Nucleo F429 boot - what happens?

[turboscrew]

I have a weird problem. I have the Nucleo board with a TLE493 magnetometer connected. When I connect the USB cable (power the system), file manager opens. If I then close the window, upload my SW using Atollic, my magnetometer code works fine, but if I don't use the debugger, initialization of TLE fails - the I2C HW is already "busy". (TLE starts interrupting via the I2C clock line).

What's different in the boots? The boot0-pin is low by default, and it hasn't been touched.

What code initializes the USB and what else does it do? I'm using I2C2 with PF1 as the clock line.

[Lubos KOUDELKA]

Hello,

please try to verify, if you're clearing I2C interrupt flags properly. It would explain why your code is running in debugger - debugger is reading out registers to display them to the user, and some flags are cleared using read access.

Best regards,

Lubos

[turboscrew]

That I need to check! Thanks! That didn't think of.

I got the darn thing to work by initializing the TLE in the 'main' right after the HW initializations. Not good, but has to do for now.

I still want to fix that in a 'proper' way later.