A bootloader on a microcontroller is a very useful thing. It allows me to update the firmware in the field if necessary. There are many ways to use and make a bootloader (see “Serial Bootloader for the Freedom Board with Processor Expert“). But such a bootloader needs some space in FLASH, plus it needs to be programmed first on a blank device, so a JTAG programmer is needed. That’s why vendors have started including a ROM bootloader into their devices: the microcontroller comes out of the factory with a bootloader in FLASH. So instead writing my bootloader, I can use the one in the ROM.
FRDM-KL03Z with ROM Bootloader
And as with everything, there are pros and cons of that approach.
During Embedded World 2017 in Nürnberg I was lucky to get a handful LPC800-DIP boards. To get all students who were lucky to get one, here is a tutorial to make that very exciting ‘blinky’ application on that board:
I’m using Eclipse based IDE’s to develop and debug my embedded applications. This works great, as Eclipse has all the necessary tools to edit, build and debug it. But when it comes just to download/flash a binary to the board, then things are pretty much specific to the tools used. With the advent of the new MCUXpresso IDE, here is how that Eclipse IDE can be used for this.
The MCUXpresso IDE (see “MCUXpresso IDE: Unified Eclipse IDE for NXPs ARM Cortex-M Microcontrollers“) has one great feature: it includes debug support for the popular LPC-Link2 debug probes. That way I have yet another powerful debug probe with extra features for ARM based boards. That LPC-Link2 circuit is present on many LPCXpresso boards from NXP. So why not using it to debug it my custom hardware?
Looking for a small, inexpensive ($25-30) ARM development board (say 120-180 MHz ARM Cortex-M4 with FPU, 512kB-1MB of FLASH and 256 KByte of RAM? Then have a look at the Teensy 3.5 and Teensy 3.6 by PJRC/Paul Stoffregen:
Teensy 3.5 with NXP K64F ARM Cortex-M4F
The only problem? it is not possible to debug it :-(. At least not in the traditional sense. This article is about how to change the board to use it with any normal SWD debugging tool e.g. Eclipse and the Segger J-Link :-).
There are many mergers going on in the industry, and one of the largest one was in 2016 the integration of Freescale Semiconductor with NXP Semiconductors, with both providing Eclipse based IDE’s to their customer base. Consequently, the company merger triggered a merger of the IDE’s, and last week NXP has released the result: the MCUXpresso IDE.
For a CubeSat project we only have a single board available. But multiple universities and developers need to have access to that board for developing and debugging the firmware. We cannot easily ship around the board: that takes a lot of time and during shipment nobody can use the board.
There is a nice feature in the Segger J-Link software which allows to share the debug connection over the network: the J-Link Remote Server. It even works nicely between different networks without complicated firewall setup:
When working and debugging a bootloader, debugging can be a challenge: During debugging the bootloader, a new binary gets loaded into the microcontroller address space which is unknown to the debugger. As soon as I step into the newly loaded binary, I only see assembly code, with that ugly “No source available” in Eclipse:
No Source Available, debugging in assembly
But wait: GDB is able to do pretty much everything you can imagine, so here is how to debug multiple binaries with GDB and Eclipse, and to turn the above into something which is easy to debug:
The Hexiwear docking station would have a nice feature: it has embedded a debug circuit (OpenSDA). That way I would not need an external debug probe to debug the Hexiwear. However, a debug probe is required to reprogram the docking station itself:
Repgrogramming the Mikroelektronika Docking Station
The Hexiwear device is a great and versatile device with two microcontrollers on it. Developing firmware on a Hexiwear means changing what was originally on it. And sometimes it happens that I’m not sure if the changes are for good. Or that I accidentally destroyed the firmware on the NXP Kinetis KW40 BLE microcontroller :-(. So I had to find a way to restore the original firmware, and this is what this post is about.
Restoring the Hexiwear Firmware with a Segger J-Link
My Toradex i.MX7Dual module comes with a preflashed Linux distribution (see “Tutorial: First Steps with NXP i.MX7 and Toradex Colibri Board“). As with any other things, Linux gets updated from time to time, and Toradex publishes new firmware. In this article I’m documenting how I can update Linux in the external FLASH on that module.
From time to time I face some problems which are really hard to find. Mostly these kind of bugs are very timing sensitive and depend on interrupt execution order. Maybe a dangling pointer is overwriting memory, code is running wild, or some functions are not reentrant as they should be. For these kind of bugs, good tools are worth their weight in gold. The Percepio FreeRTOS+Trace and the Segger SystemView have helped me many times to narrow down such kind problems in my applications. Another ultimate tools is hardware trace: Now I have a Segger J-Trace Pro for ARM Cortex-M in my arsenal of bug extinguishing weapons on my desk:
Dear bugs, look what I have on my desk. Your hiding time is over! 🙂
Getting a board from a distributor like Farnell/Element14/Mouser (add your own distributor) means that chances are high that the default firmware on it is written years from now because the inventory has not been updated, or because boards are still produced with that original firmware (because of testing?). So what happens if I use board with a firmware developed pre-Windows 8/10 area?
Freshly Unboxed NXP FRDM-KL25Z Board
It might work, but chances are high that the bootloader and firmware is not ready for the ‘modern age’, and as a result the board might be bricked. If you still have a Windows 7 machine around (I do!), you are lucky. If not, then you need to read this article….
More and more of my students are using Microsoft Windows 10 machines, and my computer has been upgraded to Windows 10 a couple of week ago too. From my work and experience, a new operating system causes always some challenges, and Windows 10 is no difference. And no, this is not about Microsoft vs. Apple vs. Linux, this post is about addressing a potential and painful problem which I have observed with Windows 10 machines, and to my understanding it could happen with any other operating system too. The problem is that somehow on several student machines the bootloader and OpenSDA application on their FRDM boards did not work any more.
FRDM-K64F (top) programming the OpenSDA Bootloader (bottom)
Sometimes it is very convenient to load a new firmware to a board without the need for a hardware debugger. This is usually done with a bootloader. The NXP Freedom and Tower evaluation boards have on-board debug device/microcontroller (OpenSDA) which can load different firmware implementations like CMSIS-DAP/mbed, P&E Multilink or a Segger J-Link OpenSDA applications. Both mbed and P&E implemenations support to program the board with drag&drop: simply send a file to a virtual MSD (Mass Storage Device) to get it programmed. The latest Segger OpenSDA firmware has this ability added now too: Programming the board with a virtual MSD device:
Related to my earlier article about using OpenOCD, I want to share something I have learned (again) with OpenOCD v0.10.0:
I was running often into the following error:
Warn : Cannot communicate... target not halted.
Error: auto_probe failed
Error: Connect failed. Consider setting up a gdb-attach event for the target to prepare target for GDB connect, or use 'gdb_memory_map disable'.
Error: attempted 'gdb' connection rejected