Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 17 Dec 2020 10:04:16 +0000
From:      bugzilla-noreply@freebsd.org
To:        bugs@FreeBSD.org
Subject:   [Bug 251915] TOCTOU race between tty_signal_sessleader() and killjobc()
Message-ID:  <bug-251915-227@https.bugs.freebsd.org/bugzilla/>

next in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D251915

            Bug ID: 251915
           Summary: TOCTOU race between tty_signal_sessleader() and
                    killjobc()
           Product: Base System
           Version: CURRENT
          Hardware: Any
                OS: Any
            Status: New
          Severity: Affects Some People
          Priority: ---
         Component: kern
          Assignee: bugs@FreeBSD.org
          Reporter: j.piecuch96@gmail.com

In tty_signal_sessleader():

if (tp->t_session !=3D NULL && tp->t_session->s_leader !=3D NULL) {
        p =3D tp->t_session->s_leader;
        PROC_LOCK(p);
        kern_psignal(p, sig);
        PROC_UNLOCK(p);
}

We're holding the tty lock, but not the session lock, so the s_leader may be
changed to NULL right after the !=3D NULL check by a concurrent invocation =
of
killjobc() by the session leader. The compiler *might* optimize this and on=
ly
read s_leader a single time, but that's far from guaranteed.

I don't have a patch because I'm not sure what the right way to deal with t=
his
is.
We could read s_leader a single time, like this:

if (tp->t_session !=3D NULL && (p =3D tp->t_session->s_leader) !=3D NULL) {
        PROC_LOCK(p);
        kern_psignal(p, sig);
        PROC_UNLOCK(p);
}

...but the compiler may in theory still output vulnerable code. I don't know
what assumptions are made in FreeBSD about what compilers can and can't do.

--=20
You are receiving this mail because:
You are the assignee for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-251915-227>