Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 14 Jun 2019 18:05:19 +0200
From: Jann Horn <jannh@...gle.com>
To: "Denis 'GNUtoo' Carikli" <GNUtoo@...erdimension.org>
Cc: Kees Cook <keescook@...omium.org>, Emese Revfy <re.emese@...il.com>, 
	Paul Kocialkowski <paul.kocialkowski@...tlin.com>, 
	Kernel Hardening <kernel-hardening@...ts.openwall.com>, linux-arm-kernel@...ts.infradead.org, 
	Russell King <linux@...linux.org.uk>
Subject: Re: [PATCH] security: do not enable CONFIG_GCC_PLUGINS by default

+32-bit ARM folks

On Fri, Jun 14, 2019 at 5:10 PM Denis 'GNUtoo' Carikli
<GNUtoo@...erdimension.org> wrote:
> On a Galaxy SIII (I9300), the patch mentioned below broke boot:
> - The display still had the bootloader logo, while with this
>   patch, the 4 Tux logo appears.
> - No print appeared on the serial port anymore after the kernel
>   was loaded, whereas with this patch, we have the serial
>   console working, and the device booting.
>
> Booting was broken by the following commit:
>   9f671e58159a ("security: Create "kernel hardening" config area")
>
> As the bootloader of this device enables the MMU, I had the following
> patch applied during the tests:
>   Author: Arve Hjønnevåg <arve@...roid.com>
>   Date:   Fri Nov 30 17:05:40 2012 -0800
>
>       ANDROID: arm: decompressor: Flush tlb before swiching domain 0 to client mode
>
>       If the bootloader used a page table that is incompatible with domain 0
>       in client mode, and boots with the mmu on, then swithing domain 0 to
>       client mode causes a fault if we don't flush the tlb after updating
>       the page table pointer.
>
>       v2: Add ISB before loading dacr.
>
>   diff --git a/arch/arm/boot/compressed/head.S b/arch/arm/boot/compressed/head.S
>   index 7135820f76d4..6e87ceda3b29 100644
>   --- a/arch/arm/boot/compressed/head.S
>   +++ b/arch/arm/boot/compressed/head.S
>   @@ -837,6 +837,8 @@ __armv7_mmu_cache_on:
>                   bic     r6, r6, #1 << 31        @ 32-bit translation system
>                   bic     r6, r6, #(7 << 0) | (1 << 4)    @ use only ttbr0
>                   mcrne   p15, 0, r3, c2, c0, 0   @ load page table pointer
>   +               mcrne   p15, 0, r0, c8, c7, 0   @ flush I,D TLBs
>   +               mcr     p15, 0, r0, c7, c5, 4   @ ISB
>                   mcrne   p15, 0, r1, c3, c0, 0   @ load domain access control
>                   mcrne   p15, 0, r6, c2, c0, 2   @ load ttb control
>    #endif
>
> Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@...erdimension.org>
> ---
>  scripts/gcc-plugins/Kconfig | 1 -
>  1 file changed, 1 deletion(-)
>
> diff --git a/scripts/gcc-plugins/Kconfig b/scripts/gcc-plugins/Kconfig
> index e9c677a53c74..afa1db3d3471 100644
> --- a/scripts/gcc-plugins/Kconfig
> +++ b/scripts/gcc-plugins/Kconfig
> @@ -18,7 +18,6 @@ config GCC_PLUGINS
>         bool
>         depends on HAVE_GCC_PLUGINS
>         depends on PLUGIN_HOSTCC != ""
> -       default y
>         help
>           GCC plugins are loadable modules that provide extra features to the
>           compiler. They are useful for runtime instrumentation and static analysis.

I don't think GCC_PLUGINS alone is supposed to generate any code? It
just makes it possible to enable a bunch of other kconfig flags that
can generate code.

STACKPROTECTOR_PER_TASK defaults to y and depends on GCC_PLUGINS, so
is that perhaps what broke? Can you try whether disabling just that
works for you?

My guess is that maybe there is some early boot code that needs to
have the stack protector disabled, or something like that.

Powered by blists - more mailing lists

Your e-mail address:

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.