[WIP 3/4] bpf: Add a way to mark functions as requiring privilege

[Andy Lutomirski <luto@xxxxxxxxxx>]

This is horribly incomplete:

 - I only marked one function as requiring privilege, and there are
   surely more.

 - Checking is_priv is probably not the right thing to do.  This should
   probably do something more clever.  At the very lease, it needs to
   integrate with the upcoming lockdown LSM infrastructure.

 - The seen_privileged_funcs mechanism is probably not a good solution.
   Instead we should check something while we still have enough context
   to give a good error message.  But we *don't* want to check for
   capabilities up front before even seeing a function call, since we
   don't want to inadvertently generate audit events for privileges that
   are never used.

So it's the idea that counts :)

Signed-off-by: Andy Lutomirski <luto@xxxxxxxxxx>
---
 include/linux/bpf.h          | 15 +++++++++++++++
 include/linux/bpf_verifier.h |  1 +
 kernel/bpf/verifier.c        |  8 ++++++++
 kernel/trace/bpf_trace.c     |  1 +
 4 files changed, 25 insertions(+)

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 2d5e1a4dff6c..de31b9888b6c 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -229,6 +229,7 @@ struct bpf_func_proto {
 	u64 (*func)(u64 r1, u64 r2, u64 r3, u64 r4, u64 r5);
 	bool gpl_only;
 	bool pkt_access;
+	u16 privilege;
 	enum bpf_return_type ret_type;
 	enum bpf_arg_type arg1_type;
 	enum bpf_arg_type arg2_type;
@@ -237,6 +238,20 @@ struct bpf_func_proto {
 	enum bpf_arg_type arg5_type;
 };
 
+/*
+ * Some functions should require privilege to call at all, even in a test
+ * run.  These flags indicate why privilege is required.  The core BPF
+ * code will verify that the creator of such a program has the requisite
+ * privilege.
+ *
+ * NB: This means that anyone who creates a privileged program (due to
+ * such a call or due to a privilege-requiring pointer-to-integer conversion)
+ * is responsible for restricting access to the program in an appropriate
+ * manner.
+ */
+#define BPF_FUNC_PRIV_READ_KERNEL_MEMORY BIT(0)
+#define BPT_FUNC_PRIV_WRITE_GLOBAL_LOGS BIT(1)
+
 /* bpf_context is intentionally undefined structure. Pointer to bpf_context is
  * the first argument to eBPF programs.
  * For socket filters: 'struct bpf_context *' == 'struct sk_buff *'
diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h
index 5fe99f322b1c..9877f5753cf4 100644
--- a/include/linux/bpf_verifier.h
+++ b/include/linux/bpf_verifier.h
@@ -363,6 +363,7 @@ struct bpf_verifier_env {
 	u32 id_gen;			/* used to generate unique reg IDs */
 	bool allow_ptr_leaks;
 	bool seen_direct_write;
+	u16 seen_privileged_funcs;
 	struct bpf_insn_aux_data *insn_aux_data; /* array of per-insn state */
 	const struct bpf_line_info *prev_linfo;
 	struct bpf_verifier_log log;
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 5900cbb966b1..5e048688fd8d 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -4129,6 +4129,9 @@ static int check_helper_call(struct bpf_verifier_env *env, int func_id, int insn
 
 	if (changes_data)
 		clear_all_pkt_pointers(env);
+
+	env->seen_privileged_funcs |= fn->privilege;
+
 	return 0;
 }
 
@@ -9371,6 +9374,11 @@ int bpf_check(struct bpf_prog **prog, union bpf_attr *attr,
 	if (ret == 0)
 		adjust_btf_func(env);
 
+	if (env->seen_privileged_funcs && !is_priv) {
+		ret = -EPERM;
+		goto err_release_maps;
+	}
+
 err_release_maps:
 	if (!env->prog->aux->used_maps)
 		/* if we didn't copy map pointers into bpf_prog_info, release
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index ca1255d14576..d9454588d9e8 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -152,6 +152,7 @@ BPF_CALL_3(bpf_probe_read, void *, dst, u32, size, const void *, unsafe_ptr)
 static const struct bpf_func_proto bpf_probe_read_proto = {
 	.func		= bpf_probe_read,
 	.gpl_only	= true,
+	.privilege	= BPF_FUNC_PRIV_READ_KERNEL_MEMORY,
 	.ret_type	= RET_INTEGER,
 	.arg1_type	= ARG_PTR_TO_UNINIT_MEM,
 	.arg2_type	= ARG_CONST_SIZE_OR_ZERO,
-- 
2.21.0