Hi,

I'm trying to rationalize some very common I2C code and came across the following:

Every controller except cpu/mpc8xx implements functions like this:

---

uchar i2c_reg_read(uchar i2c_addr, uchar reg) { uchar buf;

i2c_read(i2c_addr, reg, 1, &buf, 1); return buf; }

void i2c_reg_write(uchar i2c_addr, uchar reg, uchar val) { i2c_write(i2c_addr, reg, 1, &val, 1); }

---

cpu/mpc8xx, however, adds i2c_init() calls:

---

uchar i2c_reg_read(uchar i2c_addr, uchar reg) { uchar buf;

i2c_init(CFG_I2C_SPEED, CFG_I2C_SLAVE); i2c_read(i2c_addr, reg, 1, &buf, 1); return (buf); }

void i2c_reg_write(uchar i2c_addr, uchar reg, uchar val) { i2c_init(CFG_I2C_SPEED, CFG_I2C_SLAVE); i2c_write(i2c_addr, reg, 1, &val, 1); }

---

Can somebody lend historical perspective as to why these i2c_init() calls are needed?

thanks, Ben