Enable cache on F15

What

UART Rx could no longer receive new data

How

Flood Rx with many data in a short time

Add hardware I2C support along the current bit-bang implementation

See #20

Currently SPI calls are blocking, which is not good

Currently we simply disable it

The vector table on F15 is not the same as D10

Add the missing SDHC support

A very few modules in libbase are depending on libsc, which was a quick dirty hack at the time. We now employ a no exception policy to that and thus should be eliminated

inno14-d-2014-inno.elf
Thread [1] (Suspended : User Request)
HardFaultHandlerC() at vectors.c:443 0xb7fa
signal handler called() at 0xfffffff9
_malloc_r() at 0x12808
operator new() at 0x12370
libsc::k60::UartDevice::SendBuffer() at uart_device.cpp:205 0x9872
main() at main.cpp:78 0x711a

// Wait until data is received
while (!GET_BIT(MEM_MAPS[m_module]->S, I2C_S_IICIF_SHIFT))
{
    if (GET_BIT(MEM_MAPS[m_module]->SMB, I2C_SMB_SLTF_SHIFT))
    {
        LOG_DL("i2c scl timeout");
        return false;
    }
}

The interrupt flag in I2C status register was 0. The if statement was false.
The I2C interrupt enable flag in I2C control register 1 was found being 0.

Trying to simply set the I2C interrupt flag to 1 if it is 0 to fix this problem.

Timeout while waiting the peripheral to protect against malfunctioning

Oh girl...

A hack is committed in 890c3a4 to workaround the issue, but seriously it's extremely hackish

OnReceiveListener in your SmartCar always gets the first byte at the first callback

For example:

Android: output_stream.write("Hi.");
output_stream.flush();

SmartCar:
1st OnReceiveListener: bytes = 'H' , size = 1
2nd OnReceiveListener: bytes = { 'i', '.' } , size = 2