Hi Tamas,
> I do not know whether the this two phase setting of MSP_LIMIT is still in use or not. If not the no need for S_MSP_STACK_SIZE_INIT.
The only place where it's used is the GCC linker file:
__msp_init_stack_size__ = S_MSP_STACK_SIZE_INIT
So for GCC __msp_init_stack_size__ is in 2 times less (0x400) than for Keil and IAR (0x800) "__msp_init_stack_size__ = S_MSP_STACK_SIZE".
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of Tamas Ban via TF-M
Sent: Tuesday, June 16, 2020 10:17 AM
To: tf-m(a)lists.trustedfirmware.org
Cc: nd <nd(a)arm.com>
Subject: Re: [TF-M] S_MSP_STACK_SIZE_INIT vs S_MSP_STACK_SIZE
Hi Andrej,
The BOOT_TFM_SHARED_DATA and the MSP_STACK area are overlapping on purpose. The partitions which are expecting to receive data from bootloader is intended to get their own data during the partition init phase with calling tfm_core_get_boot_data() with the partition's major_type. Then the data will be copied from shared area to partition's memory.
So after all partition's init is executed then the data from the shared buffer is distributed to the owning partitions and resides in their memory. At this point the shared_data area can be overwritten by growing MSP, without destroying shared data.
Originally there was an S_MSP_STACK_SIZE_INIT size which was used to setup the MSP_LIMIT for the init phase to avoid overwriting the shared data area. After the init phase the MSP_LIMIT was set again with its full size S_MSP_STACK_SIZE.
I do not know whether the this two phase setting of MSP_LIMIT is still in use or not. If not the no need for S_MSP_STACK_SIZE_INIT.
+-> +-> +--------------+ <- Shared boot data base, S_MSP_STACK_SIZE
| Shared| | |
M | Data | | |
S | | | |
P | +-> +--------------+ <- S_MSP_STACK_SIZE_INIT
| | |
| | |
| | |
+-> +--------------+ <- Top of MSP
Tamas
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org<mailto:tf-m-bounces@lists.trustedfirmware.org>> On Behalf Of Andrej Butok via TF-M
Sent: 16 June 2020 09:52
To: tf-m(a)lists.trustedfirmware.org<mailto:tf-m@lists.trustedfirmware.org>
Subject: [TF-M] S_MSP_STACK_SIZE_INIT vs S_MSP_STACK_SIZE
Hello,
What is a difference between S_MSP_STACK_SIZE_INIT and S_MSP_STACK_SIZE defined in partition\region_defs.h:
#define S_MSP_STACK_SIZE_INIT (0x0000400)
#define S_MSP_STACK_SIZE (0x0000800)
S_MSP_STACK_SIZE_INIT is used only for gcc.
S_MSP_STACK_SIZE is used for armclang and iar.
Guess, it should be used only one definition. So in our platform code we are going to use only S_MSP_STACK_SIZE.
Should you fix it for all platforms in the original TFM?
Thanks,
Andrej Butok
SW Tech Lead
Security & Connectivity, Microcontrollers
NXP Semiconductors
Hi Andrej,
The BOOT_TFM_SHARED_DATA and the MSP_STACK area are overlapping on purpose. The partitions which are expecting to receive data from bootloader is intended to get their own data during the partition init phase with calling tfm_core_get_boot_data() with the partition's major_type. Then the data will be copied from shared area to partition's memory.
So after all partition's init is executed then the data from the shared buffer is distributed to the owning partitions and resides in their memory. At this point the shared_data area can be overwritten by growing MSP, without destroying shared data.
Originally there was an S_MSP_STACK_SIZE_INIT size which was used to setup the MSP_LIMIT for the init phase to avoid overwriting the shared data area. After the init phase the MSP_LIMIT was set again with its full size S_MSP_STACK_SIZE.
I do not know whether the this two phase setting of MSP_LIMIT is still in use or not. If not the no need for S_MSP_STACK_SIZE_INIT.
+-> +-> +--------------+ <- Shared boot data base, S_MSP_STACK_SIZE
| Shared| | |
M | Data | | |
S | | | |
P | +-> +--------------+ <- S_MSP_STACK_SIZE_INIT
| | |
| | |
| | |
+-> +--------------+ <- Top of MSP
Tamas
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of Andrej Butok via TF-M
Sent: 16 June 2020 09:52
To: tf-m(a)lists.trustedfirmware.org
Subject: [TF-M] S_MSP_STACK_SIZE_INIT vs S_MSP_STACK_SIZE
Hello,
What is a difference between S_MSP_STACK_SIZE_INIT and S_MSP_STACK_SIZE defined in partition\region_defs.h:
#define S_MSP_STACK_SIZE_INIT (0x0000400)
#define S_MSP_STACK_SIZE (0x0000800)
S_MSP_STACK_SIZE_INIT is used only for gcc.
S_MSP_STACK_SIZE is used for armclang and iar.
Guess, it should be used only one definition. So in our platform code we are going to use only S_MSP_STACK_SIZE.
Should you fix it for all platforms in the original TFM?
Thanks,
Andrej Butok
SW Tech Lead
Security & Connectivity, Microcontrollers
NXP Semiconductors
Hello,
What is a difference between S_MSP_STACK_SIZE_INIT and S_MSP_STACK_SIZE defined in partition\region_defs.h:
#define S_MSP_STACK_SIZE_INIT (0x0000400)
#define S_MSP_STACK_SIZE (0x0000800)
S_MSP_STACK_SIZE_INIT is used only for gcc.
S_MSP_STACK_SIZE is used for armclang and iar.
Guess, it should be used only one definition. So in our platform code we are going to use only S_MSP_STACK_SIZE.
Should you fix it for all platforms in the original TFM?
Thanks,
Andrej Butok
SW Tech Lead
Security & Connectivity, Microcontrollers
NXP Semiconductors
Hi,
This design addressing to share the code of the common crypto primitives(SHA256, RSA, later AES) between MCUboot and runtime SPE. The goal is to reduce the flash footprint of SPE.
Design proposal:
https://review.trustedfirmware.org/c/TF-M/trusted-firmware-m/+/4587
Implementation:
https://review.trustedfirmware.org/c/TF-M/trusted-firmware-m/+/4583/
TL;DR:
SPE binary size can be reduced with 10KB. If image encryption will be turned on then ~12-13KB is the gain. Please let us know if you think this improvements would be useful.
@Thomas Törnblom<mailto:thomas.tornblom@iar.com>:
Could you check the porting to IAR toolchain?
BR,
Tamas Ban
Hi Thomas,
My bad, I merged a patch today that broke it. The fix is here https://review.trustedfirmware.org/c/TF-M/trusted-firmware-m/+/4562. I will merge it once the CI has verified it.
Kind regards,
Jamie
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of Thomas Törnblom via TF-M
Sent: 12 June 2020 14:00
To: tf-m(a)lists.trustedfirmware.org
Subject: [TF-M] Does FVP_SSE300_MPS2 target build?
I just noticed the merge of the FVP_SSE300_MPS2 support and I'm trying to build, but it fails with:
---
[ 23%] Building C object test/secure_test/CMakeFiles/tfm_secure_tests.dir/suites/ps/secure/psa_ps_s_interface_testsuite.o
C:/Users/thomasto/Projects/tf-m1/trusted-firmware-m/test/suites/ps/secure/psa_ps_s_interface_testsuite.c:46:39: error:
use of undeclared identifier 'PS_MAX_ASSET_SIZE'
static const uint8_t write_asset_data[PS_MAX_ASSET_SIZE] = {0xBF};
...
---
Known issue?
Thomas
--
Thomas Törnblom, Product Engineer
IAR Systems AB
Box 23051, Strandbodgatan 1
SE-750 23 Uppsala, SWEDEN
Mobile: +46 76 180 17 80 Fax: +46 18 16 78 01
E-mail: thomas.tornblom(a)iar.com<mailto:thomas.tornblom@iar.com> Website: www.iar.com<http://www.iar.com>
Twitter: www.twitter.com/iarsystems<http://www.twitter.com/iarsystems>
I just noticed the merge of the FVP_SSE300_MPS2 support and I'm trying
to build, but it fails with:
---
[ 23%] Building C object
test/secure_test/CMakeFiles/tfm_secure_tests.dir/suites/ps/secure/psa_ps_s_interface_testsuite.o
C:/Users/thomasto/Projects/tf-m1/trusted-firmware-m/test/suites/ps/secure/psa_ps_s_interface_testsuite.c:46:39:
error:
use of undeclared identifier 'PS_MAX_ASSET_SIZE'
static const uint8_t write_asset_data[PS_MAX_ASSET_SIZE] = {0xBF};
...
---
Known issue?
Thomas
--
*Thomas Törnblom*, /Product Engineer/
IAR Systems AB
Box 23051, Strandbodgatan 1
SE-750 23 Uppsala, SWEDEN
Mobile: +46 76 180 17 80 Fax: +46 18 16 78 01
E-mail: thomas.tornblom(a)iar.com <mailto:thomas.tornblom@iar.com>
Website: www.iar.com <http://www.iar.com>
Twitter: www.twitter.com/iarsystems <http://www.twitter.com/iarsystems>
Zephyr uses DT as a means to describe the hardware to the system for things like register address, irq numbers, etc. Instead of the typical runtime usage, it generates code to be used at build time for Cortex-M (and other) devices to address the size cost concern.
On the KConfig side Zephyr uses a python implementation of the KConfig parser to allow portability across host build systems.
- k
> On Jun 12, 2020, at 4:35 AM, Gyorgy Szing via TF-M <tf-m(a)lists.trustedfirmware.org> wrote:
>
> Hi,
>
> Just some minor additions. In the TF-M build system Kconfig will most likely generate input for cmake scripts, so the compiler may not have a connection to it.
>
> The original intention with DT is capturing configuration data to allow hardware independent binaries (for defining base address of peripherals, boot information, etc..), and is mostly parsed runtime. In the Cortex-M space this might be a bit out-of scope. Most likely it will be used as a “standard” language to define “retargeting” information, and a tool will be developed which runs compile time and which allows access to the retargeting data for cmake. Something like CMSIS zone, but closer to established open-source technologies.
> In multicore systems having bot Cortex-A and M core DT would allow having a consistent solution for both A and M firmware. This is also a reason why cmake is a better build solution for TF-M than CMSIS Build.
>
> /George
>
> From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of David Hu via TF-M
> Sent: 12 June 2020 11:16
> To: Thomas Törnblom <thomas.tornblom(a)iar.com>; tf-m(a)lists.trustedfirmware.org
> Cc: nd <nd(a)arm.com>
> Subject: Re: [TF-M] CMake flags
>
> Hi Thomas,
>
> DT is another very attractive topic. TF-M currently doesn’t include any DT support.
> So maybe there is mixed preprocessing of DT and Kconfig in Zephyr, but Kconfig can be parsed without any compiler.
>
> Best regards,
> Hu Ziji
>
> From: Thomas Törnblom <thomas.tornblom(a)iar.com>
> Sent: Friday, June 12, 2020 5:09 PM
> To: David Hu <David.Hu(a)arm.com>; tf-m(a)lists.trustedfirmware.org
> Cc: nd <nd(a)arm.com>
> Subject: Re: [TF-M] CMake flags
>
> Perhaps I'm mixing dts and Kconfig up.
>
> In Zephyr, dtc is used to generate output that is run through the gcc preprocessor in assembly mode.
>
> The output from dtc looks like:
> ---
> /dts-v1/;
>
> / {
> ��� #address-cells = < 0x1 >;
> ��� #size-cells = < 0x1 >;
> ��� model = "STMicroelectronics STM32F746G DISCOVERY board";
> ��� compatible = "st,stm32f746g-disco", "st,stm32f746";
> ��� chosen {
> ��� ��� zephyr,console = &usart1;
> ��� ��� zephyr,shell-uart = &usart1;
> ��� ��� zephyr,sram = &sram0;
> ��� ��� zephyr,flash = &flash0;
> ��� ��� zephyr,dtcm = &dtcm;
> ��� };
> ...
> ---
>
> Our preprocessor can not be run in assembler mode and it pukes on the "#address-cells" type elements in the file.
>
> If this is not what Kconfig is about, then please ignore my comments.
>
> Thomas
>
> Den 2020-06-12 kl. 09:43, skrev David Hu:
> Hi Thomas,
> �
> I�ve not met such a problem yet. Could you please give more details about the portability issues in Kconfig?
> In our PoC, it firstly parses the Kconfig files and then convert the output Kconfig options into CMake flags. Thus the whole process is not related to any compiler.
> I checked Zephyr previously. It looks like Zephyr has the similar idea as our PoC does.
> �
> Best regards,
> Hu Ziji
> �
> From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of Thomas T�rnblom via TF-M
> Sent: Friday, June 12, 2020 3:36 PM
> To: tf-m(a)lists.trustedfirmware.org
> Subject: Re: [TF-M] CMake flags
> �
> Introducing a dependency on Kconfig may cause portability issues with some toolchains (IAR for instance)
>
> I am working on adding support for the IAR toolchain to Zephyr, and Kconfig is one of the issues I've run into.
>
> At least in Zephyr, the output from Kconfig is run through the C preprocessor and this causes issues with the IAR preprocessor so for now I use gcc for this step.
>
> Thomas
>
> Den 2020-06-12 kl. 07:32, skrev David Hu via TF-M:
> Hi Adrian,
> �
> Actually we have been investigating configuration alternatives and making some prototypes of Kconfig.
> �
> Kconfig can bring some advantages:
> • It can setup configuration and check dependency before CMake starts and therefore it can help CMake simplify dealing with the mutual dependencies. For example, if CPU and arch options can be configured in Kconfig, CMake can avoid the workaround to re-configure compiler after parsing platform files.
> • Kconfig can collect together the options which are distributed in various CMake files. Developers can more easily understand the options and their dependencies, compared with searching them all around the build system.
> �
> But Kconfig may not cover all the flags in command line. Flags such as compiler selections, build type and project config file, may vary in each build. It can be inconvenient to put them into Kconfig.
> �
> Best regards,
> Hu Ziji
> �
> From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of Adrian Shaw via TF-M
> Sent: Friday, June 12, 2020 12:24 AM
> To: Raef Coles <Raef.Coles(a)arm.com>
> Cc: tf-m(a)lists.trustedfirmware.org
> Subject: [TF-M] CMake flags
> �
> Hi Raef. Thanks for the presentation today.
> �
> If I remember correctly, TF-M's CMake has quite a lot of command line flags now (the ones that are set with -D). And I think this will grow in the future as more options are needed.
> �
> From a command line point of view, this is kind of annoying because the commands end up being quite long. Do you guys think something like Kconfig could help?
> �
> Adrian
> �
>
> �
> --
>
> Thomas T�rnblom, Product Engineer
> IAR Systems AB
> Box 23051, Strandbodgatan 1
> SE-750 23 Uppsala, SWEDEN
> Mobile: +46 76 180 17 80 Fax: +46 18 16 78 01
> E-mail: thomas.tornblom(a)iar.com Website: www.iar.com
> Twitter: www.twitter.com/iarsystems
>
> --
>
> Thomas T�rnblom, Product Engineer
> IAR Systems AB
> Box 23051, Strandbodgatan 1
> SE-750 23 Uppsala, SWEDEN
> Mobile: +46 76 180 17 80 Fax: +46 18 16 78 01
> E-mail: thomas.tornblom(a)iar.com Website: www.iar.com
> Twitter: www.twitter.com/iarsystems
> --
> TF-M mailing list
> TF-M(a)lists.trustedfirmware.org
> https://lists.trustedfirmware.org/mailman/listinfo/tf-m
Note that Zephyr has a Cmake function that does this exact thing, if you're interested: https://github.com/zephyrproject-rtos/zephyr/blob/master/cmake/extensions.c…
Øyvind Rønningstad
-----Original Message-----
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of David Hu via TF-M
Sent: Friday, June 12, 2020 11:45
To: Raef Coles <Raef.Coles(a)arm.com>; Gyorgy Szing <Gyorgy.Szing(a)arm.com>; Adrian Shaw <Adrian.Shaw(a)arm.com>; tf-m(a)lists.trustedfirmware.org
Cc: nd <nd(a)arm.com>
Subject: Re: [TF-M] CMake flags
Hi Raef,
My current prototype is to parse the options from Kconfig output file and "inject" them into CMake (by set()) in a CMake file running at the beginning of build process.
Kconfig generation is an optional step before build starts, only if you request to change the configuration. Otherwise, build just starts with previous Kconfig output config as normal.
Best regards,
Hu Ziji
-----Original Message-----
From: Raef Coles <Raef.Coles(a)arm.com>
Sent: Friday, June 12, 2020 5:34 PM
To: Gyorgy Szing <Gyorgy.Szing(a)arm.com>; David Hu <David.Hu(a)arm.com>; Adrian Shaw <Adrian.Shaw(a)arm.com>; tf-m(a)lists.trustedfirmware.org
Cc: nd <nd(a)arm.com>
Subject: Re: CMake flags
I think I'd advocate for using a workflow of generating a cfg.cmake file from the kconfig, instead of a txt file with an associated cmake parser.
I think it'd be advantageous to not have a parser for an arbitrary config file format in the cmake, since the cmake set(CACHE) statements are already fairly user-readable. It would be different if we could only generate something less readable from the kconfig (i.e. xml/json). It would be ideal if we could generate the statements such that:
set(<variable_name> <default_option> CACHE <TYPE> "<docstring> (<valid_option> | <valid_option> | ...)"
I'm not sure if we can do this in Kconfig, but it seems plausible, and imo is the ideal solution since it gives the user of later editing the cmake config (even using a cache editor like ccmake) while documenting which choices are sane.
Raef
________________________________________
From: Gyorgy Szing <Gyorgy.Szing(a)arm.com>
Sent: 12 June 2020 06:56
To: David Hu; Adrian Shaw; Raef Coles; tf-m(a)lists.trustedfirmware.org
Cc: nd; nd
Subject: RE: CMake flags
Hi,
This requirement could be addressed with a user specific metadata file which is processed by the build-system. Implementation wise the easiest would be to include a file like UserCfg.cmake. This file could contain set(<option> CACHE <STRING|BOOL|PATH> "") commands to set the option variables. User experience wise a text file parsed by cmake scripts would be the best. Such a file could contain variable=value lines.
Longer term when Kconfig support is ready, it could cover these options too. Running Kconfig is a an act done less often, and most of the times, the user is expected to save the Kconfig input and output files. Most of the time the build-system would just use the pre-generated config. I imagine a workflow similar to the one used by the Linux Kernel.
Just my two cent's though.
/George
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org> On Behalf Of David Hu via TF-M
Sent: 12 June 2020 07:33
To: Adrian Shaw <Adrian.Shaw(a)arm.com>; Raef Coles <Raef.Coles(a)arm.com>; tf-m(a)lists.trustedfirmware.org
Cc: nd <nd(a)arm.com>
Subject: Re: [TF-M] CMake flags
Hi Adrian,
Actually we have been investigating configuration alternatives and making some prototypes of Kconfig.
Kconfig can bring some advantages:
* It can setup configuration and check dependency before CMake starts and therefore it can help CMake simplify dealing with the mutual dependencies. For example, if CPU and arch options can be configured in Kconfig, CMake can avoid the workaround to re-configure compiler after parsing platform files.
* Kconfig can collect together the options which are distributed in various CMake files. Developers can more easily understand the options and their dependencies, compared with searching them all around the build system.
But Kconfig may not cover all the flags in command line. Flags such as compiler selections, build type and project config file, may vary in each build. It can be inconvenient to put them into Kconfig.
Best regards,
Hu Ziji
From: TF-M <tf-m-bounces(a)lists.trustedfirmware.org<mailto:tf-m-bounces@lists.trustedfirmware.org>> On Behalf Of Adrian Shaw via TF-M
Sent: Friday, June 12, 2020 12:24 AM
To: Raef Coles <Raef.Coles(a)arm.com<mailto:Raef.Coles@arm.com>>
Cc: tf-m(a)lists.trustedfirmware.org<mailto:tf-m@lists.trustedfirmware.org>
Subject: [TF-M] CMake flags
Hi Raef. Thanks for the presentation today.
If I remember correctly, TF-M's CMake has quite a lot of command line flags now (the ones that are set with -D). And I think this will grow in the future as more options are needed.
From a command line point of view, this is kind of annoying because the commands end up being quite long. Do you guys think something like Kconfig could help?
Adrian
--
TF-M mailing list
TF-M(a)lists.trustedfirmware.org
https://lists.trustedfirmware.org/mailman/listinfo/tf-m
