Add ST_E stage: serialize t_hat[0..1] -> ek_mem[0..767], s_hat[0..1] ->
dkp_mem[0..767] via byteEncode12 (2 coeffs -> 3 bytes, LSB-first 12-bit:
b0=c0[7:0], b1={c1[3:0],c0[11:8]}, b2=c1[11:4]), then copy rho into
ek_mem[768..799]. Byte readback tap (dbg_byte_sel/idx -> dbg_byte_o).
Verified vs KAT-derived golden: ek 800B (== KAT pk) + dk_pke 768B
(== KAT sk prefix) byte-exact (20430 cyc). Completes Stage 2 datapath.