How to connect FPGA and µC

That depends totally on the function that the FPGA will be performing, and also to some extent the micro that you're using!

More info please! Regards,

Look at how peripherals are connected to micro controllers, and consider using the same methods for your FPGA. The three popular methods are memory mapped, SPI (Serial Peripheral Interface, but you may see it under other names) bus, and I2C (Inter Integrated Circuit -- called "eye squared cee") bus.

Memory mapped you know of. It's fast, but it uses the most FPGA pins and quite a few FPGA resources. None the less the actual FPGA design is straightforward, although if the processor speed approaches that of the FPGA one can end up sweating timing details to a large degree.

SPI is a byte-oriented synchronous serial link. It uses three or four wires from the micro to the peripheral -- clock, data in, data out, and some sort of framing information which is often bit-banged in software. If the FPGA is set up as the slave it can be little more than an externally clocked shift register and a latch to let the FPGA know that the transfer is done. It's slower than memory mapped, and it only supports point-point links.

I2C is an invention of the devil -- er, Philips. It's this way-complex thing that allows two-wire communication between a micro and a number of peripheral chips on a board. It would probably require more space inside the FPGA than memory mapped, you'd tear your hair out making it work, and when you were done you couldn't call it "I2C" without paying Philips royalties (they trademarked the name, but if they patented it it's expired -- so some vendors make their stuff compatible, then figure out how to tell you so without saying "I2C").

also port mapped, or embedded (nios, microblaze)

-- Mike Treseler