Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 11 Jul 2014 12:55:51 +0300
From:      Konstantin Belousov <kostikbel@gmail.com>
To:        Mateusz Guzik <mjguzik@gmail.com>
Cc:        svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org, Mateusz Guzik <mjg@FreeBSD.org>
Subject:   Re: svn commit: r267760 - head/sys/kern
Message-ID:  <20140711095551.GA93733@kib.kiev.ua>
In-Reply-To: <20140711024351.GA18214@dft-labs.eu>
References:  <201406230128.s5N1SIYK097224@svn.freebsd.org> <20140623064044.GD93733@kib.kiev.ua> <20140623070652.GA27040@dft-labs.eu> <20140623072519.GE93733@kib.kiev.ua> <20140623080501.GB27040@dft-labs.eu> <20140623081823.GG93733@kib.kiev.ua> <20140623131653.GC27040@dft-labs.eu> <20140623163523.GK93733@kib.kiev.ua> <20140711024351.GA18214@dft-labs.eu>

next in thread | previous in thread | raw e-mail | index | archive | help

--HrG051WEavH+0eil
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Jul 11, 2014 at 04:43:51AM +0200, Mateusz Guzik wrote:
> On Mon, Jun 23, 2014 at 07:35:23PM +0300, Konstantin Belousov wrote:
> > On Mon, Jun 23, 2014 at 03:16:53PM +0200, Mateusz Guzik wrote:
> > > If traversal while transition to P_INEXEC is allowed, execve dealing
> > > with a setuid binary is problematic. This is more of hypothetical nat=
ure,
> > > but with sufficienly long delay it could finish the syscall and start
> > > opening some files, which paths would now be visible for an unprivile=
ged
> > > reader.
> > >=20
> > > That said, I propose adding a counter to struct proc which would which
> > > would block execve. It would be quite similar to p_lock.
> > I thought about this too.  In fact, I considered using PHOLD for this.
> >=20
> > >=20
> > > iow execve would:
> > >=20
> > >         PROC_LOCK(p);
> > > 	p->p_flag |=3D P_INEXEC;=20
> > >         while (p->p_execlock > 0)
> > >                 msleep(&p->p_execlock, &p->p_mtx, PWAIT, "execlock", =
0);
> > > 	PROC_UNLOCK(p);
> > >=20
> > > And it would be mandatory for external fdp consumers to grab the coun=
ter.
> > >=20
> > > I'm tempted to add P_GETPIN which would both increase p_lock and p_ex=
eclock,
> > > that way the process is guaranteed not to exit and not to execve even
> > > after proc lock is dropped.
> > See above about PHOLD.
> >=20
> > >=20
> > > There is a separate question if p_execlock should be renamed and
> > > extended to also block any kind of credential changes.
> > >=20
> > > Then the guarantee is even stronger since we know that credentials we
> > > checked against are not going to change for the duration of our
> > > operations, but it is unclear if we need this.
> >=20
> > If doing separate execlock/p_lock, I think that it could be possible
> > to use per-process sx lock instead of hand-rolling the counter.  The
> > accessors would lock sx shared, while kern_execve would take it in
> > exclusive mode.
>=20
> Both patches need some cleaning up. The name 'keeplock' is no exactly
> the best either.
>=20
> In both cases the same mechanism blocks both exec and exit, this can be
> split if needed (p_lock would still cover exit, p_something would cover
> exec).
>=20
> Here is a version with sx lock:
>=20
> http://people.freebsd.org/~mjg/patches/exec-exit-hold-wait.patch
>=20
> I'm not really happy with this. Reading foreign fdt is very rare and
> this adds lock + unlock for every exec and exit.
>=20
> On the other hand mere counter version is rather simple:
>=20
> http://people.freebsd.org/~mjg/patches/exec-exit-hold-nolock.patch
>=20
> I don't have strong opinion here, but prefer the latter.

I suggest the name 'imagelock' for the beast.

The nolock version requires two atomics on both entry and leave from the
protected region, while sx-locked variant requires only one atomic for
entry and leave.

I am not sure why you decided to acquire p->p_keeplock in after the
proc lock in pget(), which indeed causes the complications of dropping
the proc_lock and rechecking to avoid LOR.  Did you tried to add a flag
to pfind*() functions to indicate that p_keeplock should be acquired,
instead ?

--HrG051WEavH+0eil
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=QOFj
-----END PGP SIGNATURE-----

--HrG051WEavH+0eil--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20140711095551.GA93733>