Are all USB-2 devices compatible with USB-3 port?

@AliChen seems to be on the right path, actually this is a good line of thinking I hadn't considered. I never checked the motherboard specs when I posted my comment. This board uses a NEC (now Renesas) D720200F1 USB chipset for it's USB 3.0 ports, which seems to need a firmware update for full compatibility with published USB 3.0 standards, more information is available here. There is also information on TechPowerUp forums on more updated firmware information for this chipset as well.

Renesas does not offer the firmware update directly from what I can see, so it might take a little trial and error. Gigabyte seems to have no official update on their site that I could find either and this firmware is not part of your BIOS.

Remember that anytime you mess with firmware, especially in older hardware like this, there is a potential of something to fail or go wrong. Just giving you fair warning, do with this information what you will.

However the other answer of replacing the motherboard should be considered given the age and the relatively cheap availability of AM3/AM3+ motherboards that are comparable or superior to your existing board.

The Gigabyte 880GM-USB-3 motherboard is based on AMD 700-series chipset, which is nearly a 10 years-old design. The board likely uses a third-party USB 3.0 chip for USB 3 functionality. Ten years ago there were no functional USB 3.0 at all, and there were various odd ideas on how to blend USB 2.0 ports numbering into USB 3.0 port numbers, so massive confusion in architecture of drivers existed. My advice would be to get a modern board and do not waste your time debugging earlier prototypes of USB 3.0.

Actually, asking generically if USB2 devices are compatible with USB3 or not doesn't make much sense, because every USB3 port is the old USB 2.0 port with extra USB3 Tx and Rx pair of wires on top, literally. They are run by nearly totally separate controllers. The USB 3-style connect detection is supposed to happen first, which is based on presence of terminated links. The host transmitter is supposed to send special pulses to detect the line load. If load (USB3 device) is not detected, USB3 should shut off completely, and listen to pull-ups on old USB2 wires. Then follow the standard USB2 procedures. This sounds as fail-proof, but apparently older USB3 designs had implemented the process in somehow deficient way. A USB2 device has no idea about any extra wires or any connect detect procedures, and fires the old-style USB2 connect signaling right away, or with some delay, depending on device. But the old Renesas/NEC host controller apparently doesn't like the premature connect on USB2 lines, and apparently fails to properly report the connect event to host system registers.