rtos | A cobbled together RTOS for ARM microcontrollers

Solved. The solution was to generate new rsa and dsa keys with:

It is installed directly into the interrupt vector table. If the vector table uses CMSIS names for the handlers then you can map the CMSIS name to the name of the FreeRTOS systick handler in FreeRTOSConfig.h, as per the FAQ - see the red "special note to ARM Cortex-M users" here: https://www.freertos.org/FAQHelp.html

There's no easy way to mix freeRTOS and deep_sleep mode. During deep sleep the CPU is powered down and its context is lost, yet the RTC memory can be retained. As all SRAM's content is lost, there's no easy backup-restore we can do here to safely restore everything after coming out from deep sleep.

But what you can do is to bring everything down to a safe state before entering deep-sleep, you can signal all your tasks to finish what they're doing and exit, and then take some advantage of ESP32's relatively low wake-up latency. It is a very unpleasant inconvenient for a Wi-Fi connected device, but more or less acceptable for BLE devices that will wake up and send a beacon once in a few seconds.

You would also want to fine-tune the second stage bootloader's configuration by enabling the CONFIG_BOOTLOADER_SKIP_VALIDATE_IN_DEEP_SLEEP option so waking-up from deep-sleep would be faster than booting from a cold reset.

(1) An RTOS is a Real Time Operating System. Implementing an operating system is not the same thing as using an operating system. It seems like you'd know whether you're programming using a Real Time Operating System or nothing. And that's the difference between using a RTOS and bare metal.

Note that the RTOS code is bare-metal programming, because it's not using any lower-level software. And then when you write your code using the RTOS, it's not bare-metal programming, because you're using the services of the RTOS.

(2) It seems like you'd know whether you're implementing an operating system or an embedded application And that's the other difference.

(3) As regards an SoC - that's a hardware category. Is there one integrated circuit containing the CPU and a bunch of associated functions (interrupt controller, maybe an MMU, peripheral interfaces, network, etc.)? Then it may be a SoC. Or are there a few other ICs providing these functions? Then it's not a SoC.

You did not specify if you had any requirements in terms of FreeRTOS version, but you can either:

• use the demo provided in file SW-EK-TM4C1294XL-2.1.4.178.exe available on TI WEB site as is - you will find it in directory examples\boards\ek-tm4c1294xl-boostxl-senshub\senshub_iot
• use this example as a basis to use a more recent version of FreeRTOS: you just would have to replace FreeRTOS source code by the most recent one, and maybe to modify some code in the demo: some function names/signatures may have changed across major versions.

The procedure hereafter describes step by step how to create a minimalist FreeRTOS program with Code Composer Studio 10.3.0 and FreeRTOS v202104.00 in a Windows 10 environment using the second approach. You may have to adjust the drive letter to you specific setup, I am using D: for the purpose of this example..

• Download Code Composer Studio 10.3.0, FreeRTOS v202104.00 and SW-EK-TM4C1294XL-2.1.4.178.exe.

• Install Code Composer Studio with support for the Tiva-C MCU familly. When prompted for a workspace name, specify D:\ti\workspace_v10.

• Unzip FreeRTOSv202104.00.zipinto D:\.

• Unzip SW-EK-TM4C1294XL-2.1.4.178.exe into D:\SW-EK-TM4C1294XL-2.1.4.178.

• Launch CCS

• Use the menu item File/New/CCS Project, and create an 'Empty Project (with main.c).

[]

• Click on the Finishbutton.
• Create the following directories:
  D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\driverlib
D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\inc
D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\FreeRTOS-Kernel
D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\FreeRTOS-Kernel\include
D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\FreeRTOS-Kernel\portable\GCC
D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\FreeRTOS-Kernel\portable\GCC\ARM_CM4F
D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\FreeRTOS-Kernel\portable\MemMang
• Copy D:\SW-EK-TM4C1294XL-2.1.4.178\examples\boards\ek-tm4c1294xl-boostxl-senshub\senshub_iot\FreeRTOSConfig.h into D:\ti\workspace_v10\TM4C129EXL-FreeRTOS
• Copy all .h files from D:\SW-EK-TM4C1294XL-2.1.4.178\driverlib into D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\driverlib.
• Copy D:\SW-EK-TM4C1294XL-2.1.4.178\driverlib\gcc\libdriver.a into D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\driverlib.
• Copy all .hfiles from D:\SW-EK-TM4C1294XL-2.1.4.178\inc into D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\inc.
• Copy all files present in D:\FreeRTOSv202104.00\FreeRTOS\Source\include into D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\FreeRTOS-Kernel\include.
• Copy all .cfiles present in D:\FreeRTOSv202104.00\FreeRTOS\Source into D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\FreeRTOS-Kernel.
• Copy all file present in D:\FreeRTOSv202104.00\FreeRTOS\Source\portable\GCC\ARM_CM4F into D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\FreeRTOS-Kernel\portable\GCC\ARM_CM4F
• Copy D:\FreeRTOSv202104.00\FreeRTOS\Source\portable\MemMang\heap_4.c into D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\FreeRTOS-Kernel\portable\MemMang.
• Edit D:\ti\workspace_v10\TM4C129EXL-FreeRTOS\main.c, and replace its content by:

Could it be the change in size from the last link?

Yes.

There is no particular reason that you should use pointer to signed rather than unsigned.

You are probably never going to dereference this pointer directly to access the words on the stack. If you do want to access the data on the stack, some of it will be signed, some unsigned, and some neither (strings etc) so the pointer type will not help you with that.

When you want to pass around a pointer but never dereference it then one convention is to use a pointer to void, but that convention isn't so popular in embedded system code.

One reason to use a pointer to a 32-bit integer is to suggest that the pointer is at least word-aligned. If you intend on complying with the ARM EABI (which you should) then the stack should be doubleword (64-bit) aligned at the entry to every EABI compliant function. To hint that that is the case you might want to even use a (u)int64_t pointer. This might be misleading though because not everything on the stack is 64-bit or 32-bit aligned, just the whole frames.

Here is the full first batch file rewritten:

Going top-to-bottom, you have your application code which in the end will likely want to talk to some server: resolve DNS name, open a tcp connection etc. This is what LwIP provides - a set of API functions: socket functions, DNS functions and other. For example, when you do a TCP socket connect to a given IP address and port, it decides which out of available network interfaces to use, constructs a frame in form of array of bytes and sends those bytes over that interface. This is where LwIP part ends - it requesting given network interface to output X bytes it provides, as well as consuming any bytes that may arrive on that interface. It doesn't know how exactly to make the bytes "come out", but it knows how to construct the data to send and understands the data you give it.

Going bottom-to-top, you have your modem - LTE, 3G, 2G, doesn't really matter. Modems provide a set of AT commands to talk to them to perform a set of functions: set SIM card PIN, get signal quality, list available operators, select an operator etc., as well as a way to switch it to PPP mode, which also done ATD command. Assuming the modem has been configured properly before, once you switch it to PPP mode it'll be able to send data using PPP protocol and will also "spit out" any data it receives. This is where the modem part ends - when switched to PPP mode, the modem is able to send and receive raw network traffic using PPP protocol. It knows how to output the raw bytes you give it, but it doesn't understand them on a very high level.

Your role is to connect (interface) both parts. Your modem uses PPP to produce and consume network traffic. LwIP has the capability to understand and produce PPP frames.

In case of sending data out, after preparing a PPP frame LwIP is going to call the sio_write function that is expected to send provided bytes to modem that's already in PPP mode. This is the part that you need to fill in. In case of reads sio_read is used and it's your job to fill it in so that it returns bytes that were received from the modem. How you're going to do that - using RTOS and queues for bytes, no RTOS or any other way - doesn't matter. It's what you find more convenient or what fits your overall project structure better. As long as those functions do what they're supposed to, LwIP will be happy to use them.

This interfacing is discussed in more detail here: https://lwip.fandom.com/wiki/PPP#PPP_over_serial. Again, the general concept is that LwIP is just a software library and in the end it'll want to send and receive bytes. Your job is enabling it to do so, by filling in the functions it expects.