That's how every enthusiast board I've seen does it. It drives PWM duty against temperature with no concern for fan RPM. You can usually control the curve in the BIOS. You get like 5 or so adjustment points where you can specify the minimum idle duty (regardless of how cold the CPU is, a few points linking a certain duty to a certain temp, and then a temp at which it should run 100%. A few I've seen will let you enable a minimum fan RPM warning and set that RPM if you want.Are you saying that the motherboard is simply going to continue increasing fan speed based on temp with no care of RPM / tricked RPM?
Even the Intel Server board I have doesn't run closed loop feedback on RPM. It differs from an enthusiast board and uses a closed loop system, but it's temperature based. It will simply ramp the fan's duty to whatever is necessary for a given cooling domain to maintain a programmed thermal margin. I "hacked" the .SDR file for my Intel board to make it hold a lot more thermal margin since I didn't think ~80C was a good operating temp for my CPUs when fully loaded. The Intel .SDR does define some characteristics for the fans, which include min and max RPM, so I guess if you changed the fans or faked the fan RPM on the Intel board it could detect a fault if the RPM went outside the expected range, but I don't see how the Supermicro board could have a similar mechanism.
Regardless, I can always change the RPM scaling factor being applied to the fans or make the factor non linear based on the fan RPM if I need to further outsmart Supermicro's inflexible fan control. It's a microcontroller running software and I can program it to do anything I want (as long as the code fits in the flash). Multiples of 2 are just really easy because you simply do a bitwise shift of the measured value to scale it.