sd_init_driver() bricking the Pico

[mchobby]

Hi,
I'm trying to include the no-OS-FatFS-SD-SPI-RPI-Pico into the Picoterm project .
Constraint: because of the VGA rendering and USB Host (for Keyboard), I cannot use the UART debugging, Neither USB debugging.
So I do send debugging message over GP28 with a PIO TX based uart (see the debug_print() call below in the code).

As current configuration I'm using the standard hw_config configuration together with an Adafruit microSD breakout (for SPI).
The config & wiring are: SO->GP16, CS->GP17, CLK->GP18, SI->GP19,Detect->GP22, 3.3V & GND.

From my main() I do call the test_sdcard() available here below after the sdio_init_all() call.
I can see the message "SD Card driver initializing" then the Pico stops running!
This means that sd_init_driver() does crashed.

The result is the same with an SD inserted and without SD.
Did I miss something obvious ?

void test_sdcard() {
	FRESULT fr;
	FATFS fs;
	FIL fil;
	int ret;
	char buf[100];
	char filename[] = "test02.txt";

	// https://www.digikey.com/en/maker/projects/raspberry-pi-pico-rp2040-sd-card-example-with-micropython-and-cc/e472c7f578734bfd96d437e68e670050
	 debug_print("------BEGIN SDCARD test----");
	 debug_print("SD Card driver initializing");
	 // Initialize SD card
	if (!sd_init_driver()) {
		 debug_print("ERROR: Could not initialize SD card\r\n");
		 return;
	}
	debug_print("SD Card driver initialized"),
	// Mount drive
	fr = f_mount(&fs, "0:", 1);
	if (fr != FR_OK) {
			sprintf( debug_msg, "ERROR: Could not mount filesystem (%d)\r\n", fr);
			debug_print( debug_msg );
			return;
	}

	// Open file for writing ()
	fr = f_open(&fil, filename, FA_WRITE | FA_CREATE_ALWAYS);
	if (fr != FR_OK) {
			sprintf( debug_msg, "ERROR: Could not open file (%d)\r\n", fr);
			debug_print( debug_msg );
			return;
	}

	debug_print("------END OF SDCARD test----");

}

[carlk3]

I didn't dig into how you've implemented everything, but people have had similar sounding problems with exclusive DMA_IRQ conflicts in the past. (For example, see #8.) If you're using the sdio branch, you could try setting the DMA_IRQ_num in the spi_t instance to DMA_IRQ_1. (It defaults to DMA_IRQ_0.) For example, see https://github.com/carlk3/no-OS-FatFS-SD-SPI-RPi-Pico/blob/sdio/examples/command_line/hw_config.c.

If that's not it, you might have to resort to enabling or adding some printfs. For example, (again, assuming sdio branch), there is a USE_DBG_PRINTF compiler define that you can enable (e.g., uncomment https://github.com/carlk3/no-OS-FatFS-SD-SPI-RPi-Pico/blob/sdio/examples/command_line/CMakeLists.txt#L37).

[evolutional]

fwiw I'm using main branch and I get the same problem, soon after calling the init function (and doing nothing else), I get this output over usb serial shortly after.

*** PANIC ***

Hard assert

I can get to the point of mounting the drive, but it also panics soon after.

EDIT: I think you may be right about the DMA IRQ conflict; the panic went away after I disabled the video output. Investigating...

EDIT2: Yes, it was the DMA + assumptions about the channels in use; I changed the qvga video code to claim unused channels and use DMA irq1 and it all works with this lib nicely