[PATCH bpf-next] bpf: Tighten ptr_to_btf_id checks.

[PATCH bpf-next] bpf: Tighten ptr_to_btf_id checks.

[Alexei Starovoitov]

From: Alexei Starovoitov <ast@kernel.org>

The networking programs typically don't require CAP_PERFMON, but through kfuncs
like bpf_cast_to_kern_ctx() they can access memory through PTR_TO_BTF_ID. In
such case enforce CAP_PERFMON. Also make sure that those programs are GPL if
they access kernel data structures. All kfuncs require GPL anyway.

Also remove allow_ptr_to_map_access. It's the same as allow_ptr_leaks and
different name for the same check only causes confusion.

Fixes: fd264ca02094 ("bpf: Add a kfunc to type cast from bpf uapi ctx to kernel ctx")
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
---
 include/linux/bpf.h          |  5 -----
 include/linux/bpf_verifier.h |  1 -
 kernel/bpf/verifier.c        | 17 ++++++++++++++---
 3 files changed, 14 insertions(+), 9 deletions(-)

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index c6aa6912ea16..4235ac4ed1c8 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -1891,11 +1891,6 @@ static inline bool bpf_allow_uninit_stack(void)
 	return perfmon_capable();
 }
 
-static inline bool bpf_allow_ptr_to_map_access(void)
-{
-	return perfmon_capable();
-}
-
 static inline bool bpf_bypass_spec_v1(void)
 {
 	return perfmon_capable();
diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h
index c05aa6e1f6f5..b5090e89cb3f 100644
--- a/include/linux/bpf_verifier.h
+++ b/include/linux/bpf_verifier.h
@@ -531,7 +531,6 @@ struct bpf_verifier_env {
 	bool explore_alu_limits;
 	bool allow_ptr_leaks;
 	bool allow_uninit_stack;
-	bool allow_ptr_to_map_access;
 	bool bpf_capable;
 	bool bypass_spec_v1;
 	bool bypass_spec_v4;
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 6599d25dae38..69040c09f4f5 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -4703,6 +4703,18 @@ static int check_ptr_to_btf_access(struct bpf_verifier_env *env,
 	u32 btf_id;
 	int ret;
 
+	if (!env->allow_ptr_leaks) {
+		verbose(env,
+			"'struct %s' access is allowed only to CAP_PERFMON and CAP_SYS_ADMIN\n",
+			tname);
+		return -EPERM;
+	}
+	if (!env->prog->gpl_compatible && btf_is_kernel(reg->btf)) {
+		verbose(env,
+			"Cannot access kernel 'struct %s' from non-GPL compatible program\n",
+			tname);
+		return -EINVAL;
+	}
 	if (off < 0) {
 		verbose(env,
 			"R%d is ptr_%s invalid negative access: off=%d\n",
@@ -4823,9 +4835,9 @@ static int check_ptr_to_map_access(struct bpf_verifier_env *env,
 	t = btf_type_by_id(btf_vmlinux, *map->ops->map_btf_id);
 	tname = btf_name_by_offset(btf_vmlinux, t->name_off);
 
-	if (!env->allow_ptr_to_map_access) {
+	if (!env->allow_ptr_leaks) {
 		verbose(env,
-			"%s access is allowed only to CAP_PERFMON and CAP_SYS_ADMIN\n",
+			"'struct %s' access is allowed only to CAP_PERFMON and CAP_SYS_ADMIN\n",
 			tname);
 		return -EPERM;
 	}
@@ -16675,7 +16687,6 @@ int bpf_check(struct bpf_prog **prog, union bpf_attr *attr, bpfptr_t uattr)
 
 	env->allow_ptr_leaks = bpf_allow_ptr_leaks();
 	env->allow_uninit_stack = bpf_allow_uninit_stack();
-	env->allow_ptr_to_map_access = bpf_allow_ptr_to_map_access();
 	env->bypass_spec_v1 = bpf_bypass_spec_v1();
 	env->bypass_spec_v4 = bpf_bypass_spec_v4();
 	env->bpf_capable = bpf_capable();
-- 
2.30.2

Re: [PATCH bpf-next] bpf: Tighten ptr_to_btf_id checks.

[Yonghong Song]

On 11/25/22 10:35 AM, Alexei Starovoitov wrote:
> From: Alexei Starovoitov <ast@kernel.org>
> 
> The networking programs typically don't require CAP_PERFMON, but through kfuncs
> like bpf_cast_to_kern_ctx() they can access memory through PTR_TO_BTF_ID. In
> such case enforce CAP_PERFMON. Also make sure that those programs are GPL if
> they access kernel data structures. All kfuncs require GPL anyway.
> 
> Also remove allow_ptr_to_map_access. It's the same as allow_ptr_leaks and
> different name for the same check only causes confusion.
> 
> Fixes: fd264ca02094 ("bpf: Add a kfunc to type cast from bpf uapi ctx to kernel ctx")
> Signed-off-by: Alexei Starovoitov <ast@kernel.org>

Acked-by: Yonghong Song <yhs@fb.com>