I2C (advanced) Read Issue

[Mark]

Hello,

In searching through this forum I found a reference to the advanced I2C functions having an issue with the I2CRead, specifically that the first read did not return correctly and that an I2C_SET_NO_ACK had to be issued after the first I2CRead and then another I2CRead to obtain the correct value. This was stated to be fixed but still looks to be broken. Will there be a proper fix to I2CRead?

IMG_4736.JPG (1017.24 KiB) Viewed 11901 times

Thanks,
Mark

[TomNB]

Hi Mark. Can you send a link to the post you are referring to?

[Mark]

http://forum.embeddedethernet.com/viewt ... w=previous

[Mark]

Code: Select all

{
	int iRtn;

	BYTE address = 0x22;		//mpc23008 address, U13
	BYTE addr_gpio = 0x09;		//gpio register address
	BYTE data_gpio = 0;			//variable to hold gpio read

	iRtn = I2CStart(address, I2C_START_WRITE, I2C_START_TIMEOUT);
	iprintf( "I2CStart = %d\r\n>", iRtn );

	iRtn = I2CSend(addr_gpio, I2C_RX_TX_TIMEOUT);
	iprintf( "I2CSend = %d,I2CRtn=  %d\r\n>", addr_gpio, iRtn );

	iRtn = I2CRestart(address, I2C_START_READ);
	iprintf( "I2CRestart = %d\r\n>", iRtn );

	iRtn = I2CRead(&data_gpio, I2C_RX_TX_TIMEOUT);
	iprintf( "I2CRead = %d,I2CRtn=  %d\r\n>", data_gpio, iRtn );
    iprintf("LED Value= %X\r\n>", data_gpio);

    //This is an issue that was supposed to be resolved (see NB forum and i2cmaster.h description)
	I2C_SET_NO_ACK;

	iRtn = I2CRead(&data_gpio, I2C_RX_TX_TIMEOUT);
	iprintf( "I2CRead = %d,I2CRtn=  %d\r\n>", data_gpio, iRtn );
    iprintf("LED Value= %X\r\n>", data_gpio);

	iRtn = I2CStop();
	printf( "I2CStop = %x\r\n>", iRtn );

	return data_gpio;
}

[TomNB]

Thanks Mark, but what I was really looking for was just a link to the thread so we can read the details.

[Mark]

I gave the link!

[TomNB]

Sorry, I scrolled to the last entry and saw your code post. Didn't realize you made two separate ones. What release are you using?

[Mark]

Release 2.7.5

[Mark]

BTW, the code snippet that I provided is my code that will help demonstrate the issue. If the remove the I2C_SET_NO_ACK along with the second read you will see that the first read provides incorrect data.

[Mark]

Any updates on this issue? Has it been verified/corrected?