[Dubbie]

That looks really good – much better.

Thanks for the help.

[Dubbie]

The optimizing worked great. I now have enough memory to finish my project. It did break my LCD drivers by removing my useless wait loops 😀 However, adding asm (“”) inside the loop made the compiler leave them alone.

Thanks for the tip.

[Dubbie]

Thanks for the tip.

As you can see from my second screenshot, I did find the embedded memory explorer. Very handy. I am also using the nano lib. Unfortunately, I have to use the float printf support.

 

I will try the optimise for size flag.

[Dubbie]

Here are the worst offenders from a build that does work (when I comment out a library I wrote)

Attachments:

You must be logged in to view attached files.

[Dubbie]

Just as a followup.

The problem was indeed that I had forgotten to enable SWD in the firmware, (I had temporarily disabled it for a different reason and forgotten to turn it back on) so as soon as the new image booted, then nothing worked.

Thanks for the nudge in the right direction!

[Dubbie]

The disabling does sound plausible. I do have to manually delete the flash before it will work again. Thanks, will investigate in the morning.

[Dubbie]

Self reply…

I have fixed this. As expected, it was an “incompetent user fault”

VS2015 suggested I add a bunch of paths to the VisualGDB include paths. I had said “Sure, why not” One of those paths contained the wrong version of the cmsis_os.h

Deleted that path and everything sprang into life.

😀

[Dubbie]

It does include a system_stm32xxx.c file. However, I just realised that in that file, the definition for AHBPrescTable[16] is present, but there is nothing for APBPrescTable[8].

No idea why this is the case.
I added a new line:

const uint8_t APBPrescTable[8] = { 0, 0, 0, 0, 1, 2, 3, 4 };

And now it seems to build fine.
Thankyou for pointing me in the right direction.

R

[Author]

Posts