[bpf-next,v2,03/14] bpf: Allow struct_ops prog to return referenced kptr

Commit Message

Amery Hung Dec. 20, 2024, 7:55 p.m. UTC

From: Amery Hung <[e-mail]>

Allow a struct_ops program to return a referenced kptr if the struct_ops
operator's return type is a struct pointer. To make sure the returned
pointer continues to be valid in the kernel, several constraints are
required:

1) The type of the pointer must matches the return type
2) The pointer originally comes from the kernel (not locally allocated)
3) The pointer is in its unmodified form

Implementation wise, a referenced kptr first needs to be allowed to _leak_
in check_reference_leak() if it is in the return register. Then, in
check_return_code(), constraints 1-3 are checked. During struct_ops
registration, a check is also added to warn about operators with
non-struct pointer return.

In addition, since the first user, Qdisc_ops::dequeue, allows a NULL
pointer to be returned when there is no skb to be dequeued, we will allow
a scalar value with value equals to NULL to be returned.

In the future when there is a struct_ops user that always expects a valid
pointer to be returned from an operator, we may extend tagging to the
return value. We can tell the verifier to only allow NULL pointer return
if the return value is tagged with MAY_BE_NULL.

Signed-off-by: Amery Hung <[e-mail]>
---
 kernel/bpf/bpf_struct_ops.c | 12 +++++++++++-
 kernel/bpf/verifier.c       | 36 ++++++++++++++++++++++++++++++++----
 2 files changed, 43 insertions(+), 5 deletions(-)

Patch

diff --git a/kernel/bpf/bpf_struct_ops.c b/kernel/bpf/bpf_struct_ops.c
index d9e0af00580b..27d4a170df84 100644
--- a/kernel/bpf/bpf_struct_ops.c
+++ b/kernel/bpf/bpf_struct_ops.c
@@ -386,7 +386,7 @@  int bpf_struct_ops_desc_init(struct bpf_struct_ops_desc *st_ops_desc,
 	st_ops_desc->value_type = btf_type_by_id(btf, value_id);
 
 	for_each_member(i, t, member) {
-		const struct btf_type *func_proto;
+		const struct btf_type *func_proto, *ret_type;
 
 		mname = btf_name_by_offset(btf, member->name_off);
 		if (!*mname) {
@@ -409,6 +409,16 @@  int bpf_struct_ops_desc_init(struct bpf_struct_ops_desc *st_ops_desc,
 		if (!func_proto)
 			continue;
 
+		if (func_proto->type) {
+			ret_type = btf_type_resolve_ptr(btf, func_proto->type, NULL);
+			if (ret_type && !__btf_type_is_struct(ret_type)) {
+				pr_warn("func ptr %s in struct %s returns non-struct pointer, which is not supported\n",
+					mname, st_ops->name);
+				err = -EOPNOTSUPP;
+				goto errout;
+			}
+		}
+
 		if (btf_distill_func_proto(log, btf,
 					   func_proto, mname,
 					   &st_ops->func_models[i])) {
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 26305571e377..0e6a3c4daa7d 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -10707,6 +10707,8 @@  record_func_key(struct bpf_verifier_env *env, struct bpf_call_arg_meta *meta,
 static int check_reference_leak(struct bpf_verifier_env *env, bool exception_exit)
 {
 	struct bpf_verifier_state *state = env->cur_state;
+	enum bpf_prog_type type = resolve_prog_type(env->prog);
+	struct bpf_reg_state *reg = reg_state(env, BPF_REG_0);
 	bool refs_lingering = false;
 	int i;
 
@@ -10716,6 +10718,12 @@  static int check_reference_leak(struct bpf_verifier_env *env, bool exception_exi
 	for (i = 0; i < state->acquired_refs; i++) {
 		if (state->refs[i].type != REF_TYPE_PTR)
 			continue;
+		/* Allow struct_ops programs to return a referenced kptr back to
+		 * kernel. Type checks are performed later in check_return_code.
+		 */
+		if (type == BPF_PROG_TYPE_STRUCT_OPS && !exception_exit &&
+		    reg->ref_obj_id == state->refs[i].id)
+			continue;
 		verbose(env, "Unreleased reference id=%d alloc_insn=%d\n",
 			state->refs[i].id, state->refs[i].insn_idx);
 		refs_lingering = true;
@@ -16320,13 +16328,14 @@  static int check_return_code(struct bpf_verifier_env *env, int regno, const char
 	const char *exit_ctx = "At program exit";
 	struct tnum enforce_attach_type_range = tnum_unknown;
 	const struct bpf_prog *prog = env->prog;
-	struct bpf_reg_state *reg;
+	struct bpf_reg_state *reg = reg_state(env, regno);
 	struct bpf_retval_range range = retval_range(0, 1);
 	enum bpf_prog_type prog_type = resolve_prog_type(env->prog);
 	int err;
 	struct bpf_func_state *frame = env->cur_state->frame[0];
 	const bool is_subprog = frame->subprogno;
 	bool return_32bit = false;
+	const struct btf_type *reg_type, *ret_type = NULL;
 
 	/* LSM and struct_ops func-ptr's return type could be "void" */
 	if (!is_subprog || frame->in_exception_callback_fn) {
@@ -16335,10 +16344,26 @@  static int check_return_code(struct bpf_verifier_env *env, int regno, const char
 			if (prog->expected_attach_type == BPF_LSM_CGROUP)
 				/* See below, can be 0 or 0-1 depending on hook. */
 				break;
-			fallthrough;
+			if (!prog->aux->attach_func_proto->type)
+				return 0;
+			break;
 		case BPF_PROG_TYPE_STRUCT_OPS:
 			if (!prog->aux->attach_func_proto->type)
 				return 0;
+
+			if (frame->in_exception_callback_fn)
+				break;
+
+			/* Allow a struct_ops program to return a referenced kptr if it
+			 * matches the operator's return type and is in its unmodified
+			 * form. A scalar zero (i.e., a null pointer) is also allowed.
+			 */
+			reg_type = reg->btf ? btf_type_by_id(reg->btf, reg->btf_id) : NULL;
+			ret_type = btf_type_resolve_ptr(prog->aux->attach_btf,
+							prog->aux->attach_func_proto->type,
+							NULL);
+			if (ret_type && ret_type == reg_type && reg->ref_obj_id)
+				return __check_ptr_off_reg(env, reg, regno, false);
 			break;
 		default:
 			break;
@@ -16360,8 +16385,6 @@  static int check_return_code(struct bpf_verifier_env *env, int regno, const char
 		return -EACCES;
 	}
 
-	reg = cur_regs(env) + regno;
-
 	if (frame->in_async_callback_fn) {
 		/* enforce return zero from async callbacks like timer */
 		exit_ctx = "At async callback return";
@@ -16460,6 +16483,11 @@  static int check_return_code(struct bpf_verifier_env *env, int regno, const char
 	case BPF_PROG_TYPE_NETFILTER:
 		range = retval_range(NF_DROP, NF_ACCEPT);
 		break;
+	case BPF_PROG_TYPE_STRUCT_OPS:
+		if (!ret_type)
+			return 0;
+		range = retval_range(0, 0);
+		break;
 	case BPF_PROG_TYPE_EXT:
 		/* freplace program can return anything as its return value
 		 * depends on the to-be-replaced kernel func or bpf program.