How is that possible? A local hardware I2C interface makes life a lot easier. Some microcontrollers even have 2 or more and SoCs have many (6 or more isn't unheard of).
Dunno, but - thinking about what I usually use
- m683xx, has 1xSPI line, no-i2c -> i2c is implemented by two TPU-pins for bit-banging
- m68hc11 has 1xSPI line, no-i2c -> i2c is implemented by two PORT-pins for bits-banging
- the Ath5k (Linux/MIPS SoC) used in Fonera2 has 1xSPI line, no-i2c, -> i2c by SPI-I2C bridge
- the Ath9k (Linux/MIPS SoC) used in the RSP has 1xSPI line, no-i2c, -> i2c by SPI-I2C bridge
- the Ath9k (Linux/MIPS SoC) used in the WL703 has 1xSPI line, no-i2c, -> i2c by SPI-I2C bridge
- my fpga SoftCore uses 3 digital pins to implement SPI -> i2c is implemented by SPI-I2C bridge
I think ... I am using weird things