Comments by Member 12995087

Member 12995087 30-Mar-24 16:16pm View

I am quite sure your VS version is outdated.

Member 12995087 28-Jun-23 4:23am View

May be you can print out via serial or whatever interface a history of Malloc/Free/New/Delete like M0x12345678 F0x12345678... to find out, if a double free/delete or a free/delete of a never allocated memory block is causing the problem?

However, there is always the chance that a bug somewhere else causes corruption in a completely other software module.

If debugging on the ESP32 is so hard, it may help to create a simple Windows project containing your code, where the functions accessing hardware are filled with dummy code delivering the expected function results. In this way, you can use the full power of Visual Studio to find out what is going wrong.

Member 12995087 27-Jun-23 11:34am View

Do you have example code for screen_info_t usage which can reproduce the heap corruption?
You didn't say something if you do new/delete in IRQ or different tasks.

You are sure, it comes from screen_info_t and you call the class only from one task or from main loop?

Memory at address 0x00000000 is valid? iMXRT can be configured in this dangerous way...

When you have the possibility to reproduce the corruption, may be you can set a write breakpoint for the corrupted address and find out, which code part is writing to it.

Comments by Member 12995087 (Top 3 by date)