STM32 MPUs Embedded software and solutions

During build with DISTRO=openstlinux-x11 IMAGE=st-example-image-x11 MACHINE=stm32mp1-disco on Ecosystem-1.2 errors arose:------------------------------------------------------->ERROR: st-image-vendorfs-1.0-r0 do_rootfs: The postinstall intercept hook 'update_udev_hwdb' failed, details in /home/support/STM32MP15-Ecosystem-v1.2.0/Distribution-Package/openstlinux-20-02-19/build-openstlinuxx11-stm32mp1-disco/tmp-glibc/work/stm32mp1_disco-ostl-linux-gnueabi/st-image-vendorfs/1.0-r0/temp/log.do_rootfsERROR: st-image-vendorfs-1.0-r0 do_rootfs: Function failed: do_rootfsERROR: Logfile of failure stored in: /home/support/STM32MP15-Ecosystem-v1.2.0/Distribution-Package/openstlinux-20-02-19/build-openstlinuxx11-stm32mp1-disco/tmp-glibc/work/stm32mp1_disco-ostl-linux-gnueabi/st-image-vendorfs/1.0-r0/temp/log.do_rootfs.16879ERROR: Task (/home/support/STM32MP15-Ecosystem-v1.2.0/Distribution-Package/openstlinux-20-02-19/layer

I've included the stock OpenAMP echo code for the M4 from CubeIDE 1.10-2 and am able to echo a message to /dev/ttyRPMSG0 and have cat print the echoed message.I've written a simple app in C to open the RPMSG0 device, write(2) a structure and then read(2) the structure back. I've modified the OpenAMP callback code to modify a sructure member and a debugger on the M4 shows that VIRT_UART_Transmit() buffer arg(2) contains the changed structure member value, however, my read(2) on the A7 displays the original structure member value.This is for a POC so I'm hoping to simply use /dev/RPMSG driver subsystem as is.Can the standard st-linux RPMSG facility be used to pass structures to M4 and back to A7 as I'm attempted to do .... open(); write(); read() - no close() as I'm expecting to write and read more messages ?Are there /dev/RPMSGn settings I need to use in order for this to work as intended ?Are there examples of how

Hi! I am following the tutorial How to optimize the boot time for stm32mp157c-ev1 I have enabled FastMap but the time is not reduce (about 10s to attaching ubi) I want to avoid attaching ubi... Is it posible load kernel and device tree in nand using STM32_Programmer?

Using the Distribution Package for STM32MP1 I'm able to access to the rootfs partition, example:my_postprocess_function() { cp -dr /home/osboxes/www ${IMAGE_ROOTFS}/opt/ chown -R daemon:daemon ${IMAGE_ROOTFS}/opt/www/ }But how to access to bootfs partition? For example to add the `boot.scr.uimg` or `rproc-m4-fw.elf`. I cannot find a `${IMAGE_BOOTFS}` variable.

Trying to build image using unmodified distribution package. Hardware crypto devices are enabled in config:root@stm32mp1:/etc# zcat /proc/config.gz | grep CONFIG_CRYPTO_DEV_STM32CONFIG_CRYPTO_DEV_STM32_CRC=yCONFIG_CRYPTO_DEV_STM32_HASH=yCONFIG_CRYPTO_DEV_STM32_CRYP=yBut 'cat /proc/crypto' shows no 'stm32' drivers. What am I doing wrong?

Hello,I use STM32MP157C-DK2 board with freeRTOS.I run my bord in engeneering mode with default freeRTOS parameters provided by cubeMXI create tree threads in my application but the last does not start.May be it is a memory space problem. So I try to increase the parameters TOTAL_HEAP_SIZE but when I run again no threads works.What parameters could I change to have more free space ?Best regards,Arnaud Bousquet

Hello,I wish use FreeRTOS with STM32MP157C. When I configure my bord on CubeMX a message pop up says I should configure SYS parameter with an others time source than sysTick. But if I configure TIM1 for example I cannot assign TIM1 in SYS section .cardHave you any ideas to solve this problem ?Best Regards,Arnaud Bousquet

I guess, it has something to do with the secure monitor inited by TF-A, since GPIOZ is protected.   Maybe there is something wrong with the default device tree using the GPIOZ?   What can I do?  

I've run a simple graphical program with Mono and OpenTK. It contains a simple fragment shader (demo). One frame renders about 7 seconds and CPU usage is 100%. Is it normal?vec2 p = ( gl_FragCoord.xy / resolution.xy ) - 0.5; float sx = (amp)*1.9 * sin( 4.0 * (freq) * (p.x-phase) - 6.0 * (speed)*time); float dy = 43./ ( 60. * abs(4.9*p.y - sx - 1.2)); dy += 1./ (60. * length(p - vec2(p.x, 0.))); gl_FragColor = vec4( (p.x + 0.05) * dy, 0.2 * dy, dy, 2.0 );​

Hello ST communityI am having trouble configuring the Ethernet PHY on my custom board using the stm32mp153aac. The MPU is supposed to generate a 50 MHz reference clock for the transceiver, but when I measure the pin with the oscilloscope I get nothing. The reset pin also seems to be configured as input and never changes.The ethernet transceiver is connected according to "Figure 46. 10/100M Ethernet PHY connection with REFCLK from RCC example" (external ETH_CLK 50 MHz reference, RMII mode, 100 Mbit).https://www.st.com/content/ccc/resou

When I run my M4 Application every thing is working fine. But while the Board boots until i start the M4 remoteproc the pin is on high level. How can I set the pins reset(boot) configuration to low? I'am using the developer package.

Hi, I need to save a variable from u-boot console but is not possible because the command SAVEENV is disabled (STM32MP1 module) .   How can I do?  

Hello,i use STM32mpu1-dk2 and want to implement an update over the air mechanism. For that i would duplicate all partitions from fsbl to u-boot, that meansfsbl, teeh, teed, teep, u-boot.Now i would like to know how one can tell TF-A which partion should be choosen for the next boot stage. That means, imagine we have 2 partioitns for OP-Tee and we have updated Op-TEE (teeh, teed teep), now we want fsbl to starts from the new updatetd OP-Tee partition.Any Idea how can this goal be achieved using TF-A?Thanks

Hello, I received the STM32MP1 discovery kit and I started to use it.I installed SDK and I compiled successfully some hello world applications.Now, I would like to build some extra libraries to add, or in the future I may want to rebuild existing ones with newer versions.These packages are typically distributed with a configure script made with autotools, or with CMake build scripts, or with other less widespread tools.I would like to know what are the options to be used for correct integration with SDK and inside the OpenSTLinux distribution.For example, with autotools you may need to set options like "--bindir", "--includedir", etc, or CMake needs to receive the right information from an external file assigned to the CMAKE_TOOLCHAIN_FILE macro.After that, you usually build with "make" and "make install" and you did the job.What are the recommended options for expanding SDK?Thank you very much for your time.

Looking at this code:https://github.com/ARM-software/arm-trusted-firmware/tree/af1ac83e0fa4e77aad13e1e8e47b6fcafeb17dbc/drivers/st/mmcI don't understand if the methods are named after the sdmmc2 bus of the MPU (#1 for sd card, #2 for emmc) or it's just a suffix without any special meaning. In other words, they are used also to access to the sd-card?

Hi there,I am trying to get the ID of the MP1 from a given address. But when I try to do so in Linux, using the microprocessor core A7, it gives me a Segmentation fault (core dumped) error.When I run the same program on the microcontroller, core M4, it prints the ID just like this:[0]2359324[1]859328773[2]876098352How can I get that same ID when running my code on Linux?Here is my code:#include <stdio.h> #include <stdint.h>   // 0x5C005234 is the address where the ID is stored #define VAL ((uint32_t *) 0x5C005234)   void main(){   printf("[0] %d [1] %d [2] %d", VAL[0], VAL[1], VAL[2]);   }

I'm working on getting a Distribution Package build working on an aarch64 machine, specifically a Nvidia Jetson Nano. After downloading the 1.1.0 ecosystem package and sourcing the bitbake environment, bitbake st-image-core runs successfully for a while but fails while executing the strip command in the populate-sysroot section of gcc-arm-none-eabi-native-7-r0. When I execute the strip command by hand the failure is actually an "unknown format" error on the file being stripped. When I execute the "file" command against the file giving the error it has been compiled for x86_64. This would imply, I think that the recipes for gcc have be set up for a 64-bit Intel architecture. My question is where do I modify the recipes to ensure they are using the local machines architecture to build the components?An example of the error I am getting is:ERROR: gcc-arm-none-eabi-native-7-r0 do_populate_sysroot: Fatal errors occurred in subprocesses:Command '['strip', '--re

Hello ST communityI'm trying to configure the USB in OTG mode so I can upload linux to the NAND flash on my board (I'm using the STM32MP153AAC on a custom board).Here's the output from the debug serial:https://pastebin.com/f1UYcyDqAnd here is dmesg on the host computer:https://pastebin.com/Mj4fMGT2From what I can tell, stm32prog is being executed by U-Boot on the board with the arguments "stm32prog usb 0". This causes the stm32-usbphyc device driver to be probed (successfully), and as can be seen in the dmesg output on the host computer, a new USB device is detected, but it doesn't seem to be communicating properly. The error is "device descriptor read/64,

Hello communityProblem in short:The boot process hangs when U-Boot tries to load the stm32_fmc2_nand driver. Serial output can be found here:https://pastebin.com/VgJG6rSFI have designed and built a custom board using the STM32MP153AAC. The peripherals share a lot in common with the STM32MP157C-DK2 dev board (serial is also connected to uart4, PMIC is also connected to i2c4) so I used the STM32MP157C-DK2.dts device tree source file as reference for a lot of the changes I made to my own device tree source file.My top-level dts can be found here. It was auto-generated by CubeMX and then manually edited here and there:https://pastebin.com/pUJsdDzHYou can

Hi guys,I just wanted to share my work: STM32MP1 port for Arduino is merged! I have spent months on this work.https://github.com/stm32duino/Arduino_Core_STM32/pull/717With this merge, you can run Arduino on the M4 coporcessor and it can use the Arduino header pins, while the host Linux still have access to the RPi HAT header pin.It is not landed to a release yet, so you cannot use it out-of-the-box yet but additional tricks are needed to build Arduino_Core_STM32 from source.This merge includes basic peripherals like SPI, I2C, PWM, ADC, and there is no direct communication between A7 and M4 yet. I'm working on OpenAMP RPMsg support (/dev/ttyRPMSG0) for now:

$ stm32mp1setup [1/3]: Apply patches Apply patches based on required configuration ERROR: not possible to apply kernel-headers/0001-stm32mp1-Align-kernel-headers-to-CK-4.19.17 on external/kernel-headers [2/3]: Load graphics libraries => Extracting data to device/stm/stm32mp1/graphics => Extracting data to device/stm/stm32mp1-kernel/gcnano-driver => Applying required patches to device/stm/stm32mp1-kernel/gcnano-driver [3/3]: Load op-tee user module sourcesWhen I try to build kernel:$ load_kernel   Start loading the kernel source (Linux) ERROR: Not possible to apply patch .../build/device/stm/stm32mp1-kernel/source/patch/4.19/0001-ARM-stm32mp1-r2-MACHINE.patch, please review android_kernel.configWhen I try to make:​$ make -j ============================================ PLATFORM_VERSION_CODENAME=REL PLATFORM_VERSION=9 TARGET_PRODUCT=aosp_eval TARGET_BUILD_VARIANT=userdebug TARGET_BUILD_TYPE=r

When signing OP-TEE pager and pageable binaries, the information written to the BinaryType field of the STM32 header (byte 255), unless I have wrong view about semantics (teex->pager, teed->pageable) and typical values (pager->0x21, pageable->0x22), is erroneous.Signing tool command `--type teed` writes 0x21 and `--type teex` 0x22, while 0x22 and 0x21 are expected respectively.e.g., signing OP-TEE pageable binaryRegards,Otto

I am looking for an example using spi6 which is only accessible by A7 core. Explicitly how to enable it in the device tree and a quick example for sending a byte.

Hi,I've been looking at the following NAND flash to use as a "mass" memory for my MP1 project (to hold Linux and data storage during run time). Booting and setup is done via QSPI NOR Flash.https://www.mouser.fi/datasheet/2/877/datasheet_en_20181214-1484073.pdfThe NAND Flash does, however, have two sets of chip selects and Ready/Busy outputs.Is there a way to interface this with the FMC to gain access to all 16Gb of storage using the MP153C? The reference manuals do refer to multiple NCEs but I don't find them on the pin list in CubeMX.Do I just need to use ADDR[25:24] for the chip selects? And then mux the Ready/Busy outputs from the memory together to the NWAIT input on the FMC?Thank you for your help,Tero

..