Re: [RESEND PATCH 0/7] Introduce bus domains controller framework
From: Sudeep Holla
Date: Tue Apr 23 2019 - 09:55:59 EST
On Tue, Apr 23, 2019 at 01:33:19PM +0000, Benjamin GAIGNARD wrote:
>
> On 4/23/19 3:21 PM, Sudeep Holla wrote:
> > On Mon, Mar 18, 2019 at 12:05:54PM +0100, Benjamin Gaignard wrote:
> >> Le lun. 18 mars 2019 Ã 11:43, Sudeep Holla <sudeep.holla@xxxxxxx> a Ãcrit :
> >>> On Mon, Mar 18, 2019 at 11:05:58AM +0100, Benjamin Gaignard wrote:
> >>>> Bus domains controllers allow to divided system on chip into multiple domains
> >>>> that can be used to select by who hardware blocks could be accessed.
> >>>> A domain could be a cluster of CPUs (or coprocessors), a range of addresses or
> >>>> a group of hardware blocks.
> >>>>
> >>>> Framework architecture is inspirated by pinctrl framework:
> >>>> - a default configuration could be applied before bind the driver
> >>>> - configurations could be apllied dynamically by drivers
> >>>> - device node provides the bus domains configurations
> >>>>
> >>>> An example of bus domains controller is STM32 ETZPC hardware block
> >>>> which got 3 domains:
> >>>> - secure: hardware blocks are only accessible by software running on trust
> >>>> zone.
> >>>> - non-secure: hardware blocks are accessible by non-secure software (i.e.
> >>>> linux kernel).
> >>>> - coprocessor: hardware blocks are only accessible by the corpocessor.
> >>>> Up to 94 hardware blocks of the soc could be managed by ETZPC and
> >>>> assigned to one of the three domains.
> >>>>
> >>> You fail to explain why do we need this in non-secure Linux ?
> >>> You need to have solid reasons as why this can't be done in secure
> >>> firmware. And yes I mean even on arm32. On platforms with such hardware
> >>> capabilities you will need some secure firmware to be running and these
> >>> things can be done there. I don't want this enabled for ARM64 at all,
> >>> firmware *has to deal* with this.
> >> We use ETZPC to define if hardware blocks can be used by Cortex A7 or Cortex
> >> M4 (both non-secure) on STM32MP1 SoC, this new framework allow to change
> >> hardware block split at runtime. This could be done even on non-secure world
> >> because their is nothing critical to change hardware blocks users.
> > OK, that's interesting, assuming Cortex M4 execution as non-secure. I would
> > expect otherwise. Even if it's configurable, I would see that happen in
> > secure entity via OPTEE or something similar from non-secure side.
> Your assumption is correct Cortex M4 execution is non-secure.
Sorry if I was not clear. I told Cortex M4 non-secure execution is interesting
as I expected it to be secure.
> >
> > Do you have any documents that I can refer to get the overall security
> > design for such platforms ?
>
> SoC datasheet is here:
>
> https://www.st.com/resource/en/datasheet/stm32mp157a.pdf
>
> with just few words about ETZPC:
>
> 3.14 TrustZone protection controller (ETZPC)
> ETZPC is used to configure TrustZone security of bus masters and slaves with
> programmable-security attributes (securable resources) such as:
> â On-chip SYSRAM with programmable secure region size
> â AHB and APB peripherals to be made secure
> Notice that by default, SYSRAM and peripheral are set to secure access
> only, so, not accessible by non-secure masters such as Cortex-M4 or DMA1/DMA2.
> ETZPC can also allocate peripherals and SRAM to be accessible only by
> the Cortex-M4 and/or DMA1/DMA2. This ensures the safe execution of the
> Cortex-M4 firmware, protected from other masters (e.g. Cortex-A7) unwanted
> accesses.
>
The above statement makes me wonder if Cortex-M4 firmware is really
non-secure, if so why does it need such an isolation from other masters
like Cortex-A7. For me Cortex-M4 is secure and Cortex-A7 can execute
in non-secure hence Cortex-M4 needs to be isolated from Cortex-A7 as
mentioned in the above excerpts from the datasheet.
--
Regards,
Sudeep