mlkem test diagnostics and allocator updates

rtl/ip/fft/FFT1024_32B.v

@@ -1,7 +1,7 @@
 //----------------------------------------------------------------------
 //  FFT: 1024-Point FFT Using Radix-2^2 Single-Path Delay Feedback
 //----------------------------------------------------------------------
-(* use_dsp = "no" *)
+
 module FFT #(
     parameter   WIDTH = 32
 )(

rtl/ip/fft/Multiply.v

@@ -1,7 +1,7 @@
 //----------------------------------------------------------------------
 //  Multiply: Complex Multiplier
 //----------------------------------------------------------------------
-(* use_dsp = "no" *)
+
 module Multiply #(
     parameter   WIDTH = 16
 )(
@@ -13,7 +13,7 @@ module Multiply #(
     output  signed  [WIDTH-1:0] m_im
 );
 
-(* use_dsp = "no" *) wire signed [WIDTH*2-1:0]   arbr, arbi, aibr, aibi;
+ wire signed [WIDTH*2-1:0]   arbr, arbi, aibr, aibi;
 wire signed [WIDTH-1:0]     sc_arbr, sc_arbi, sc_aibr, sc_aibi;
 
 //  Signed Multiplication

rtl/ip/fft/SdfUnit.v

@@ -1,7 +1,7 @@
 //----------------------------------------------------------------------
 //  SdfUnit: Radix-2^2 Single-Path Delay Feedback Unit for N-Point FFT
 //----------------------------------------------------------------------
-(* use_dsp = "no" *)
+
 module SdfUnit #(
     parameter   N = 64,     //  Number of FFT Point
     parameter   M = 64,     //  Twiddle Resolution
@@ -86,7 +86,7 @@ reg             bf2_do_en;  //  2nd Butterfly Output Data Enable
 //  Multiplication
 wire[1:0]       tw_sel;     //  Twiddle Select (2n/n/3n)
 wire[LOG_N-3:0] tw_num;     //  Twiddle Number (n)
-(* use_dsp = "no" *) wire[LOG_N-1:0] tw_addr;    //  Twiddle Table Address
+ wire[LOG_N-1:0] tw_addr;    //  Twiddle Table Address
 wire[WIDTH-1:0] tw_re;      //  Twiddle Factor (Real)
 wire[WIDTH-1:0] tw_im;      //  Twiddle Factor (Imag)
 reg             mu_en;      //  Multiplication Enable

rtl/ip/fft/axi_fft_wrapper.v

@@ -7,7 +7,7 @@
 (* use_dsp = "no" *)
 module fft_controller #(
     parameter FFT_LEN         = 1024, // FFT处理长度
-    parameter BRAM_RD_LATENCY = 2,    // 适配 BRAM 的读延迟 (通常 Block RAM 开启输出寄存器为 2)
+    parameter BRAM_RD_LATENCY = 1,    // 适配 BRAM 的读延迟 (通常 Block RAM 开启输出寄存器为 1)
     parameter FFT_DATA_WIDTH  = 16
 )(
     input               aclk,