dist/testbed-support: use '.elf' binary for dwm1001 board

[blueted2]

Contribution description

When uploading firmware for the dwm1001 to the test-bed, this Makefile attempts to use the .hex binary, however it appears that one isn't generated for this board, thus it generates an error.

Command:

make iotlab-exp

Error:

iotlab-experiment: error: [Errno 2] No such file or directory: '<firmware>.hex'

This commit makes it so that the .elf binary is used for both m3:* and dwm1001:* boards. I don't know if this issue exists for any other boards.

---

My top-level Makefile configuration:

IOTLAB_SITE ?= lille
IOTLAB_TYPE ?= dwm1001:dw1000

---

Testing procedure

This commit now allows the firmware to be uploaded for the dwm1001 board using make iotlab-exp.

Issues/PRs references

[maribu]

@aabadie: Could you take a look at this one?

[aabadie]

I never tried that part of the build system with IoT-LAB. Looking at the changes I see 2 things:

• this Makefile.iotlab could be used not only with m3 and dwm1001 but also with nrf52xxxdk and other boards. Only zigduino should use .hex and firefly should use .bin
• Instead of using an elf file, I think it's more optimal to use .bin files because the firmware is much smaller, so it saves some bandwidth when uploading the file to IoT-LAB

[blueted2]

Good points. It seems fit iot lab complains if you try to start an experiment directly with a .bin (without uploading the firmware first, it shows a pop-up with Wrong format for this type of nodes. (expected elf)), but once it's uploaded it's fine.

[blueted2]

Okay, so it sees that fit iot doesn't like starting experiments directly with the .bin file, giving the following error on the dashboard: screenshot

It does however seem that first launching the experiment with no firmware, then flashing the firmware seems to do the trick.