return;
 	}
@@ -7549,7 +7541,7 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
 						}
 
 						ue->priv = elem_priv;
-						nft_trans_commit_list_add_elem(ctx->net, trans, GFP_KERNEL);
+						nft_trans_commit_list_add_elem(ctx->net, trans);
 						goto err_elem_free;
 					}
 				}
@@ -7573,7 +7565,7 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
 	}
 
 	nft_trans_container_elem(trans)->elems[0].priv = elem.priv;
-	nft_trans_commit_list_add_elem(ctx->net, trans, GFP_KERNEL);
+	nft_trans_commit_list_add_elem(ctx->net, trans);
 	return 0;
 
 err_set_full:
@@ -7839,7 +7831,7 @@ static int nft_del_setelem(struct nft_ctx *ctx, struct nft_set *set,
 	nft_setelem_data_deactivate(ctx->net, set, elem.priv);
 
 	nft_trans_container_elem(trans)->elems[0].priv = elem.priv;
-	nft_trans_commit_list_add_elem(ctx->net, trans, GFP_KERNEL);
+	nft_trans_commit_list_add_elem(ctx->net, trans);
 	return 0;
 
 fail_ops:
@@ -7864,9 +7856,8 @@ static int nft_setelem_flush(const struct nft_ctx *ctx,
 	if (!nft_set_elem_active(ext, iter->genmask))
 		return 0;
 
-	trans = nft_trans_alloc_gfp(ctx, NFT_MSG_DELSETELEM,
-				    struct_size_t(struct nft_trans_elem, elems, 1),
-				    GFP_ATOMIC);
+	trans = nft_trans_alloc(ctx, NFT_MSG_DELSETELEM,
+				struct_size_t(struct nft_trans_elem, elems, 1));
 	if (!trans)
 		return -ENOMEM;
 
@@ -7877,7 +7868,7 @@ static int nft_setelem_flush(const struct nft_ctx *ctx,
 	nft_trans_elem_set(trans) = set;
 	nft_trans_container_elem(trans)->nelems = 1;
 	nft_trans_container_elem(trans)->elems[0].priv = elem_priv;
-	nft_trans_commit_list_add_elem(ctx->net, trans, GFP_ATOMIC);
+	nft_trans_commit_list_add_elem(ctx->net, trans);
 
 	return 0;
 }
@@ -7894,7 +7885,7 @@ static int __nft_set_catchall_flush(const struct nft_ctx *ctx,
 
 	nft_setelem_data_deactivate(ctx->net, set, elem_priv);
 	nft_trans_container_elem(trans)->elems[0].priv = elem_priv;
-	nft_trans_commit_list_add_elem(ctx->net, trans, GFP_KERNEL);
+	nft_trans_commit_list_add_elem(ctx->net, trans);
 
 	return 0;
 }
-- 
2.49.1

[PATCH v2 net-next 5/7] netfilter: nft_set_pipapo: remove redundant test for avx feature bit

[Florian Westphal]

Sebastian points out that avx2 depends on avx, see check_cpufeature_deps()
in arch/x86/kernel/cpu/cpuid-deps.c:
avx2 feature bit will be cleared when avx isn't available.

No functional change intended.

Reported-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Reviewed-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
---
 net/netfilter/nft_set_pipapo.c      | 3 +--
 net/netfilter/nft_set_pipapo_avx2.c | 2 +-
 2 files changed, 2 insertions(+), 3 deletions(-)

diff --git a/net/netfilter/nft_set_pipapo.c b/net/netfilter/nft_set_pipapo.c
index b385cfcf886f..4b64c3bd8e70 100644
--- a/net/netfilter/nft_set_pipapo.c
+++ b/net/netfilter/nft_set_pipapo.c
@@ -530,8 +530,7 @@ static struct nft_pipapo_elem *pipapo_get(const struct nft_pipapo_match *m,
 	local_bh_disable();
 
 #if defined(CONFIG_X86_64) && !defined(CONFIG_UML)
-	if (boot_cpu_has(X86_FEATURE_AVX2) && boot_cpu_has(X86_FEATURE_AVX) &&
-	    irq_fpu_usable()) {
+	if (boot_cpu_has(X86_FEATURE_AVX2) && irq_fpu_usable()) {
 		e = pipapo_get_avx2(m, data, genmask, tstamp);
 		local_bh_enable();
 		return e;
diff --git a/net/netfilter/nft_set_pipapo_avx2.c b/net/netfilter/nft_set_pipapo_avx2.c
index 29326f3fcaf3..7559306d0aed 100644
--- a/net/netfilter/nft_set_pipapo_avx2.c
+++ b/net/netfilter/nft_set_pipapo_avx2.c
@@ -1099,7 +1099,7 @@ bool nft_pipapo_avx2_estimate(const struct nft_set_desc *desc, u32 features,
 	    desc->field_count < NFT_PIPAPO_MIN_FIELDS)
 		return false;
 
-	if (!boot_cpu_has(X86_FEATURE_AVX2) || !boot_cpu_has(X86_FEATURE_AVX))
+	if (!boot_cpu_has(X86_FEATURE_AVX2))
 		return false;
 
 	est->size = pipapo_estimate_size(desc);
-- 
2.49.1

[PATCH v2 net-next 6/7] netfilter: nf_reject: remove unneeded exports

[Florian Westphal]

These functions have no external callers and can be static.

Signed-off-by: Florian Westphal <fw@strlen.de>
---
 include/net/netfilter/ipv4/nf_reject.h |  8 ------
 include/net/netfilter/ipv6/nf_reject.h | 10 -------
 net/ipv4/netfilter/nf_reject_ipv4.c    | 27 ++++++++++++-------
 net/ipv6/netfilter/nf_reject_ipv6.c    | 37 +++++++++++++++++---------
 4 files changed, 42 insertions(+), 40 deletions(-)

diff --git a/include/net/netfilter/ipv4/nf_reject.h b/include/net/netfilter/ipv4/nf_reject.h
index c653fcb88354..09de2f2686b5 100644
--- a/include/net/netfilter/ipv4/nf_reject.h
+++ b/include/net/netfilter/ipv4/nf_reject.h
@@ -10,14 +10,6 @@
 void nf_send_unreach(struct sk_buff *skb_in, int code, int hook);
 void nf_send_reset(struct net *net, struct sock *, struct sk_buff *oldskb,
 		   int hook);
-const struct tcphdr *nf_reject_ip_tcphdr_get(struct sk_buff *oldskb,
-					     struct tcphdr *_oth, int hook);
-struct iphdr *nf_reject_iphdr_put(struct sk_buff *nskb,
-				  const struct sk_buff *oldskb,
-				  __u8 protocol, int ttl);
-void nf_reject_ip_tcphdr_put(struct sk_buff *nskb, const struct sk_buff *oldskb,
-			     const struct tcphdr *oth);
-
 struct sk_buff *nf_reject_skb_v4_unreach(struct net *net,
                                          struct sk_buff *oldskb,
                                          const struct net_device *dev,
diff --git a/include/net/netfilter/ipv6/nf_reject.h b/include/net/netfilter/ipv6/nf_reject.h
index d729344ba644..94ec0b9f2838 100644
--- a/include/net/netfilter/ipv6/nf_reject.h
+++ b/include/net/netfilter/ipv6/nf_reject.h
@@ -9,16 +9,6 @@ void nf_send_unreach6(struct net *net, struct sk_buff *skb_in, unsigned char cod
 		      unsigned int hooknum);
 void nf_send_reset6(struct net *net, struct sock *sk, struct sk_buff *oldskb,
 		    int hook);
-const struct tcphdr *nf_reject_ip6_tcphdr_get(struct sk_buff *oldskb,
-					      struct tcphdr *otcph,
-					      unsigned int *otcplen, int hook);
-struct ipv6hdr *nf_reject_ip6hdr_put(struct sk_buff *nskb,
-				     const struct sk_buff *oldskb,
-				     __u8 protocol, int hoplimit);
-void nf_reject_ip6_tcphdr_put(struct sk_buff *nskb,
-			      const struct sk_buff *oldskb,
-			      const struct tcphdr *oth, unsigned int otcplen);
-
 struct sk_buff *nf_reject_skb_v6_tcp_reset(struct net *net,
 					   struct sk_buff *oldskb,
 					   const struct net_device *dev,
diff --git a/net/ipv4/netfilter/nf_reject_ipv4.c b/net/ipv4/netfilter/nf_reject_ipv4.c
index 0d3cb2ba6fc8..05631abe3f0d 100644
--- a/net/ipv4/netfilter/nf_reject_ipv4.c
+++ b/net/ipv4/netfilter/nf_reject_ipv4.c
@@ -12,6 +12,15 @@
 #include <linux/netfilter_ipv4.h>
 #include <linux/netfilter_bridge.h>
 
+static struct iphdr *nf_reject_iphdr_put(struct sk_buff *nskb,
+					 const struct sk_buff *oldskb,
+					 __u8 protocol, int ttl);
+static void nf_reject_ip_tcphdr_put(struct sk_buff *nskb, const struct sk_buff *oldskb,
+				    const struct tcphdr *oth);
+static const struct tcphdr *
+nf_reject_ip_tcphdr_get(struct sk_buff *oldskb,
+			struct tcphdr *_oth, int hook);
+
 static int nf_reject_iphdr_validate(struct sk_buff *skb)
 {
 	struct iphdr *iph;
@@ -136,8 +145,9 @@ struct sk_buff *nf_reject_skb_v4_unreach(struct net *net,
 }
 EXPORT_SYMBOL_GPL(nf_reject_skb_v4_unreach);
 
-const struct tcphdr *nf_reject_ip_tcphdr_get(struct sk_buff *oldskb,
-					     struct tcphdr *_oth, int hook)
+static const struct tcphdr *
+nf_reject_ip_tcphdr_get(struct sk_buff *oldskb,
+			struct tcphdr *_oth, int hook)
 {
 	const struct tcphdr *oth;
 
@@ -163,11 +173,10 @@ const struct tcphdr *nf_reject_ip_tcphdr_get(struct sk_buff *oldskb,
 
 	return oth;
 }
-EXPORT_SYMBOL_GPL(nf_reject_ip_tcphdr_get);
 
-struct iphdr *nf_reject_iphdr_put(struct sk_buff *nskb,
-				  const struct sk_buff *oldskb,
-				  __u8 protocol, int ttl)
+static struct iphdr *nf_reject_iphdr_put(struct sk_buff *nskb,
+					 const struct sk_buff *oldskb,
+					 __u8 protocol, int ttl)
 {
 	struct iphdr *niph, *oiph = ip_hdr(oldskb);
 
@@ -188,10 +197,9 @@ struct iphdr *nf_reject_iphdr_put(struct sk_buff *nskb,
 
 	return niph;
 }
-EXPORT_SYMBOL_GPL(nf_reject_iphdr_put);
 
-void nf_reject_ip_tcphdr_put(struct sk_buff *nskb, const struct sk_buff *oldskb,
-			  const struct tcphdr *oth)
+static void nf_reject_ip_tcphdr_put(struct sk_buff *nskb, const struct sk_buff *oldskb,
+				    const struct tcphdr *oth)
 {
 	struct iphdr *niph = ip_hdr(nskb);
 	struct tcphdr *tcph;
@@ -218,7 +226,6 @@ void nf_reject_ip_tcphdr_put(struct sk_buff *nskb, const struct sk_buff *oldskb,
 	nskb->csum_start = (unsigned char *)tcph - nskb->head;
 	nskb->csum_offset = offsetof(struct tcphdr, check);
 }
-EXPORT_SYMBOL_GPL(nf_reject_ip_tcphdr_put);
 
 static int nf_reject_fill_skb_dst(struct sk_buff *skb_in)
 {
diff --git a/net/ipv6/netfilter/nf_reject_ipv6.c b/net/ipv6/netfilter/nf_reject_ipv6.c
index cb2d38e80de9..6b022449f867 100644
--- a/net/ipv6/netfilter/nf_reject_ipv6.c
+++ b/net/ipv6/netfilter/nf_reject_ipv6.c
@@ -12,6 +12,19 @@
 #include <linux/netfilter_ipv6.h>
 #include <linux/netfilter_bridge.h>
 
+static struct ipv6hdr *
+nf_reject_ip6hdr_put(struct sk_buff *nskb,
+		     const struct sk_buff *oldskb,
+		     __u8 protocol, int hoplimit);
+static void
+nf_reject_ip6_tcphdr_put(struct sk_buff *nskb,
+			 const struct sk_buff *oldskb,
+			 const struct tcphdr *oth, unsigned int otcplen);
+static const struct tcphdr *
+nf_reject_ip6_tcphdr_get(struct sk_buff *oldskb,
+			 struct tcphdr *otcph,
+			 unsigned int *otcplen, int hook);
+
 static bool nf_reject_v6_csum_ok(struct sk_buff *skb, int hook)
 {
 	const struct ipv6hdr *ip6h = ipv6_hdr(skb);
@@ -146,9 +159,10 @@ struct sk_buff *nf_reject_skb_v6_unreach(struct net *net,
 }
 EXPORT_SYMBOL_GPL(nf_reject_skb_v6_unreach);
 
-const struct tcphdr *nf_reject_ip6_tcphdr_get(struct sk_buff *oldskb,
-					      struct tcphdr *otcph,
-					      unsigned int *otcplen, int hook)
+static const struct tcphdr *
+nf_reject_ip6_tcphdr_get(struct sk_buff *oldskb,
+			 struct tcphdr *otcph,
+			 unsigned int *otcplen, int hook)
 {
 	const struct ipv6hdr *oip6h = ipv6_hdr(oldskb);
 	u8 proto;
@@ -192,11 +206,11 @@ const struct tcphdr *nf_reject_ip6_tcphdr_get(struct sk_buff *oldskb,
 
 	return otcph;
 }
-EXPORT_SYMBOL_GPL(nf_reject_ip6_tcphdr_get);
 
-struct ipv6hdr *nf_reject_ip6hdr_put(struct sk_buff *nskb,
-				     const struct sk_buff *oldskb,
-				     __u8 protocol, int hoplimit)
+static struct ipv6hdr *
+nf_reject_ip6hdr_put(struct sk_buff *nskb,
+		     const struct sk_buff *oldskb,
+		     __u8 protocol, int hoplimit)
 {
 	struct ipv6hdr *ip6h;
 	const struct ipv6hdr *oip6h = ipv6_hdr(oldskb);
@@ -216,11 +230,11 @@ struct ipv6hdr *nf_reject_ip6hdr_put(struct sk_buff *nskb,
 
 	return ip6h;
 }
-EXPORT_SYMBOL_GPL(nf_reject_ip6hdr_put);
 
-void nf_reject_ip6_tcphdr_put(struct sk_buff *nskb,
-			      const struct sk_buff *oldskb,
-			      const struct tcphdr *oth, unsigned int otcplen)
+static void
+nf_reject_ip6_tcphdr_put(struct sk_buff *nskb,
+			 const struct sk_buff *oldskb,
+			 const struct tcphdr *oth, unsigned int otcplen)
 {
 	struct tcphdr *tcph;
 
@@ -248,7 +262,6 @@ void nf_reject_ip6_tcphdr_put(struct sk_buff *nskb,
 				      csum_partial(tcph,
 						   sizeof(struct tcphdr), 0));
 }
-EXPORT_SYMBOL_GPL(nf_reject_ip6_tcphdr_put);
 
 static int nf_reject6_fill_skb_dst(struct sk_buff *skb_in)
 {
-- 
2.49.1

[PATCH v2 net-next 7/7] netfilter: nft_payload: extend offset to 65535 bytes

[Florian Westphal]

From: Fernando Fernandez Mancera <fmancera@suse.de>

In some situations 255 bytes offset is not enough to match or manipulate
the desired packet field. Increase the offset limit to 65535 or U16_MAX.

In addition, the nla policy maximum value is not set anymore as it is
limited to s16. Instead, the maximum value is checked during the payload
expression initialization function.

Tested with the nft command line tool.

table ip filter {
	chain output {
		@nh,2040,8 set 0xff
		@nh,524280,8 set 0xff
		@nh,524280,8 0xff
		@nh,2040,8 0xff
	}
}

Signed-off-by: Fernando Fernandez Mancera <fmancera@suse.de>
Signed-off-by: Florian Westphal <fw@strlen.de>
---
 include/net/netfilter/nf_tables_core.h |  2 +-
 net/netfilter/nft_payload.c            | 18 +++++++++++-------
 2 files changed, 12 insertions(+), 8 deletions(-)

diff --git a/include/net/netfilter/nf_tables_core.h b/include/net/netfilter/nf_tables_core.h
index 6c2f483d9828..7644cfe9267d 100644
--- a/include/net/netfilter/nf_tables_core.h
+++ b/include/net/netfilter/nf_tables_core.h
@@ -73,7 +73,7 @@ struct nft_ct {
 
 struct nft_payload {
 	enum nft_payload_bases	base:8;
-	u8			offset;
+	u16			offset;
 	u8			len;
 	u8			dreg;
 };
diff --git a/net/netfilter/nft_payload.c b/net/netfilter/nft_payload.c
index 059b28ffad0e..b0214418f75a 100644
--- a/net/netfilter/nft_payload.c
+++ b/net/netfilter/nft_payload.c
@@ -40,7 +40,7 @@ static bool nft_payload_rebuild_vlan_hdr(const struct sk_buff *skb, int mac_off,
 
 /* add vlan header into the user buffer for if tag was removed by offloads */
 static bool
-nft_payload_copy_vlan(u32 *d, const struct sk_buff *skb, u8 offset, u8 len)
+nft_payload_copy_vlan(u32 *d, const struct sk_buff *skb, u16 offset, u8 len)
 {
 	int mac_off = skb_mac_header(skb) - skb->data;
 	u8 *vlanh, *dst_u8 = (u8 *) d;
@@ -212,7 +212,7 @@ static const struct nla_policy nft_payload_policy[NFTA_PAYLOAD_MAX + 1] = {
 	[NFTA_PAYLOAD_SREG]		= { .type = NLA_U32 },
 	[NFTA_PAYLOAD_DREG]		= { .type = NLA_U32 },
 	[NFTA_PAYLOAD_BASE]		= { .type = NLA_U32 },
-	[NFTA_PAYLOAD_OFFSET]		= NLA_POLICY_MAX(NLA_BE32, 255),
+	[NFTA_PAYLOAD_OFFSET]		= { .type = NLA_BE32 },
 	[NFTA_PAYLOAD_LEN]		= NLA_POLICY_MAX(NLA_BE32, 255),
 	[NFTA_PAYLOAD_CSUM_TYPE]	= { .type = NLA_U32 },
 	[NFTA_PAYLOAD_CSUM_OFFSET]	= NLA_POLICY_MAX(NLA_BE32, 255),
@@ -797,7 +797,7 @@ static int nft_payload_csum_inet(struct sk_buff *skb, const u32 *src,
 
 struct nft_payload_set {
 	enum nft_payload_bases	base:8;
-	u8			offset;
+	u16			offset;
 	u8			len;
 	u8			sreg;
 	u8			csum_type;
@@ -812,7 +812,7 @@ struct nft_payload_vlan_hdr {
 };
 
 static bool
-nft_payload_set_vlan(const u32 *src, struct sk_buff *skb, u8 offset, u8 len,
+nft_payload_set_vlan(const u32 *src, struct sk_buff *skb, u16 offset, u8 len,
 		     int *vlan_hlen)
 {
 	struct nft_payload_vlan_hdr *vlanh;
@@ -940,14 +940,18 @@ static int nft_payload_set_init(const struct nft_ctx *ctx,
 				const struct nft_expr *expr,
 				const struct nlattr * const tb[])
 {
+	u32 csum_offset, offset, csum_type = NFT_PAYLOAD_CSUM_NONE;
 	struct nft_payload_set *priv = nft_expr_priv(expr);
-	u32 csum_offset, csum_type = NFT_PAYLOAD_CSUM_NONE;
 	int err;
 
 	priv->base        = ntohl(nla_get_be32(tb[NFTA_PAYLOAD_BASE]));
-	priv->offset      = ntohl(nla_get_be32(tb[NFTA_PAYLOAD_OFFSET]));
 	priv->len         = ntohl(nla_get_be32(tb[NFTA_PAYLOAD_LEN]));
 
+	err = nft_parse_u32_check(tb[NFTA_PAYLOAD_OFFSET], U16_MAX, &offset);
+	if (err < 0)
+		return err;
+	priv->offset = offset;
+
 	if (tb[NFTA_PAYLOAD_CSUM_TYPE])
 		csum_type = ntohl(nla_get_be32(tb[NFTA_PAYLOAD_CSUM_TYPE]));
 	if (tb[NFTA_PAYLOAD_CSUM_OFFSET]) {
@@ -1069,7 +1073,7 @@ nft_payload_select_ops(const struct nft_ctx *ctx,
 	if (tb[NFTA_PAYLOAD_DREG] == NULL)
 		return ERR_PTR(-EINVAL);
 
-	err = nft_parse_u32_check(tb[NFTA_PAYLOAD_OFFSET], U8_MAX, &offset);
+	err = nft_parse_u32_check(tb[NFTA_PAYLOAD_OFFSET], U16_MAX, &offset);
 	if (err < 0)
 		return ERR_PTR(err);
 
-- 
2.49.1