linux-kernel - Re: Does CONFIG_HARDENED

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 22 Nov 2017 09:56:30 -0800
From:   Kees Cook <keescook@...omium.org>
To:     Michael Holzheu <holzheu@...ux.vnet.ibm.com>
Cc:     Tycho Andersen <tycho@...ker.com>, Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Heiko Carstens <heiko.carstens@...ibm.com>,
        Martin Schwidefsky <schwidefsky@...ibm.com>,
        Vasily Gorbik <gor@...ux.vnet.ibm.com>
Subject: Re: Does CONFIG_HARDENED_USERCOPY break /dev/mem?

On Wed, Nov 22, 2017 at 9:43 AM, Kees Cook <keescook@...omium.org> wrote:
> On Wed, Nov 22, 2017 at 1:28 AM, Michael Holzheu
> <holzheu@...ux.vnet.ibm.com> wrote:
>> So what's your plan now? How will you fix this issue?
>
> I think the best plan here would be to use the Kconfig "imply
> STRICT_DEVMEM" in HARDENED_USERCOPY. That would make STRICT_DEVMEM
> enabled by default but still configurable. Then the kernel-text check
> in hardened usercopy could be skip when !STRICT_DEVMEM.

Hmm, this doesn't work the way I was expecting...

diff --git a/security/Kconfig b/security/Kconfig
index e8e449444e65..3b4effd8bbc2 100644
--- a/security/Kconfig
+++ b/security/Kconfig
@@ -143,6 +143,7 @@ config HARDENED_USERCOPY
        bool "Harden memory copies between kernel and userspace"
        depends on HAVE_HARDENED_USERCOPY_ALLOCATOR
        select BUG
+       imply STRICT_DEVMEM
        help
          This option checks for obviously wrong memory regions when
          copying memory to/from the kernel (via copy_to_user() and

$ make defconfig
$ egrep '(DEVMEM|HARDENED_USERCOPY)\b' .config
CONFIG_DEVMEM=y
# CONFIG_STRICT_DEVMEM is not set
# CONFIG_HARDENED_USERCOPY is not set
$ ./scripts/config -e CONFIG_HARDENED_USERCOPY
$ egrep '(DEVMEM|HARDENED_USERCOPY)\b' .config
CONFIG_DEVMEM=y
# CONFIG_STRICT_DEVMEM is not set
CONFIG_HARDENED_USERCOPY=y

Maybe make STRICT_DEVMEM's default tie to HARDENED_USERCOPY? Bleh,
this doesn't seem to work either:

diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index 947d3e2ed5c2..8cc05033ba65 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -1985,7 +1985,7 @@ config STRICT_DEVMEM
        bool "Filter access to /dev/mem"
        depends on MMU && DEVMEM
        depends on ARCH_HAS_DEVMEM_IS_ALLOWED
-       default y if TILE || PPC
+       default y if TILE || PPC || HARDENED_USERCOPY
        ---help---
          If this option is disabled, you allow userspace (root) access to all
          of memory, including kernel and userspace memory. Accidental

I'm open to suggestions...

-Kees

-- 
Kees Cook
Pixel Security