I'm attempting to enable DCI on a NUC kit -- specifically, a NUC7i7BNH. I've enabled the "Silicon Debug Features" in the BIOS, which appears to have enabled DCI on the CPU: MSR 0xC80 (IA32_DEBUG_INTERFACE) reads 0xC0000001. However, it seems that it is not enabled in the chipset: the ECTRL register of the DCI PCR (0xFDB80004) reads 0x00000001, which does not have bit 4 set (page 1706 of https://www.intel.com/content/dam/www/public/us/en/documents/datasheets/100-series-chipset-datasheet-vol-2.pdf).
That manual indicates that HDCIEN can only be reset to a 1 by asserting RTCRST#. I've tried disassembling the NUC and removing the CMOS battery, but to no avail -- the BIOS detects the CMOS failure and notifies me on screen, but by the time I can run an EFI shell again to check the bit, it's back to 0. The way I see it, there are a few possibilities here (in order of good to bad):
- There's some code in the BIOS that writes a 0 to the bit.
- I've looked through it a bunch in IDA, but I didn't see anything obvious (i.e. no literal references to 0xfdb80004).
- In any case, if it is there, I should be able to patch it out -- boot guard is not enabled.
- Removing the CMOS battery did not actually assert RTCRST#.
- If this is the case, how can I forcefully assert RTCRST#? Is there a jumper somewhere on the board? I took everything off, but I didn't see one.
- The PCH does not allow this bit to be set for some reason (a blown fuse perhaps?)
Has anyone gotten DCI working on these NUC kits before? Does anyone have any suggestions for debugging this further, or contact information for someone who might?
Any help would be greatly appreciated.