Thank you for this report. We are investigating the problem, since there definitely should not be any Comport listed for this board.
I reinstalled Win7 yesterday... the COM-port is still existing (now it's called COM1 instead of COM3).
I booted a Mint Linux 12 Live CD and checked the IRQ table plus lshw. There is no COM-port detected by Linux.
Here are two more pictures taken from the Device Manager. As you can see, the "fake" COM-port even has got an IRQ asigned (IRQ 4). Also the COM-port is connected to the Intel LPC Controller.
I flashed the new BIOS version (0037), but the COM-port is still existing. Will there be any solution by the next BIOS release?
Yes, the plan is to roll the fix for this into the next production BIOS release. However, I do not have an estimated release date for it yet.
I installed the new 0039 revision... the "ghost" serial port is still there When will Intel fix this bug?
Further I encountered a new bug after installing the 0039 BIOS: I use "low voltage" DDR3-SDRAM modules (I own 2x2GB Samsung and 2x2GB Crucial 1.35v memory kits). With revision 0037 and below, I was able to use my memory with 1.35v. Although the DH61AG was not able to detect the correct voltage, I could change the value manually within the BIOS. Directly after updating the BIOS to revision 0039, the DH61AG failed to boot. First I thought, the BIOS update process failed. But I recognised the power LED blinking as an error code referring to the memory. I could revive the DH61AG by changing the jumper to maintance mode and change the SDRAM voltage to 1.50v.
After some trial and error, I can say: When I change the voltage manually to 1.35v (which is the optimal voltage for low voltage RAM), the DH61AG fails to boot (with revision 0039). This sucks...
This is my 4th Intel mainboard (DH67CF, DG45FC, DG33TL). But this is my first Intel mainboard with such serious bugs.