Odd SD_MBR_COMMAND_INIT_SD behavior

[ImQ009]

So I have the following bootloader code:

int main() {
    uint32_t code;
    
    sd_mbr_command_t mbr_command = { SD_MBR_COMMAND_INIT_SD, };
    
    code = sd_mbr_command(&mbr_command);
    APP_ERROR_CHECK(code);
    
    code = sd_softdevice_vector_table_base_set(BOOTLOADER_REGION_START);
    APP_ERROR_CHECK(code);
    
    SOFTDEVICE_HANDLER_INIT(NRF_CLOCK_LFCLKSRC_RC_250_PPM_4000MS_CALIBRATION, NULL);
    
    
    
    serial_out("SD INITIALIZED - BOOTING APP\r\n");
    
    
    
    sd_softdevice_disable();
    
    code = sd_softdevice_vector_table_base_set(CODE_REGION_1_START);
    APP_ERROR_CHECK(code);
    
    bootloader_util_app_start(CODE_REGION_1_START);
    
    return 0;
}

Pretty straightforward. Initializing SD, disabling it and then booting the application. However it appears something REALLY odd happens with the sd_mbr_command call. It makes the whole system crash.

So, I set up a breakpoint at that exact line and launched the debugger. I slowly trace the execution, line by line, trying to locate the culprit. It works, sd_mbr_command returns just fine and then the bootloader proceeds as normal.

Head scratching intensifies...

Okay, one more time, re-starting the chip and debugging again, with the same breakpoint. Only this time immediately resuming execution after hitting the breakpoint.

What happens? Absolutely nothing, sd_mbr_command never returns. It's always getting stuck on this exact point: http://i.imgur.com/DzNshAD.png

So. It appears sd_mbr_command works only when executed slowly by stepping into each line manually in the debugger. What.

Any assistance, please? Thank you

[ImQ009]

As I was writing this I had a theory. Checked it immediately after posting.

The solution was to add a short delay before calling sd_mbr_command