@Kai-Heng Feng, the fix is not showing up in 4.4.0-72

carl@carl-5520:~$ uname -a
Linux carl-5520 4.4.0-72-generic #93-Ubuntu SMP Fri Mar 31 14:07:41 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux

carl@carl-5520:~$ sudo powertop


          Package   |             Core    |            CPU 0       CPU 4
POLL        0.0%    | POLL        0.0%    | POLL        0.0%    0.0 ms  0.0%    0.0 ms
C1          2.7%    | C1          2.2%    | C1          4.3%    1.3 ms  0.1%    0.3 ms
C2          7.3%    | C2          7.3%    | C2          7.0%    1.6 ms  7.7%    4.1 ms
C3         87.9%    | C3         88.4%    | C3         85.2%   14.8 ms 91.5%   23.1 ms

                    |             Core    |            CPU 1       CPU 5
                    | POLL        0.0%    | POLL        0.0%    0.0 ms  0.0%    0.0 ms
                    | C1          2.6%    | C1          0.0%    0.1 ms  5.2%    1.2 ms
                    | C2          4.4%    | C2          2.8%    1.9 ms  6.1%    1.1 ms
                    | C3         91.3%    | C3         96.0%   10.2 ms 86.6%    7.4 m

                    |             Core    |            CPU 2       CPU 6
                    | POLL        0.0%    | POLL        0.0%    0.0 ms  0.0%    0.0 ms
                    | C1          5.1%    | C1          8.8%    0.7 ms  1.4%    1.1 ms
                    | C2          8.4%    | C2         14.7%    2.5 ms  2.1%    1.5 ms
                    | C3         84.6%    | C3         73.7%    6.9 ms 95.4%   10.6 ms

                    |             Core    |            CPU 3       CPU 7
                    | POLL        0.0%    | POLL        0.0%    0.0 ms  0.0%    0.0 ms
                    | C1          0.9%    | C1          0.5%    0.2 ms  1.3%    3.3 ms
                    | C2          8.9%    | C2         13.6%    1.6 ms  4.2%    2.5 ms
                    | C3         87.1%    | C3         80.3%    6.8 ms 93.9%   10.8 ms


carl@carl-5520:~$ sudo i7z

Cpu speed from cpuinfo 2902.00Mhz
cpuinfo might be wrong if cpufreq is enabled. To guess correctly try estimating via tsc
Linux's inbuilt cpu_khz code emulated now
True Frequency (without accounting Turbo) 2902 MHz
  CPU Multiplier 29x || Bus clock frequency (BCLK) 100.07 MHz

Socket [0] - [physical cores=4, logical cores=8, max online cores ever=4]
  TURBO ENABLED on 4 Cores, Hyper Threading ON
  Max Frequency without considering Turbo 3002.07 MHz (100.07 x [30])
  Max TURBO Multiplier (if Enabled) with 1/2/3/4 Cores is  39x/37x/36x/35x
  Real Current Frequency 1395.28 MHz [100.07 x 13.94] (Max of below)
        Core [core-id]  :Actual Freq (Mult.)      C0%   Halt(C1)%  C3 %   C6 %  Temp      VCore
        Core 1 [0]:       1351.45 (13.51x)      5.21    97.6       0       0    44      0.8995
        Core 2 [1]:       1395.28 (13.94x)      1.37    99.3       0       0    43      0.8981
        Core 3 [2]:       1310.44 (13.10x)      2.49    98.9       0       0    44      0.8981
        Core 4 [3]:       1328.40 (13.27x)         1    99.6       0       0    44      0.8954



C0 = Processor running without halting
C1 = Processor running with halts (States >C0 are power saver modes with cores idling)
C3 = Cores running with PLL turned off and core cache turned off
C6, C7 = Everything in C3 + core state saved to last level cache, C7 is deeper than C6
  Above values in table are in percentage over the last 1 sec
[core-id] refers to core-id number in /proc/cpuinfo
'Garbage Values' message printed when garbage values are read
  Ctrl+C to exit