apparmor: don't try to replace stale label in ptrace access check

As a comment above begin_current_label_crit_section() explains,
begin_current_label_crit_section() must run in sleepable context because
when label_is_stale() is true, aa_replace_current_label() runs, which uses
prepare_creds(), which can sleep.
Until now, the ptrace access check (which runs with a task lock held)
violated this rule.

Also add a might_sleep() assertion to begin_current_label_crit_section(),
because asserts are less likely to be ignored than comments.

Fixes: b2d09ae ("apparmor: move ptrace checks to using labels")
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>

security/apparmor/include/cred.h

@@ -151,6 +151,8 @@ static inline struct aa_label *begin_current_label_crit_section(void)
 {
 	struct aa_label *label = aa_current_raw_label();
 
+	might_sleep();
+
 	if (label_is_stale(label)) {
 		label = aa_get_newest_label(label);
 		if (aa_replace_current_label(label) == 0)

security/apparmor/lsm.c

@@ -114,13 +114,13 @@ static int apparmor_ptrace_access_check(struct task_struct *child,
 	struct aa_label *tracer, *tracee;
 	int error;
 
-	tracer = begin_current_label_crit_section();
+	tracer = __begin_current_label_crit_section();
 	tracee = aa_get_task_label(child);
 	error = aa_may_ptrace(tracer, tracee,
 			(mode & PTRACE_MODE_READ) ? AA_PTRACE_READ
 						  : AA_PTRACE_TRACE);
 	aa_put_label(tracee);
-	end_current_label_crit_section(tracer);
+	__end_current_label_crit_section(tracer);
 
 	return error;
 }