Skip to main content
RPipi.1
Associate II
November 5, 2020
Question

TouchGFX bootloader, dfu posibilities

  • November 5, 2020
  • 1 reply
  • 1067 views

Hi,

I have been building on a demo STM32F746g-DISCO touchgfx application for a little while now.

My next thing I want to acheive is software update. I realised that the application do not fit in the internal flash.

So I am thinking about my options. Either:

  1. Run the Application on the QSPI, this should be possible right?
  2. Split up the hex generation for one internal and one external binary, have a custom loader.

The approach I have tried going with is 1). But I am having issues jumping to the application but with no success re-enabling interrupts etc, normaly Hal_Delay() stop working, or some other strange issues, from turning on CPU cache or initiating SystemClock Config twice.

Another question is why are there 2 ld files in the touchgfx project, one ram and one flash.

Which way would you guys have gone solving this?

BR

Robin

This topic has been closed for replies.

1 reply

Martin KJELDSEN
Principal III
November 18, 2020

What won't fit in internal flash? application + assets? Just application? You can place assets in qspi, leaving code in internal.

the .ld files are generated by CubeMX. You want to use the _FLASH version. The RAM version is for putting startup code etc into ram.

e.g.

/* Sections */
SECTIONS
{
 /* The startup code into "RAM" Ram type memory */
 .isr_vector :
 {
 . = ALIGN(4);
 KEEP(*(.isr_vector)) /* Startup code */
 . = ALIGN(4);
 } >RAM

/Martin