Hi Tetsuo,
On Tue, Jun 22, 2010 at 09:28:37AM +0900, Tetsuo Handa wrote:
> Kees Cook wrote:
> > + /* require ptrace target be a child of ptracer on attach */
> > + if (mode == PTRACE_MODE_ATTACH && ptrace_scope &&
> > + !capable(CAP_SYS_PTRACE)) {
> > + struct task_struct *walker = child;
> > +
> > + read_lock(&tasklist_lock);
>
> Holding tasklist_lock does not imply rcu protection.
> Don't you need rcu_read_lock() like setpriority() and getppid()?
You're totally right, thanks for the catch! Looks like setpriority() does
a similar kind of thing, so I've wrapped the whole thing in rcu_ now:
...
+ rcu_read_lock();
read_lock(&tasklist_lock);
while (walker->pid > 0) {
...
rc = -EPERM;
read_unlock(&tasklist_lock);
+ rcu_read_unlock();
...
-Kees
--
Kees Cook
Ubuntu Security Team
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
