[PATCH v3 3/6] bpf: add bpf_list_add_impl to insert node after a given list node

Chengkaitao posted 6 patches 1 month, 1 week ago
There is a newer version of this series
[PATCH v3 3/6] bpf: add bpf_list_add_impl to insert node after a given list node
Posted by Chengkaitao 1 month, 1 week ago
From: Kaitao Cheng <chengkaitao@kylinos.cn>

Add a new kfunc bpf_list_add_impl(prev, node, meta, off) that inserts
'node' after 'prev' in the BPF linked list. Both must be in the same
list; 'prev' must already be in the list. The new node must be an
owning reference (e.g. from bpf_obj_new); the kfunc consumes that
reference and the node becomes non-owning once inserted.

Returns 0 on success, -EINVAL if prev is not in a list or node is
already in a list (or duplicate insertion). On failure, the kernel
drops the passed-in node.

Signed-off-by: Kaitao Cheng <chengkaitao@kylinos.cn>
---
 kernel/bpf/helpers.c  | 34 ++++++++++++++++++++++++++++++++++
 kernel/bpf/verifier.c | 23 ++++++++++++++++-------
 2 files changed, 50 insertions(+), 7 deletions(-)

diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
index 577af62a9f7a..d212962d4ed6 100644
--- a/kernel/bpf/helpers.c
+++ b/kernel/bpf/helpers.c
@@ -2495,6 +2495,39 @@ __bpf_kfunc struct bpf_list_node *bpf_list_back(struct bpf_list_head *head)
 	return (struct bpf_list_node *)h->prev;
 }
 
+static int __bpf_list_add_after(struct bpf_list_node_kern *prev,
+				struct bpf_list_node_kern *node,
+				struct btf_record *rec, u64 off)
+{
+	struct bpf_list_head *head;
+	struct list_head *n = &node->list_head, *p = &prev->list_head;
+
+	head = READ_ONCE(prev->owner);
+	if (unlikely(!head))
+		goto fail;
+
+	if (cmpxchg(&node->owner, NULL, BPF_PTR_POISON))
+		goto fail;
+
+	list_add(n, p);
+	WRITE_ONCE(node->owner, head);
+	return 0;
+
+fail:
+	__bpf_obj_drop_impl((void *)n - off, rec, false);
+	return -EINVAL;
+}
+
+__bpf_kfunc int bpf_list_add_impl(struct bpf_list_node *prev,
+				  struct bpf_list_node *node,
+				  void *meta__ign, u64 off)
+{
+	struct bpf_list_node_kern *n = (void *)node, *p = (void *)prev;
+	struct btf_struct_meta *meta = meta__ign;
+
+	return __bpf_list_add_after(p, n, meta ? meta->record : NULL, off);
+}
+
 __bpf_kfunc struct bpf_rb_node *bpf_rbtree_remove(struct bpf_rb_root *root,
 						  struct bpf_rb_node *node)
 {
@@ -4564,6 +4597,7 @@ BTF_ID_FLAGS(func, bpf_list_pop_back, KF_ACQUIRE | KF_RET_NULL)
 BTF_ID_FLAGS(func, bpf_list_del, KF_ACQUIRE | KF_RET_NULL)
 BTF_ID_FLAGS(func, bpf_list_front, KF_RET_NULL)
 BTF_ID_FLAGS(func, bpf_list_back, KF_RET_NULL)
+BTF_ID_FLAGS(func, bpf_list_add_impl)
 BTF_ID_FLAGS(func, bpf_task_acquire, KF_ACQUIRE | KF_RCU | KF_RET_NULL)
 BTF_ID_FLAGS(func, bpf_task_release, KF_RELEASE)
 BTF_ID_FLAGS(func, bpf_rbtree_remove, KF_ACQUIRE | KF_RET_NULL)
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 8a782772dd36..f5ee11779a5c 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -12453,6 +12453,7 @@ enum special_kfunc_type {
 	KF_bpf_list_del,
 	KF_bpf_list_front,
 	KF_bpf_list_back,
+	KF_bpf_list_add_impl,
 	KF_bpf_cast_to_kern_ctx,
 	KF_bpf_rdonly_cast,
 	KF_bpf_rcu_read_lock,
@@ -12514,6 +12515,7 @@ BTF_ID(func, bpf_list_pop_back)
 BTF_ID(func, bpf_list_del)
 BTF_ID(func, bpf_list_front)
 BTF_ID(func, bpf_list_back)
+BTF_ID(func, bpf_list_add_impl)
 BTF_ID(func, bpf_cast_to_kern_ctx)
 BTF_ID(func, bpf_rdonly_cast)
 BTF_ID(func, bpf_rcu_read_lock)
@@ -12989,7 +12991,8 @@ static bool is_bpf_list_api_kfunc(u32 btf_id)
 	       btf_id == special_kfunc_list[KF_bpf_list_pop_back] ||
 	       btf_id == special_kfunc_list[KF_bpf_list_del] ||
 	       btf_id == special_kfunc_list[KF_bpf_list_front] ||
-	       btf_id == special_kfunc_list[KF_bpf_list_back];
+	       btf_id == special_kfunc_list[KF_bpf_list_back] ||
+	       btf_id == special_kfunc_list[KF_bpf_list_add_impl];
 }
 
 static bool is_bpf_rbtree_api_kfunc(u32 btf_id)
@@ -13111,7 +13114,8 @@ static bool check_kfunc_is_graph_node_api(struct bpf_verifier_env *env,
 	case BPF_LIST_NODE:
 		ret = (kfunc_btf_id == special_kfunc_list[KF_bpf_list_push_front_impl] ||
 		       kfunc_btf_id == special_kfunc_list[KF_bpf_list_push_back_impl] ||
-		       kfunc_btf_id == special_kfunc_list[KF_bpf_list_del]);
+		       kfunc_btf_id == special_kfunc_list[KF_bpf_list_del] ||
+		       kfunc_btf_id == special_kfunc_list[KF_bpf_list_add_impl]);
 		break;
 	case BPF_RB_NODE:
 		ret = (kfunc_btf_id == special_kfunc_list[KF_bpf_rbtree_remove] ||
@@ -13263,12 +13267,15 @@ __process_kf_arg_ptr_to_graph_node(struct bpf_verifier_env *env,
 		return -EINVAL;
 	}
 
-	/* bpf_list_del: require list head's lock. Use refs[] REF_TYPE_LOCK_MASK
-	 * only. At lock time we stored the locked object's btf_record in ref->
-	 * lock_rec, so we can get the list value type from the ref directly.
+	/* When there is no bpf_list_head in the parameter list, to prevent BPF
+	 * programs from calling bpf_list APIs without holding the spinlock,
+	 * we need to acquire the list head's lock. At lock time we stored the
+	 * locked object's btf_record in ref->lock_rec, so we can get the list
+	 * value type from the ref directly.
 	 */
 	if (node_field_type == BPF_LIST_NODE &&
-	    meta->func_id == special_kfunc_list[KF_bpf_list_del]) {
+	    (meta->func_id == special_kfunc_list[KF_bpf_list_del] ||
+	     meta->func_id == special_kfunc_list[KF_bpf_list_add_impl])) {
 		struct bpf_verifier_state *cur = env->cur_state;
 
 		for (int i = 0; i < cur->acquired_refs; i++) {
@@ -13281,7 +13288,7 @@ __process_kf_arg_ptr_to_graph_node(struct bpf_verifier_env *env,
 							reg->btf, reg->btf_id))
 				return 0;
 		}
-		verbose(env, "bpf_spin_lock must be held for bpf_list_del\n");
+		verbose(env, "bpf_spin_lock must be held for bpf_list api\n");
 		return -EINVAL;
 	}
 
@@ -14278,6 +14285,7 @@ static int check_kfunc_call(struct bpf_verifier_env *env, struct bpf_insn *insn,
 
 	if (meta.func_id == special_kfunc_list[KF_bpf_list_push_front_impl] ||
 	    meta.func_id == special_kfunc_list[KF_bpf_list_push_back_impl] ||
+	    meta.func_id == special_kfunc_list[KF_bpf_list_add_impl] ||
 	    meta.func_id == special_kfunc_list[KF_bpf_rbtree_add_impl]) {
 		release_ref_obj_id = regs[BPF_REG_2].ref_obj_id;
 		insn_aux->insert_off = regs[BPF_REG_2].off;
@@ -23244,6 +23252,7 @@ static int fixup_kfunc_call(struct bpf_verifier_env *env, struct bpf_insn *insn,
 		*cnt = 3;
 	} else if (desc->func_id == special_kfunc_list[KF_bpf_list_push_back_impl] ||
 		   desc->func_id == special_kfunc_list[KF_bpf_list_push_front_impl] ||
+		   desc->func_id == special_kfunc_list[KF_bpf_list_add_impl] ||
 		   desc->func_id == special_kfunc_list[KF_bpf_rbtree_add_impl]) {
 		struct btf_struct_meta *kptr_struct_meta = env->insn_aux_data[insn_idx].kptr_struct_meta;
 		int struct_meta_reg = BPF_REG_3;
-- 
2.50.1 (Apple Git-155)