I'm having a strange problem that's been holding me up for the past few days. I've been trying to configure the FMC to work with a Cypress CY7C1051DV33 SRAM on a PCB that I developed. In fact, the configuration that I used worked, and I was able to read and write successfully. However, I noticed a small hiccup with the ST-Link debugger, and I always had to hit the go button twice whenever I initialized the FMC.
I didn't understand what was happening with the debugger, but I pressed on to work on some LCD code. At that point, I began to lose the connection with the debugger. I finally realized it was because the microcontroller went into a never-ending reset.
I got rid of the LCD code and pared out a few other things--the short story is I now had a piece of code that started up, turned on and LED, and initialized the FMC for SRAM. After some additional debugging I narrowed my problem down to the initialization of 3 particular I/O pins. At that point I suspected some hardware problem and reworked a few traces on the PCB that I found thought might be problematic (but skeptical that this would have anything to do with the reset).
To simplify things, I commented out most of the FMC configuration and only configured one of my problem I/O pins. Now I have a fairly simple piece of code that lights up an LED and configures a single I/O pin with the alternate GPIO_AF12_FMC function (if I configure this pin without the alternate function, the reset never occurs). The config code is as follows:
/* GPIOE configuration */
When HAL_GPIO_Init is called it sets GPIOx->AFR= GPIO_AF12_FMC (i.e. 0x0C).
When I run this code on the STM32F429I-DISCO board the reset never occurs. I physically disconnected the PE0 pin from all circuits on both PCBs just to make sure that something physically wasn't causing the reset.
There's apparently some hardware difference. The discovery board uses the 429, my board uses the 439. The microcontroller is clocked from the internal oscillator. Power seems stable.