ptrace: ptrace_check_attach() should not do s/STOPPED/TRACED/
authorOleg Nesterov <oleg@redhat.com>
Fri, 1 Apr 2011 18:13:01 +0000 (20:13 +0200)
committerTejun Heo <tj@kernel.org>
Mon, 4 Apr 2011 00:11:05 +0000 (02:11 +0200)
After "ptrace: Clean transitions between TASK_STOPPED and TRACED"
d79fdd6d96f46fabb779d86332e3677c6f5c2a4f, ptrace_check_attach()
should never see a TASK_STOPPED tracee and s/STOPPED/TRACED/ is
no longer legal. Add the warning.

Note: ptrace_check_attach() can be greatly simplified, in particular
it doesn't need tasklist. But I'd prefer another patch for that.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
kernel/ptrace.c

index 4348586..20d5efd 100644 (file)
@@ -112,16 +112,14 @@ int ptrace_check_attach(struct task_struct *child, int kill)
         */
        read_lock(&tasklist_lock);
        if ((child->ptrace & PT_PTRACED) && child->parent == current) {
-               ret = 0;
                /*
                 * child->sighand can't be NULL, release_task()
                 * does ptrace_unlink() before __exit_signal().
                 */
                spin_lock_irq(&child->sighand->siglock);
-               if (task_is_stopped(child))
-                       child->state = TASK_TRACED;
-               else if (!task_is_traced(child) && !kill)
-                       ret = -ESRCH;
+               WARN_ON_ONCE(task_is_stopped(child));
+               if (task_is_traced(child) || kill)
+                       ret = 0;
                spin_unlock_irq(&child->sighand->siglock);
        }
        read_unlock(&tasklist_lock);