[Regression][FIRMWARE][Bisected] Navi10 power limits stopped working due to firmware change

The changes to the navi10_smu firmware in this upstream commit af76fd0ed266440ac406d5737218af7ac3cfc750 in linux-firmware caused the power limit setting via sysfs/hwmon to break.

While the value is reflected in sysfs, the behavior of the card no longer changes, and it follows its default power limit profile.

Rolling back to the initially released navi10 firmware for the SMU, from 417a9c6e197a8d3eec792494efc87a2b42f76324 in linux-firmware resolves the problem, and the power limit is followed correctly when changed.

I know this might not be the spot for this issue since it's a firmware bug, but it would be nice to get this resolved, so let me know who I need to talk to to either get it resolved on the firmware side, or update the driver code I wrote for the new firmware spec.