With -pnr, leaf 0xc and leaf 0xd subleaf 0 report the same data as leaf 0xb subleaf 0. Leaf 0xd higher subleaves similarly look to mimic leaf 0xb. Is it perhaps meant to have a maximum basic leaf of 0xb instead?
Similar anomalies can be observed for leaf 0xc with -glm, -glp, -snr, and -tnt, albeit there the maximum basic leaf needs to remain higher than 0xb (i.e. leaf 0xc would better be all zero).
For -slm the actual cpuid.def has a maximum subleaf of 0xb, but specifies (apparently for no reason) data for leaves 0xc and 0xd as well.