Sigma7

I have played around with several Netduino types (Netduino Mini, Netduino (1), Netduino Plus and Netduino 2)...

On the Netduino 2, I power it up with the button (BTN) held down (and Windows thinks the USB device is a "STM32FBOOTLOADER").  Then I'm able to download my firmware with STM's "DfuSe Demo (v3.0.3)" program.  This has worked fine for three Netduino 2's that I've gotten over the past year.  But I just ordered/received another one.. and the DfuSe download/upgrade seems (?) to work fine... but I can't get any response for the Netduino 2.  Any thoughts on what I can try?

(The firmware I'm downloading is an extremely simple boot monitor (command interpreter) that uses USART2.  And I use an ADAfruit 954 USB-to-TTL cable on Digital I/O pins 2 (RX) and 3 (TX).)

Thanks,

Tom

I have been working with Netduino Mini, Netduino 1 and Netduino Plus..  so I thought I'd try doing some Cortex-M3 "bare metal" programming on the Netduino 2 (N2).

I power-up the N2 with the pushbutton pushed, so the "STM32F2BOOTLOADER" gets invoked (and I'm able to upload/download files to flash with the STM DfuSe tool).  So far so good.

The GNU ARM toolchain I was using didn't support "-mcpu=cortex-m3 -mthumb."  So I grabbed the latest and greatest YAGARTO release.  Now things seem to build okay.

The binary file I download to flash (08000000) basically looks like:

0: 2000 e000

4: 0800 0021

20: <my program>

"My Program" basically does (all I'm trying to do is to print a character to the serial port (USART2)):

Enable USART2 in RCC_APB1ENR

Enable PIOA in RCC_AHB1ENR

for the USART2_TX pin (PIOA pin 2 Alternate Function 7)..

GPIOA_MODE(pin 2) = GPIO_MODE_AF

GPIOA_OSPEED(pin 2) = GPIO_SPEED_50MHz

GPIOA_OTYPE(pin 2) = GPIO_OTYPE_PUSHPULL

GPIOA_PUPDR(pin 2) = GPIO_PUPDR_NOPULL

GPIOA_AFRL(pin 2) = 7

the I setup the USART (USART2)

I set it up for 8-bits, no parity, 1 stop bit, no hardware flow control, and 38400 baud (which translates to DIV_Mantissa = 0x30 and DIV_Fraction = 0xd).  I enable the transmitter (CR1.TE) and receiver (CR1.RE).. then enable the USART itself (CR1.UE).

Then I just go into a loop, outputting a character as soon as the USART STATUS.TXE goes to 1.

I have one specific question, and one general one:

(1) When I look at the "Reset Sequence" in "The Definitive Guide to the ARM Cortex-M3" (pp. 44-46), it says that the reset vector should have its LSB set to 1 (for thumb).  Sounds reasonable, but I've seen many other reset examples that don't mention that at all.  Is "The Definative Guide" right on this?

(2) I know it's just pseudo-code, but does anything strike anyone as a glaring omission and/or mistake?

Thanks,

Tom