From: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
To: Oleg Nesterov <oleg@redhat.com>
Cc: serge@hallyn.com,
syzbot <syzbot+a9ac39bf55329e206219@syzkaller.appspotmail.com>,
jmorris@namei.org, keescook@chromium.org,
linux-kernel@vger.kernel.org,
linux-security-module@vger.kernel.org,
syzkaller-bugs@googlegroups.com
Subject: Re: KASAN: use-after-free Read in task_is_descendant
Date: Thu, 25 Oct 2018 11:15:48 +0900 [thread overview]
Message-ID: <201810250215.w9P2Fm2M078167@www262.sakura.ne.jp> (raw)
In-Reply-To: <20181022134634.GA7358@redhat.com>
Oleg Nesterov wrote:
> On 10/22, Tetsuo Handa wrote:
> > > And again, I do not know how/if yama ensures that child is rcu-protected, perhaps
> > > task_is_descendant() needs to check pid_alive(child) right after rcu_read_lock() ?
> >
> > Since the caller (ptrace() path) called get_task_struct(child), child itself can't be
> > released. Do we still need pid_alive(child) ?
>
> get_task_struct(child) can only ensure that this task_struct can't be freed.
The report says that it is a use-after-free read at
walker = rcu_dereference(walker->real_parent);
which means that walker was already released.
>
> Suppose that this child exits after get_task_struct(), then its real_parent exits
> too and calls call_rcu(delayed_put_task_struct).
>
> Now, when task_is_descendant() is called, rcu_read_lock() can happen after rcu gp,
> iow child->parent can be already freed/reused/unmapped.
>
> We need to ensure that child is still protected by RCU.
I wonder whether pid_alive() test helps.
We can get
[ 40.620318] parent or walker is dead.
[ 40.624146] tracee is dead.
messages using below patch and reproducer.
----------
diff --git a/kernel/ptrace.c b/kernel/ptrace.c
index 99cfddd..0d9d786 100644
--- a/kernel/ptrace.c
+++ b/kernel/ptrace.c
@@ -385,6 +385,7 @@ static int ptrace_attach(struct task_struct *task, long request,
if (mutex_lock_interruptible(&task->signal->cred_guard_mutex))
goto out;
+ schedule_timeout_killable(HZ);
task_lock(task);
retval = __ptrace_may_access(task, PTRACE_MODE_ATTACH_REALCREDS);
task_unlock(task);
diff --git a/security/yama/yama_lsm.c b/security/yama/yama_lsm.c
index ffda91a..a231ec6 100644
--- a/security/yama/yama_lsm.c
+++ b/security/yama/yama_lsm.c
@@ -283,6 +283,11 @@ static int task_is_descendant(struct task_struct *parent,
return 0;
rcu_read_lock();
+ if (!pid_alive(parent) || !pid_alive(walker)) {
+ rcu_read_unlock();
+ printk("parent or walker is dead.\n");
+ return 0;
+ }
if (!thread_group_leader(parent))
parent = rcu_dereference(parent->group_leader);
while (walker->pid > 0) {
@@ -315,6 +320,10 @@ static int ptracer_exception_found(struct task_struct *tracer,
bool found = false;
rcu_read_lock();
+ if (!pid_alive(tracee)) {
+ printk("tracee is dead.\n");
+ goto unlock;
+ }
/*
* If there's already an active tracing relationship, then make an
----------
----------
#include <unistd.h>
#include <sys/ptrace.h>
#include <poll.h>
int main(int argc, char *argv[])
{
if (fork() == 0) {
ptrace(PTRACE_ATTACH, getppid(), NULL, NULL);
_exit(0);
}
poll(NULL, 0, 100);
return 0;
}
----------
But since "child" has at least one reference, reading "child->real_parent" should
be safe. Therefore, I think that bailing out due to pid_is_alive(child) == false
(like above patch does) cannot avoid this problem...
next prev parent reply other threads:[~2018-10-25 2:16 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-10-21 7:10 KASAN: use-after-free Read in task_is_descendant syzbot
2018-10-21 7:12 ` Tetsuo Handa
2018-10-22 9:54 ` Oleg Nesterov
2018-10-22 10:06 ` Tetsuo Handa
2018-10-22 13:46 ` Oleg Nesterov
2018-10-25 2:15 ` Tetsuo Handa [this message]
2018-10-25 11:13 ` Oleg Nesterov
2018-10-25 11:36 ` Kees Cook
2018-10-25 12:05 ` Oleg Nesterov
2018-10-25 11:47 ` Tetsuo Handa
2018-10-25 12:17 ` Oleg Nesterov
2018-10-25 13:01 ` Oleg Nesterov
2018-10-26 16:09 ` Kees Cook
2018-10-29 12:23 ` Oleg Nesterov
2018-10-29 15:05 ` yama: unsafe usage of ptrace_relation->tracer Oleg Nesterov
2019-01-10 11:05 ` Tetsuo Handa
2019-01-10 18:47 ` Kees Cook
2019-01-16 17:40 ` Oleg Nesterov
2018-10-25 13:14 ` KASAN: use-after-free Read in task_is_descendant Tetsuo Handa
2018-10-25 15:55 ` Oleg Nesterov
2018-10-25 16:25 ` Oleg Nesterov
2018-10-26 12:23 ` Tetsuo Handa
2018-10-26 13:04 ` Oleg Nesterov
2018-10-26 13:51 ` Tetsuo Handa
2018-10-26 14:39 ` Oleg Nesterov
2018-10-26 15:04 ` Tetsuo Handa
2018-10-26 15:22 ` Oleg Nesterov
2018-10-25 8:19 ` Kees Cook
2018-10-25 11:52 ` Oleg Nesterov
2018-11-10 3:25 ` syzbot
2018-11-10 11:46 ` syzbot
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=201810250215.w9P2Fm2M078167@www262.sakura.ne.jp \
--to=penguin-kernel@i-love.sakura.ne.jp \
--cc=jmorris@namei.org \
--cc=keescook@chromium.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-security-module@vger.kernel.org \
--cc=oleg@redhat.com \
--cc=serge@hallyn.com \
--cc=syzbot+a9ac39bf55329e206219@syzkaller.appspotmail.com \
--cc=syzkaller-bugs@googlegroups.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).