ti-rtos i2c fast mode plus with TIVA

Hi Michael,

I've files a bug ID to track this enhancement (SDOCM00118015).

The TI-RTOS I2C does not directly support FM+, though it shouldn't be too hard to add it. The driver depends on TivaWare to set the I2C controller's 1MHz bitrate. Looking at the driverlib APIs, there is a remark in I2CMasterInitExpClk to configure it.

//*****************************************************************************
//
//! Initializes the I2C Master block.
//!
//! \param ui32Base is the base address of the I2C module.
//! \param ui32I2CClk is the rate of the clock supplied to the I2C module.
//! \param bFast set up for fast data transfers.
//!
//! This function initializes operation of the I2C Master block by configuring
//! the bus speed for the master and enabling the I2C Master block.
//!
//! If the parameter \e bFast is \b true, then the master block is set up to
//! transfer data at 400 Kbps; otherwise, it is set up to transfer data at
//! 100 Kbps.  If Fast Mode Plus (1 Mbps) is desired, software should manually
//! write the I2CMTPR after calling this function.  For High Speed (3.4 Mbps)
//! mode, a specific command is used to switch to the faster clocks after the
//! initial communication with the slave is done at either 100 Kbps or
//! 400 Kbps.
//!
//! The peripheral clock is the same as the processor clock.  This value is
//! returned by SysCtlClockGet(), or it can be explicitly hard coded if it is
//! constant and known (to save the code/execution overhead of a call to
//! SysCtlClockGet()).
//!
//! \return None.
//
//*****************************************************************************
void
I2CMasterInitExpClk(uint32_t ui32Base, uint32_t ui32I2CClk,
                    bool bFast)
{

I referred to the datasheet to find out what I2CMTPR to use and want able to get the I2C driver to operate at 833MHz (faster than fast mode). I just opened the I2C driver at I2C_400kHz and then immediately set I2CMTPR. I was able to get the i2crf430cl330_load example to run with the code shown below.

#include <xdc/runtime/Types.h>
#include <inc/hw_types.h>
#include <inc/hw_i2c.h>
#include <ti/drivers/i2c/I2CTiva.h>
extern const I2CTiva_HWAttrs i2cTivaHWAttrs[];

/*
 *  ======== taskFxn ========
 *  Task for this function is created statically. See the project's .cfg file.
 */
Void nfcLoadTask(UArg arg0, UArg arg1)
{
    I2C_Handle          i2c;
    I2C_Params          i2cParams;
    Types_FreqHz        freq;

    /* Create I2C for usage */
    I2C_Params_init(&i2cParams);
    i2cParams.bitRate = I2C_400kHz;
    i2c = I2C_open(Board_I2C_NFC, &i2cParams);
    if (i2c == NULL) {
        System_abort("Error Initializing I2C\n");
    }
    else {
        System_printf("I2C Initialized!\n");
    }

    /* Change bitrate */
    BIOS_getCpuFreq(&freq);
    if (freq.lo == 120000000) {
    	System_printf("Setting MTPR\n");
    	HWREG(i2cTivaHWAttrs[Board_I2C_NFC].baseAddr + I2C_O_MTPR) = 0x05; //From the datasheet
    }
    else {
    	System_printf("Need to find the proper dividers for %d Hz\n", freq.lo);
    }

    System_flush();

Michael Vinokur said:

And another question (also asked in a different thread):

I have 2 tasks with same priority, each using different I2C channel in blocking mode.

Task #1 calls  I2C_Transfer and blocks.

What happens when task #2 calls I2C_Transfer immediately after task #1 ? Does it wait for task #1 to complete the transfer, or these two calls are run simultaneously, because 2 different I2C channels are used?

Each I2C controller "channel" has its own internal semaphore mutex for thread-safety, so in your case, Task 2 will not block on the 2nd I2C controller.