All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Theodore Ts'o" <tytso@mit.edu>
To: "Serge E. Hallyn" <serge@hallyn.com>
Cc: David Miller <davem@davemloft.net>,
	gnomes@lxorguk.ukuu.org.uk, keescook@chromium.org,
	mcgrof@kernel.org, tixxdz@gmail.com, luto@kernel.org,
	akpm@linux-foundation.org, james.l.morris@oracle.com,
	ben.hutchings@codethink.co.uk, solar@openwall.com,
	jeyu@kernel.org, rusty@rustcorp.com.au,
	linux-kernel@vger.kernel.org,
	linux-security-module@vger.kernel.org,
	kernel-hardening@lists.openwall.com, corbet@lwn.net,
	mingo@kernel.org, netdev@vger.kernel.org, peterz@infradead.org,
	torvalds@linux-foundation.org
Subject: Re: [PATCH v5 next 1/5] modules:capabilities: add request_module_cap()
Date: Wed, 29 Nov 2017 19:35:31 -0500	[thread overview]
Message-ID: <20171130003531.gwpl22bxmweifjz2@thunk.org> (raw)
In-Reply-To: <20171129172852.GA14545@mail.hallyn.com>

On Wed, Nov 29, 2017 at 11:28:52AM -0600, Serge E. Hallyn wrote:
> 
> Just to be clear, module loading requires - and must always continue to
> require - CAP_SYS_MODULE against the initial user namespace.  Containers
> in user namespaces do not have that.
> 
> I don't believe anyone has ever claimed that containers which are not in
> a user namespace are in any way secure.

Unless the container performs some action which causes the kernel to
call request_module(), which then loads some kernel module,
potentially containing cr*p unmaintained code which was included when
the distro compiled the world, into the host kernel.

This is an attack vector that doesn't exist if you are using VM's.
And in general, the attack surface of the entire Linux
kernel<->userspace API is far larger than that which is exposed by the
guest<->host interface.

For that reason, containers are *far* more insecure than VM's, since
once the attacker gets root on the guest VM, they then have to attack
the hypervisor interface.  And if you compare the attack surface of
the two, it's pretty clear which is larger, and it's not the
hypervisor interface.

						- Ted

WARNING: multiple messages have this Message-ID (diff)
From: tytso@mit.edu (Theodore Ts'o)
To: linux-security-module@vger.kernel.org
Subject: [PATCH v5 next 1/5] modules:capabilities: add request_module_cap()
Date: Wed, 29 Nov 2017 19:35:31 -0500	[thread overview]
Message-ID: <20171130003531.gwpl22bxmweifjz2@thunk.org> (raw)
In-Reply-To: <20171129172852.GA14545@mail.hallyn.com>

On Wed, Nov 29, 2017 at 11:28:52AM -0600, Serge E. Hallyn wrote:
> 
> Just to be clear, module loading requires - and must always continue to
> require - CAP_SYS_MODULE against the initial user namespace.  Containers
> in user namespaces do not have that.
> 
> I don't believe anyone has ever claimed that containers which are not in
> a user namespace are in any way secure.

Unless the container performs some action which causes the kernel to
call request_module(), which then loads some kernel module,
potentially containing cr*p unmaintained code which was included when
the distro compiled the world, into the host kernel.

This is an attack vector that doesn't exist if you are using VM's.
And in general, the attack surface of the entire Linux
kernel<->userspace API is far larger than that which is exposed by the
guest<->host interface.

For that reason, containers are *far* more insecure than VM's, since
once the attacker gets root on the guest VM, they then have to attack
the hypervisor interface.  And if you compare the attack surface of
the two, it's pretty clear which is larger, and it's not the
hypervisor interface.

						- Ted
--
To unsubscribe from this list: send the line "unsubscribe linux-security-module" in
the body of a message to majordomo at vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

WARNING: multiple messages have this Message-ID (diff)
From: Theodore Ts'o <tytso@mit.edu>
To: "Serge E. Hallyn" <serge@hallyn.com>
Cc: David Miller <davem@davemloft.net>,
	gnomes@lxorguk.ukuu.org.uk, keescook@chromium.org,
	mcgrof@kernel.org, tixxdz@gmail.com, luto@kernel.org,
	akpm@linux-foundation.org, james.l.morris@oracle.com,
	ben.hutchings@codethink.co.uk, solar@openwall.com,
	jeyu@kernel.org, rusty@rustcorp.com.au,
	linux-kernel@vger.kernel.org,
	linux-security-module@vger.kernel.org,
	kernel-hardening@lists.openwall.com, corbet@lwn.net,
	mingo@kernel.org, netdev@vger.kernel.org, peterz@infradead.org,
	torvalds@linux-foundation.org
Subject: [kernel-hardening] Re: [PATCH v5 next 1/5] modules:capabilities: add request_module_cap()
Date: Wed, 29 Nov 2017 19:35:31 -0500	[thread overview]
Message-ID: <20171130003531.gwpl22bxmweifjz2@thunk.org> (raw)
In-Reply-To: <20171129172852.GA14545@mail.hallyn.com>

On Wed, Nov 29, 2017 at 11:28:52AM -0600, Serge E. Hallyn wrote:
> 
> Just to be clear, module loading requires - and must always continue to
> require - CAP_SYS_MODULE against the initial user namespace.  Containers
> in user namespaces do not have that.
> 
> I don't believe anyone has ever claimed that containers which are not in
> a user namespace are in any way secure.

Unless the container performs some action which causes the kernel to
call request_module(), which then loads some kernel module,
potentially containing cr*p unmaintained code which was included when
the distro compiled the world, into the host kernel.

This is an attack vector that doesn't exist if you are using VM's.
And in general, the attack surface of the entire Linux
kernel<->userspace API is far larger than that which is exposed by the
guest<->host interface.

For that reason, containers are *far* more insecure than VM's, since
once the attacker gets root on the guest VM, they then have to attack
the hypervisor interface.  And if you compare the attack surface of
the two, it's pretty clear which is larger, and it's not the
hypervisor interface.

						- Ted

  reply	other threads:[~2017-11-30  0:36 UTC|newest]

Thread overview: 266+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-11-27 17:18 [PATCH v5 next 0/5] Improve Module autoloading infrastructure Djalal Harouni
2017-11-27 17:18 ` [kernel-hardening] " Djalal Harouni
2017-11-27 17:18 ` Djalal Harouni
2017-11-27 17:18 ` [PATCH v5 next 1/5] modules:capabilities: add request_module_cap() Djalal Harouni
2017-11-27 17:18   ` [kernel-hardening] " Djalal Harouni
2017-11-27 17:18   ` Djalal Harouni
2017-11-27 18:48   ` Randy Dunlap
2017-11-27 18:48     ` [kernel-hardening] " Randy Dunlap
2017-11-27 18:48     ` Randy Dunlap
2017-11-27 21:35     ` Djalal Harouni
2017-11-27 21:35       ` [kernel-hardening] " Djalal Harouni
2017-11-27 21:35       ` Djalal Harouni
2017-11-28 19:14   ` Luis R. Rodriguez
2017-11-28 19:14     ` [kernel-hardening] " Luis R. Rodriguez
2017-11-28 19:14     ` Luis R. Rodriguez
2017-11-28 20:11     ` Kees Cook
2017-11-28 20:11       ` [kernel-hardening] " Kees Cook
2017-11-28 20:11       ` Kees Cook
2017-11-28 21:16       ` Luis R. Rodriguez
2017-11-28 21:16         ` [kernel-hardening] " Luis R. Rodriguez
2017-11-28 21:16         ` Luis R. Rodriguez
2017-11-28 21:33         ` Djalal Harouni
2017-11-28 21:33           ` [kernel-hardening] " Djalal Harouni
2017-11-28 21:33           ` Djalal Harouni
2017-11-28 22:18           ` Luis R. Rodriguez
2017-11-28 22:18             ` [kernel-hardening] " Luis R. Rodriguez
2017-11-28 22:18             ` Luis R. Rodriguez
2017-11-28 22:52             ` Djalal Harouni
2017-11-28 22:52               ` [kernel-hardening] " Djalal Harouni
2017-11-28 22:52               ` Djalal Harouni
2017-11-28 21:39         ` Kees Cook
2017-11-28 21:39           ` [kernel-hardening] " Kees Cook
2017-11-28 21:39           ` Kees Cook
2017-11-28 22:12           ` Luis R. Rodriguez
2017-11-28 22:12             ` [kernel-hardening] " Luis R. Rodriguez
2017-11-28 22:12             ` Luis R. Rodriguez
2017-11-28 22:18             ` Kees Cook
2017-11-28 22:18               ` [kernel-hardening] " Kees Cook
2017-11-28 22:18               ` Kees Cook
2017-11-28 22:48               ` Luis R. Rodriguez
2017-11-28 22:48                 ` [kernel-hardening] " Luis R. Rodriguez
2017-11-28 22:48                 ` Luis R. Rodriguez
2017-11-29  7:49                 ` Michal Kubecek
2017-11-29  7:49                   ` [kernel-hardening] " Michal Kubecek
2017-11-29  7:49                   ` Michal Kubecek
2017-11-29 13:46           ` Alan Cox
2017-11-29 13:46             ` [kernel-hardening] " Alan Cox
2017-11-29 13:46             ` Alan Cox
2017-11-29 14:50             ` David Miller
2017-11-29 14:50               ` [kernel-hardening] " David Miller
2017-11-29 14:50               ` David Miller
2017-11-29 15:54               ` Theodore Ts'o
2017-11-29 15:54                 ` [kernel-hardening] " Theodore Ts'o
2017-11-29 15:54                 ` Theodore Ts'o
2017-11-29 15:58                 ` David Miller
2017-11-29 15:58                   ` [kernel-hardening] " David Miller
2017-11-29 15:58                   ` David Miller
2017-11-29 16:29                   ` Theodore Ts'o
2017-11-29 16:29                     ` [kernel-hardening] " Theodore Ts'o
2017-11-29 16:29                     ` Theodore Ts'o
2017-11-29 22:45                   ` Linus Torvalds
2017-11-29 22:45                     ` [kernel-hardening] " Linus Torvalds
2017-11-29 22:45                     ` Linus Torvalds
2017-11-29 22:45                     ` Linus Torvalds
2017-11-30  0:06                     ` Kees Cook
2017-11-30  0:06                       ` [kernel-hardening] " Kees Cook
2017-11-30  0:06                       ` Kees Cook
2017-11-30  0:06                       ` Kees Cook
2017-11-29 17:28                 ` Serge E. Hallyn
2017-11-29 17:28                   ` [kernel-hardening] " Serge E. Hallyn
2017-11-29 17:28                   ` Serge E. Hallyn
2017-11-30  0:35                   ` Theodore Ts'o [this message]
2017-11-30  0:35                     ` [kernel-hardening] " Theodore Ts'o
2017-11-30  0:35                     ` Theodore Ts'o
2017-11-30 17:17                     ` Serge E. Hallyn
2017-11-30 17:17                       ` [kernel-hardening] " Serge E. Hallyn
2017-11-30 17:17                       ` Serge E. Hallyn
2017-11-28 20:18     ` Djalal Harouni
2017-11-28 20:18       ` [kernel-hardening] " Djalal Harouni
2017-11-28 20:18       ` Djalal Harouni
2017-11-27 17:18 ` [PATCH v5 next 2/5] modules:capabilities: add cap_kernel_module_request() permission check Djalal Harouni
2017-11-27 17:18   ` [kernel-hardening] " Djalal Harouni
2017-11-27 17:18   ` Djalal Harouni
2017-11-30  2:05   ` Luis R. Rodriguez
2017-11-30  2:05     ` [kernel-hardening] " Luis R. Rodriguez
2017-11-30  2:05     ` Luis R. Rodriguez
2017-11-27 17:18 ` [PATCH v5 next 3/5] modules:capabilities: automatic module loading restriction Djalal Harouni
2017-11-27 17:18   ` [kernel-hardening] " Djalal Harouni
2017-11-27 17:18   ` Djalal Harouni
2017-11-30  1:23   ` Luis R. Rodriguez
2017-11-30  1:23     ` [kernel-hardening] " Luis R. Rodriguez
2017-11-30  1:23     ` Luis R. Rodriguez
2017-11-30 12:22     ` Djalal Harouni
2017-11-30 12:22       ` [kernel-hardening] " Djalal Harouni
2017-11-30 12:22       ` Djalal Harouni
2017-11-27 17:18 ` [PATCH v5 next 4/5] modules:capabilities: add a per-task modules auto-load mode Djalal Harouni
2017-11-27 17:18   ` [kernel-hardening] " Djalal Harouni
2017-11-27 17:18   ` Djalal Harouni
2017-11-27 17:18 ` [PATCH v5 next 5/5] net: modules: use request_module_cap() to load 'netdev-%s' modules Djalal Harouni
2017-11-27 17:18   ` [kernel-hardening] " Djalal Harouni
2017-11-27 17:18   ` Djalal Harouni
2017-11-27 18:44   ` Linus Torvalds
2017-11-27 18:44     ` [kernel-hardening] " Linus Torvalds
2017-11-27 18:44     ` Linus Torvalds
2017-11-27 18:44     ` Linus Torvalds
2017-11-27 21:41     ` Djalal Harouni
2017-11-27 21:41       ` [kernel-hardening] " Djalal Harouni
2017-11-27 21:41       ` Djalal Harouni
2017-11-27 21:41       ` Djalal Harouni
2017-11-27 22:04       ` Linus Torvalds
2017-11-27 22:04         ` [kernel-hardening] " Linus Torvalds
2017-11-27 22:04         ` Linus Torvalds
2017-11-27 22:04         ` Linus Torvalds
2017-11-27 22:59         ` Kees Cook
2017-11-27 22:59           ` [kernel-hardening] " Kees Cook
2017-11-27 22:59           ` Kees Cook
2017-11-27 22:59           ` Kees Cook
2017-11-27 23:14           ` Linus Torvalds
2017-11-27 23:14             ` [kernel-hardening] " Linus Torvalds
2017-11-27 23:14             ` Linus Torvalds
2017-11-27 23:14             ` Linus Torvalds
2017-11-27 23:19             ` Kees Cook
2017-11-27 23:19               ` [kernel-hardening] " Kees Cook
2017-11-27 23:19               ` Kees Cook
2017-11-27 23:19               ` Kees Cook
2017-11-27 23:35               ` Linus Torvalds
2017-11-27 23:35                 ` [kernel-hardening] " Linus Torvalds
2017-11-27 23:35                 ` Linus Torvalds
2017-11-27 23:35                 ` Linus Torvalds
2017-11-28  1:23             ` Kees Cook
2017-11-28  1:23               ` [kernel-hardening] " Kees Cook
2017-11-28  1:23               ` Kees Cook
2017-11-28  1:23               ` Kees Cook
2017-11-28 12:16         ` [kernel-hardening] " Geo Kozey
2017-11-28 12:16           ` Geo Kozey
2017-11-28 12:16           ` Geo Kozey
2017-11-28 19:32           ` Theodore Ts'o
2017-11-28 19:32             ` Theodore Ts'o
2017-11-28 19:32             ` Theodore Ts'o
2017-11-28 20:08             ` Kees Cook
2017-11-28 20:08               ` Kees Cook
2017-11-28 20:08               ` Kees Cook
2017-11-28 20:12               ` Linus Torvalds
2017-11-28 20:12                 ` Linus Torvalds
2017-11-28 20:12                 ` Linus Torvalds
2017-11-28 20:20                 ` Kees Cook
2017-11-28 20:20                   ` Kees Cook
2017-11-28 20:20                   ` Kees Cook
2017-11-28 20:33                   ` Linus Torvalds
2017-11-28 20:33                     ` Linus Torvalds
2017-11-28 20:33                     ` Linus Torvalds
2017-11-28 21:10                     ` Djalal Harouni
2017-11-28 21:10                       ` Djalal Harouni
2017-11-28 21:10                       ` Djalal Harouni
2017-11-28 21:33                     ` Kees Cook
2017-11-28 21:33                       ` Kees Cook
2017-11-28 21:33                       ` Kees Cook
2017-11-28 23:23                       ` Theodore Ts'o
2017-11-28 23:23                         ` Theodore Ts'o
2017-11-28 23:23                         ` Theodore Ts'o
2017-11-28 23:29                         ` Kees Cook
2017-11-28 23:29                           ` Kees Cook
2017-11-28 23:29                           ` Kees Cook
2017-11-28 23:49                           ` Theodore Ts'o
2017-11-28 23:49                             ` Theodore Ts'o
2017-11-28 23:49                             ` Theodore Ts'o
2017-11-29  0:18                             ` Kees Cook
2017-11-29  0:18                               ` Kees Cook
2017-11-29  0:18                               ` Kees Cook
2017-11-29  6:36                               ` Theodore Ts'o
2017-11-29  6:36                                 ` Theodore Ts'o
2017-11-29  6:36                                 ` Theodore Ts'o
2017-11-29 14:46                             ` Geo Kozey
2017-11-29 14:46                               ` Geo Kozey
2017-11-29 14:46                               ` Geo Kozey
2017-12-01 15:22                             ` Marcus Meissner
2017-12-01 15:22                               ` Marcus Meissner
2017-12-01 15:22                               ` Marcus Meissner
2017-11-28 23:53                         ` Djalal Harouni
2017-11-28 23:53                           ` Djalal Harouni
2017-11-28 23:53                           ` Djalal Harouni
2017-11-28 21:51                     ` Geo Kozey
2017-11-28 21:51                       ` Geo Kozey
2017-11-28 21:51                       ` Geo Kozey
2017-11-28 23:51                       ` Linus Torvalds
2017-11-28 23:51                         ` Linus Torvalds
2017-11-28 23:51                         ` Linus Torvalds
2017-11-29  0:17                         ` Linus Torvalds
2017-11-29  0:17                           ` Linus Torvalds
2017-11-29  0:17                           ` Linus Torvalds
2017-11-29  0:26                           ` Kees Cook
2017-11-29  0:26                             ` Kees Cook
2017-11-29  0:26                             ` Kees Cook
2017-11-29  0:50                             ` Linus Torvalds
2017-11-29  0:50                               ` Linus Torvalds
2017-11-29  0:50                               ` Linus Torvalds
2017-11-29  4:26                               ` Eric W. Biederman
2017-11-29  4:26                                 ` Eric W. Biederman
2017-11-29  4:26                                 ` Eric W. Biederman
2017-11-29 18:30                               ` Kees Cook
2017-11-29 18:30                                 ` Kees Cook
2017-11-29 18:30                                 ` Kees Cook
2017-11-29 18:46                                 ` Linus Torvalds
2017-11-29 18:46                                   ` Linus Torvalds
2017-11-29 18:46                                   ` Linus Torvalds
2017-11-29 18:53                                   ` Linus Torvalds
2017-11-29 18:53                                     ` Linus Torvalds
2017-11-29 18:53                                     ` Linus Torvalds
2017-11-29 21:17                                   ` Kees Cook
2017-11-29 21:17                                     ` Kees Cook
2017-11-29 21:17                                     ` Kees Cook
2017-11-29 22:14                                     ` Linus Torvalds
2017-11-29 22:14                                       ` Linus Torvalds
2017-11-29 22:14                                       ` Linus Torvalds
2017-11-30  0:44                                       ` Kees Cook
2017-11-30  0:44                                         ` Kees Cook
2017-11-30  0:44                                         ` Kees Cook
2017-11-30  2:08                                         ` Linus Torvalds
2017-11-30  2:08                                           ` Linus Torvalds
2017-11-30  2:08                                           ` Linus Torvalds
2017-11-30  6:51                                       ` Daniel Micay
2017-11-30  6:51                                         ` Daniel Micay
2017-11-30  6:51                                         ` Daniel Micay
2017-11-30  8:50                                         ` Djalal Harouni
2017-11-30  8:50                                           ` Djalal Harouni
2017-11-30  8:50                                           ` Djalal Harouni
2017-11-30 14:16                                           ` Theodore Ts'o
2017-11-30 14:16                                             ` Theodore Ts'o
2017-11-30 14:16                                             ` Theodore Ts'o
2017-11-30 14:51                                             ` Djalal Harouni
2017-11-30 14:51                                               ` Djalal Harouni
2017-11-30 14:51                                               ` Djalal Harouni
2017-12-01  6:39                                           ` Daniel Micay
2017-12-01  6:39                                             ` Daniel Micay
2017-12-01  6:39                                             ` Daniel Micay
2017-11-29 15:28                           ` Geo Kozey
2017-11-29 15:28                             ` Geo Kozey
2017-11-29 15:28                             ` Geo Kozey
2017-11-27 18:41 ` [PATCH v5 next 0/5] Improve Module autoloading infrastructure Linus Torvalds
2017-11-27 18:41   ` [kernel-hardening] " Linus Torvalds
2017-11-27 18:41   ` Linus Torvalds
2017-11-27 18:41   ` Linus Torvalds
2017-11-27 19:02   ` Linus Torvalds
2017-11-27 19:02     ` [kernel-hardening] " Linus Torvalds
2017-11-27 19:02     ` Linus Torvalds
2017-11-27 19:02     ` Linus Torvalds
2017-11-27 19:12     ` Linus Torvalds
2017-11-27 19:12       ` [kernel-hardening] " Linus Torvalds
2017-11-27 19:12       ` Linus Torvalds
2017-11-27 19:12       ` Linus Torvalds
2017-11-27 21:31       ` Djalal Harouni
2017-11-27 21:31         ` [kernel-hardening] " Djalal Harouni
2017-11-27 21:31         ` Djalal Harouni
2017-11-27 21:31         ` Djalal Harouni
2017-11-27 19:14   ` David Miller
2017-11-27 19:14     ` [kernel-hardening] " David Miller
2017-11-27 19:14     ` David Miller
2017-11-27 22:31     ` James Morris
2017-11-27 22:31       ` [kernel-hardening] " James Morris
2017-11-27 22:31       ` James Morris
2017-11-27 23:04       ` Kees Cook
2017-11-27 23:04         ` [kernel-hardening] " Kees Cook
2017-11-27 23:04         ` Kees Cook
2017-11-27 23:44         ` James Morris
2017-11-27 23:44           ` [kernel-hardening] " James Morris
2017-11-27 23:44           ` James Morris

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=20171130003531.gwpl22bxmweifjz2@thunk.org \
    --to=tytso@mit.edu \
    --cc=akpm@linux-foundation.org \
    --cc=ben.hutchings@codethink.co.uk \
    --cc=corbet@lwn.net \
    --cc=davem@davemloft.net \
    --cc=gnomes@lxorguk.ukuu.org.uk \
    --cc=james.l.morris@oracle.com \
    --cc=jeyu@kernel.org \
    --cc=keescook@chromium.org \
    --cc=kernel-hardening@lists.openwall.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-security-module@vger.kernel.org \
    --cc=luto@kernel.org \
    --cc=mcgrof@kernel.org \
    --cc=mingo@kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=peterz@infradead.org \
    --cc=rusty@rustcorp.com.au \
    --cc=serge@hallyn.com \
    --cc=solar@openwall.com \
    --cc=tixxdz@gmail.com \
    --cc=torvalds@linux-foundation.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.