[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 4 Dec 2018 16:04:33 +0100
From: Ard Biesheuvel <ard.biesheuvel@...aro.org>
To: Nicolas Pitre <nicolas.pitre@...aro.org>
Cc: Kees Cook <keescook@...omium.org>,
linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>, Russell King <linux@...linux.org.uk>,
Emese Revfy <re.emese@...il.com>, Arnd Bergmann <arnd@...db.de>, Laura Abbott <labbott@...hat.com>,
Kernel Hardening <kernel-hardening@...ts.openwall.com>
Subject: Re: [RFC/RFT PATCH v2] ARM: smp: add support for per-task stack canaries
On Fri, 30 Nov 2018 at 17:37, Nicolas Pitre <nicolas.pitre@...aro.org> wrote:
>
> On Fri, 30 Nov 2018, Ard Biesheuvel wrote:
>
> > On Thu, 29 Nov 2018 at 23:58, Nicolas Pitre <nicolas.pitre@...aro.org> wrote:
> > >
> > > On Tue, 27 Nov 2018, Ard Biesheuvel wrote:
> > >
> > > > On ARM, we currently only change the value of the stack canary when
> > > > switching tasks if the kernel was built for UP. On SMP kernels, this
> > > > is impossible since the stack canary value is obtained via a global
> > > > symbol reference, which means
> > > > a) all running tasks on all CPUs must use the same value
> > > > b) we can only modify the value when no kernel stack frames are live
> > > > on any CPU, which is effectively never.
> > > >
> > > > So instead, use a GCC plugin to add a RTL pass that replaces each
> > > > reference to the address of the __stack_chk_guard symbol with an
> > > > expression that produces the address of the 'stack_canary' field
> > > > that is added to struct thread_info. This way, each task will use
> > > > its own randomized value.
> > > >
> > > > Cc: Russell King <linux@...linux.org.uk>
> > > > Cc: Kees Cook <keescook@...omium.org>
> > > > Cc: Emese Revfy <re.emese@...il.com>
> > > > Cc: Arnd Bergmann <arnd@...db.de>
> > > > Cc: Nicolas Pitre <nicolas.pitre@...aro.org>
> > > > Cc: Laura Abbott <labbott@...hat.com>
> > > > Cc: kernel-hardening@...ts.openwall.com
> > > > Signed-off-by: Ard Biesheuvel <ard.biesheuvel@...aro.org>
> > > > ---
> > > > v2:
> > > > - adopt powerpc Makefile approach to pass stack canary offset and thread
> > > > size to the plugin as command line arguments
> > > > - depend on !XIP_DEFLATED_DATA
> > > >
> > > > arch/arm/Kconfig | 16 +++
> > > > arch/arm/Makefile | 11 +++
> > > > arch/arm/boot/compressed/Makefile | 1 +
> > > > arch/arm/include/asm/stackprotector.h | 12 ++-
> > > > arch/arm/include/asm/thread_info.h | 3 +
> > > > arch/arm/kernel/asm-offsets.c | 1 +
> > > > arch/arm/kernel/process.c | 6 +-
> > > > scripts/Makefile.gcc-plugins | 6 ++
> > > > scripts/gcc-plugins/Kconfig | 4 +
> > > > scripts/gcc-plugins/arm_ssp_per_task_plugin.c | 103 ++++++++++++++++++++
> > > > 10 files changed, 160 insertions(+), 3 deletions(-)
> > > >
> > > > diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
> > > > index 91be74d8df65..6142e83b45da 100644
> > > > --- a/arch/arm/Kconfig
> > > > +++ b/arch/arm/Kconfig
> > > > @@ -1810,6 +1810,22 @@ config XEN
> > > > help
> > > > Say Y if you want to run Linux in a Virtual Machine on Xen on ARM.
> > > >
> > > > +config STACKPROTECTOR_PER_TASK
> > > > + bool "Use a unique stack canary value for each task"
> > > > + depends on GCC_PLUGINS && STACKPROTECTOR && SMP && !XIP_DEFLATED_DATA
> > > > + select GCC_PLUGIN_ARM_SSP_PER_TASK
> > > > + help
> > > > + Due to the fact that GCC uses an ordinary symbol reference from
> > > > + which to load the value of the stack canary, this value can only
> > > > + change at reboot time on SMP systems, and all tasks running in the
> > > > + kernel's address space are forced to use the same canary value for
> > > > + the entire duration that the system is up.
> > > > +
> > > > + Enable this option to switch to a different method that uses a
> > > > + different canary value for each task.
> > > > +
> > > > + This is not enabled by default since it relies on a GCC plugin.
> > >
> > > Why wouldn't you default it to y and let the dependencies disable it
> > > when not met?
> > >
> > > Other than that:
> > >
> > > Acked-by: Nicolas Pitre <nico@...aro.org>
> > >
> >
> > Thanks Nico.
> >
> > I wasn't sure about enabling it by default, though. Perhaps Kees has
> > any insights?
>
> I'd say that if you have CONFIG_GCC_PLUGINS=y and
> CONFIG_STACKPROTECTOR=y then it is pretty clear what the user expects
> already. The fact that CONFIG_STACKPROTECTOR requires a twist on SMP is
> an implementation detail not a feature.
>
Fair point.
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
