Well, 2022 was a long and crazy year for me (health wise as well) but I’m trying to get back on track and focus and concentrate on new things in life! I’ve been finding that the older I get, the more I think about the olden golden days of being a kid and growing up in an analogue world, before the invention of the internet and computers. Times were simpler but more basic and it’s kinda crazy to think about how much progress and change exists nowadays. I remember a time when my specific search term in Google actually showed me the exact thing I was looking for haha 🙂 Anyway, I’ll try to keep this blog going as a personal hobby of interesting things that I find out along the way in life. Just random tech craziness if you will!

~

So, I was searching around for the Linux equivalent of the BSD PF firewall rule “scrub-all no-df” and I was expecting that there would be an “iptables mangle prerouting” version – but – no luck! I just wanted to be able to clear this specific bit in the IPv4 packet header with a Linux router so that it would be able to fragment the packets along the inner routes that may have a lower MTU set. Since I couldn’t find anything I made one myself 🙂

Source: https://github.com/stoops/nf_df

Code Snippet

Test Command

ping -c 1 -M do -s 123 1.1.1.1

Packet Capture

As you can see below, the first IPv4 packet (unmodified) contains the DF bit flag set on (0x0000: …. 4000 ….) compared to the modified packet below in the “After Send” section.

Before Send

21:58:39.473143 fe:00:00:00:01:02 > fe:00:00:00:01:01, ethertype IPv4 (0x0800), 
length 165: 1.3.3.7 > 1.1.1.1: ICMP echo request, id 19298, seq 1, length 131
	0x0000:  4500 0097 0000 4000 4001 6348 0103 0307  E.....@.@.cH....
	0x0010:  0101 0101 0800 6735 4b62 0001 0f74 6d63  ......g5Kb...tmc
	0x0020:  0000 0000 2538 0700 0000 0000 1011 1213  ....%8..........
	0x0030:  1415 1617 1819 1a1b 1c1d 1e1f 2021 2223  .............!"#
	0x0040:  2425 2627 2829 2a2b 2c2d 2e2f 3031 3233  $%&'()*+,-./0123
	0x0050:  3435 3637 3839 3a3b 3c3d 3e3f 4041 4243  456789:;<=>?@ABC
	0x0060:  4445 4647 4849 4a4b 4c4d 4e4f 5051 5253  DEFGHIJKLMNOPQRS
	0x0070:  5455 5657 5859 5a5b 5c5d 5e5f 6061 6263  TUVWXYZ[\]^_`abc
	0x0080:  6465 6667 6869 6a6b 6c6d 6e6f 7071 7273  defghijklmnopqrs
	0x0090:  7475 7677 7879 7a                        tuvwxyz

Before Recv

21:58:39.475300 fe:00:00:00:01:01 > fe:00:00:00:01:02, ethertype IPv4 (0x0800), 
length 165: 1.1.1.1 > 1.3.3.7: ICMP echo reply, id 19298, seq 1, length 131
	0x0000:  4500 0097 d279 0000 3901 d7ce 0101 0101  E....y..9.......
	0x0010:  0103 0307 0000 6f35 4b62 0001 0f74 6d63  ......o5Kb...tmc
	0x0020:  0000 0000 2538 0700 0000 0000 1011 1213  ....%8..........
	0x0030:  1415 1617 1819 1a1b 1c1d 1e1f 2021 2223  .............!"#
	0x0040:  2425 2627 2829 2a2b 2c2d 2e2f 3031 3233  $%&'()*+,-./0123
	0x0050:  3435 3637 3839 3a3b 3c3d 3e3f 4041 4243  456789:;<=>?@ABC
	0x0060:  4445 4647 4849 4a4b 4c4d 4e4f 5051 5253  DEFGHIJKLMNOPQRS
	0x0070:  5455 5657 5859 5a5b 5c5d 5e5f 6061 6263  TUVWXYZ[\]^_`abc
	0x0080:  6465 6667 6869 6a6b 6c6d 6e6f 7071 7273  defghijklmnopqrs
	0x0090:  7475 7677 7879 7a                        tuvwxyz

Kernel Module

# lsmod | grep -i nf_df
nf_df    16384    0

After Send

21:52:52.646803 fe:00:00:00:01:02 > fe:00:00:00:01:01, ethertype IPv4 (0x0800), 
length 165: 1.3.3.7 > 1.1.1.1: ICMP echo request, id 52199, seq 1, length 131
	0x0000:  4500 0097 0000 0000 4001 a348 0103 0307  E.......@..H....
	0x0010:  0101 0101 0800 0e11 cbe7 0001 b472 6d63  .............rmc
	0x0020:  0000 0000 56d8 0900 0000 0000 1011 1213  ....V...........
	0x0030:  1415 1617 1819 1a1b 1c1d 1e1f 2021 2223  .............!"#
	0x0040:  2425 2627 2829 2a2b 2c2d 2e2f 3031 3233  $%&'()*+,-./0123
	0x0050:  3435 3637 3839 3a3b 3c3d 3e3f 4041 4243  456789:;<=>?@ABC
	0x0060:  4445 4647 4849 4a4b 4c4d 4e4f 5051 5253  DEFGHIJKLMNOPQRS
	0x0070:  5455 5657 5859 5a5b 5c5d 5e5f 6061 6263  TUVWXYZ[\]^_`abc
	0x0080:  6465 6667 6869 6a6b 6c6d 6e6f 7071 7273  defghijklmnopqrs
	0x0090:  7475 7677 7879 7a                        tuvwxyz

After Recv

21:52:52.669611 fe:00:00:00:01:01 > fe:00:00:00:01:02, ethertype IPv4 (0x0800), 
length 165: 1.1.1.1 > 1.3.3.7: ICMP echo reply, id 52199, seq 1, length 131
	0x0000:  4500 0097 79f9 0000 3901 304f 0101 0101  E...y...9.0O....
	0x0010:  0103 0307 0000 1611 cbe7 0001 b472 6d63  .............rmc
	0x0020:  0000 0000 56d8 0900 0000 0000 1011 1213  ....V...........
	0x0030:  1415 1617 1819 1a1b 1c1d 1e1f 2021 2223  .............!"#
	0x0040:  2425 2627 2829 2a2b 2c2d 2e2f 3031 3233  $%&'()*+,-./0123
	0x0050:  3435 3637 3839 3a3b 3c3d 3e3f 4041 4243  456789:;<=>?@ABC
	0x0060:  4445 4647 4849 4a4b 4c4d 4e4f 5051 5253  DEFGHIJKLMNOPQRS
	0x0070:  5455 5657 5859 5a5b 5c5d 5e5f 6061 6263  TUVWXYZ[\]^_`abc
	0x0080:  6465 6667 6869 6a6b 6c6d 6e6f 7071 7273  defghijklmnopqrs
	0x0090:  7475 7677 7879 7a                        tuvwxyz

Edit: In addition to this Linux kernel nftables module, I found out that there is another one running for IPv4 packet defragmentation which should also be similar to the scrub rule “fragment reassemble”. According to the documentation, if you have any stateful connection tracking rules for layer-4 (for example, postrouting nat rules or prerouting udp ports) then the fragmented packets should get reassembled again. I am looking into the scrub rule “reassemble tcp“…

# lsmod | grep -i defrag
nf_defrag_ipv4    16384    1    nf_conntrack

~

So I had this Mac Mini that I was originally using as a WiFi relay bridge (layer 2) but since then I was only using it as an rsync backup server. I decided to also turn it into a VPN tunnelling router (layer 3) for my home network setup by using OpenVPN on MacOS (connected to a Debian Linux server). I was testing out its performance and it was holding up pretty good and stable but I wanted to give WireGuard a try as well, just to see and compare. Upon initial configuration, I noticed that WireGuard was defaulting to a lower MTU (1420) compared to what I had set in OpenVPN (1450) and that some of my connections to websites were unstable/slow/hanging. Additionally, with OpenVPN, I was able to set some extra settings like MSS & fragmentation limits and I couldn’t find the equivalent of those with WireGuard. However, I do like the overall simplicity of the WG config and setup process!

I actually lowered the WG MTU (1410) but this could potentially cause larger size packets to fragment. I was searching around and found that it was possible for one to clamp the MSS values on forwarded/routed packets with an iptables forward/mangle rule. Since the Mac is the VPN client and has the BSD PF firewall, I also set a scrub rule as well:

# bsd pf
scrub all no-df random-id max-mss 1330
 
# nix nf
iptables -t mangle -A POSTROUTING -p tcp --tcp-flags SYN,RST SYN -j TCPMSS --set-mss 1330
 
# iptables-translate ftw!

I was also trying to read more about the “don’t fragment” bit that could get set on packets along the way (sometimes when the PMTUD setting is enabled) and the pf-scrub rule can help to clear that bit (no-df) where I couldn’t yet find the equivalent on the Linux side. I also tried setting these sysctl values as well:

# mac
sysctl net.inet.tcp.path_mtu_discovery=0
 
# nix
sysctl net.ipv4.ip_no_pmtu_disc=1

I’m still playing around with this setup but it’s been an interesting networking experiment so far in terms of seeing how your web connections handle and react to being encapsulated automatically by a network router…

Edit: I have written a Linux kernel module to do the equivalent of “scrub-all no-df“

Tunnelled Internet – 1G WAN – Speed Test

WireGuard:

~

Well it’s been a tough number of years for me as all of my personal belongings have been packed up in boxes the entire time, however, I have finally been able to move into a new space and a new place.

As a new decorative piece to hang on the wall, I was able to get a large size, glass print of the MacOS flying Apple desktop background wallpaper I created from Fracture Me. It helps to add a little more fun and colour inside and they have pretty good quality printers (quick service as well).

This will be a new journey for me!

~

So after many years of using and learning iptables, one of my favourite firewalls, I had to translate my command line ruleset into a new format/syntax. It wasn’t too bad but with the added power and flexibility of nftables, it can be harder to find the order/priority of the rules for a given filter hook, for example, INPUT. However, I really appreciate how the new firewall incorporates some previous features that I used to have to install via iptables modules like ipset and hashlimit, which provides the rate limiting functionality!

hook="input"
nft list ruleset                               \
  | sed -e 's/counter.*[0-9]//g'               \
  | tr '\t' ' ' | tr '\r\n' '~' | tr '{}' '\n' \
  | grep -i "hook.${hook}"                     \
  | tr '~' '\n'                                \
  | grep -i '[a-z]'                            \
  | sed -e 's/^[ ]*//'

type filter hook input priority filter; policy accept;
iifname "lo"  accept
iifname "lan" udp dport 67  accept
iifname "wan" ct state established,related  accept
drop

So on OpenWRT for example, a WAN interface can be set get an IP address via DHCP, however, this will launch a separate independent process on Linux to act as the client. On the web interface, there is a button on each network interface which allows you to restart that individual interface (and in addition, it will be smart enough to kill the previous client process and launch a new one as well as re-initialize the firewall and other stuff). I could not find the equivalent command line option to this as a simple ifdown/sleep/ifup would not capture all of these sub-actions as per the extra configurations set on the interface. So I searched and read a helpful hint that can give a similar workaround using the network init startup script. You can modify some small part of the network interface configuration and then call network reload which will be smart enough to pick up the diff/changes only. This helps greatly in case you need to script or schedule a specific interface restart:

#!/bin/sh
r=$(/bin/date | /usr/bin/md5sum | /usr/bin/cut -c 1-8)
/sbin/uci set network.wan.hostname="owrt${r}"
/sbin/uci commit network
/etc/init.d/network reload

No usage documentation, however, posting this for historical purposes:

#ok:www.google.com:false
python3 proxy.py ~/.proxy.txt host:port
export https_proxy='http://127.0.0.1:9090'

import os, sys, re
import select, socket, socketserver


class proxysrv(socketserver.BaseRequestHandler):
    def init(self):
        self.size = (8 * 1024)
        self.file = sys.argv[1]
        self.info = sys.argv[2].split(":")
        self.host = (self.info[0], int(self.info[1]), True)
        self.sock = self.request

    def config(self, file):
        proxlist = []
        try:
            fobj = open(file, "r")
            for line in fobj.readlines():
                line = line.strip().split(":")
                stat = False
                if ("true" in line[2].lower()):
                    stat = True
                proxlist.append([line[0], line[1], stat])
        except:
            print("error","config")
        return proxlist

    def checks(self, conf, data):
        try:
            proxyreq = data.decode().split("\n")
            hostreqs = proxyreq[0].strip().split(" ")
            hostinfo = hostreqs[1].split(":")
            hostname = hostinfo[0]
            hostport = int(hostinfo[1])
            lasthost = ()
            for item in conf:
                regx = re.match("^.*%s$" % (item[1]), hostname)
                if ((item[0] == "no") and (not regx)):
                    lasthost = (hostname, hostport, item[2])
                if ((item[0] == "ok") and regx):
                    lasthost = (hostname, hostport, item[2])
            if (lasthost and (not lasthost[2])):
                return lasthost
        except:
            print("error","checks")
        return self.host

    def newcon(self, srcs, data, host):
        try:
            senddata = "HTTP/1.1 200 Connection established\r\nProxy-Connection: Keep-Alive\r\n\r\n".encode()
            sockobjc = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
            sockobjc.connect((host[0], host[1]))
            if (host[2]):
                sockobjc.sendall(data)
            else:
                srcs.sendall(senddata)
            return sockobjc
        except:
            print("error","newcon")
        return None

    def relays(self, srcs, dsts):
        try:
            while True:
                (readlist, _, _) = select.select([srcs, dsts], [], [])
                for sockitem in readlist:
                    if (sockitem == dsts):
                        data = dsts.recv(self.size)
                        print("dsts",len(data))
                        if (not data):
                            raise
                        srcs.sendall(data)
                    if (sockitem == srcs):
                        data = srcs.recv(self.size)
                        print("srcs",len(data))
                        if (not data):
                            raise
                        dsts.sendall(data)
        except:
            pass

    def handle(self):
        self.init()

        proxlist = self.config(self.file)
        recvdata = self.sock.recv(self.size)
        desthost = self.checks(proxlist, recvdata)
        sockobjc = self.newcon(self.sock, recvdata, desthost)

        print("addr",self.client_address)
        print("data",recvdata)
        print("host",desthost)

        if (sockobjc):
            self.relays(self.sock, sockobjc)


if (__name__ == "__main__"):
    for e in ["http_proxy", "https_proxy", "no_proxy"]:
        for v in [e.upper(), e.lower()]:
            print("envr",v)
            os.environ[v] = ""

    sockobjc = socketserver.TCPServer(("127.0.0.1", 9090), proxysrv)
    sockobjc.serve_forever()

So there was a Reddit thread today that was diverging into bad coding tests plus grade evaluation puns (C++) – so I tried to make a purposefully confusing C program behave in a way that a coder or programmer wouldn’t exactly expect. After I wrote it, I had to put in some for-loop-printf-debug statements just to confirm and demonstrate how the right-to-left side value-lookup to delayed-increase to index-assignment works in C, it’s pretty neat to try and analyze how and why this is “working” lol!

#include <stdio.h>
int main()
{
    char A = 'A', B = 'B', C = 'C', D = 'D', E = 'E', F = 'F';
    char grades[256];
    grades[A++] = A++;
    grades[B++] = B++;
    grades[C++] = C++;
    grades[D++] = D++;
    grades[E++] = E++;
    grades[F++] = F++;
    printf("My grade is %c!\n",grades[C++]);
}

And basically this will print out “My grade is D!” – but it’s fun to trace through how each line will process the value statement first and then the assignment index is processed afterwards. In addition, each line will affect the next statement below because of the increases that take place in between processing sides, so the future lookup index values are being changed along the way! 🙂

So, there was one more app I needed to complete my personal trilogy, and it’s named BrowserBot! It’s an app that will help “curl” the web for any page you’d like to monitor for changes, and you can feed it custom commands (similar to ClippyCommand) so it will parse out any field/value you specify. I had scripts running in a crontab to do this but I wanted a more centralized and official place to quickly view the latest and refreshed web data that I was interested in tracking for security reasons!

Source Code: github.com/stoops/BrowserBot

This is meant to be like a rolling index to keep track of the MacOS apps I have created thus far:

	AppleApplications
	BrowserBot	fossjon/2022/07/16/browserbot-…	(Window)
	ClippyCommand	fossjon/2022/07/11/clippycommand-…	(Menubar)
	DynaDock	fossjon/2021/12/26/dynadock-…	(Dock)
	TurnTable	fossjon/2025/06/18/turntables-…	(App)
Bonus:
		Widgets Factory	(iOS App)

~

So as a Mac/Linux/BSD fan, I’ve posted a few different Microsoft related creations/customizations that I use as a daily reminder of the good old days! Even though I wasn’t the biggest fan of Windows and Explorer, I still tried to make some dock icons to appreciate and represent both of them in the future. I was just missing one last memory of mine, which was, Clippy! I created a new MacOS app that allows you to pipe the copy-paste buffer into a set of your own customized bash shell command line programs and it will rewrite the paste buffer so you are ready to go! This is the second custom app in my dock, along side my DynaDock application.

Source Code: github.com/stoops/ClippyCommand

Some of my more useful and commonly used editing/formatting commands!

So I recently ordered a Drop Alt keyboard with Halo True switches as it offered the ability to hot-swap them out. The case and frame is much more stable and solid compared to the Ducky Mini and it produces a much more clunky-analogue, type-writer kind of sound. I put in some Holy Panda switches and they have a greater feeling tactile bump at the start of the key travel along with a slightly lighter spring force for an easier press downwards. Both are great switches to type on overall!

Edit: For comparison, the Cherry Brown switch has a tactile resistance during the key travel, but the bump in the beginning is not as noticeable or pronounced as to what some of the other switches offer and the spring force is much lighter.

Update: The Three Tactiles for 2023

The Logitech G500 (Gaming) – A long time ago, this mouse was one of the most comfortable and capable performing ones on the market!

The Logitech MX3 (Mac) – Still one of the best feeling mice in the hand. A simple design and clean layout along with free-wheel-scrolling continued plus thumb-ledge + slightly-angled + softer-feeling!

Main Keyboard Selection

Drop-Alt+Holy-Pandas!

~

Update: fossjon/last-keyboard-mod-before-the-end-of-the-year

Intro: So, I know I’m a little late to the game in terms of getting a mechanical keyboard, however, I thought it would be a good investment to have a quality one just in case while WFH. I picked up one of those Ducky One-2-Mini keyboards in a white color and it offers some great built-in-hardware features. There are some random color-matching keys that come with it which can make it look a little bit more vintage / retro / classic. It’s not the lowest-profile one but the PBT keycaps are great quality and hug your fingers!

Pros:

• Programmable and saveable profiles
• Swappable key mappings for caps/ctrl/alt/cmd/fn
• Individually lit and bright RGB color patterns for every key
• Cherry MX Brown Switches with a serious sound and tactile travel
• Solid key stabilizers and firm case frame
• Lays flat plus two levels of height adjustment all with rubber feet
• USB-C detachable-cable + firmware updates

Mods:

• Remapped dedicated arrow keys, removed one stab set
• Added some extra PBT keycap colors
• Added one O-Ring to the space-bar keycap stem
• Removed the top case cover piece
• Replaced the bottom case feet with more medium-sized felt-pads
• Taped small pieces cut out from the keyboard box inside the bottom case
• Purchased some M3-Sized self-tapping screws for more secure plate mounting

Cons:

• Not filled with foam inside the keyboard case
• Not hot-swappable switches
• Not screw-in stabs

Mod-Sound(Impact): I placed a single O-Ring on the space-bar and it sounds much less hollow and more gentle, similar to a regular key press. I was able to replace the bottom rubber feet with felt pads and the resulting feel and sound is much softer and warmer. I also taped some thin pieces of cardboard cut out from the keyboard box to the inside of case so that the key presses sound a bit deeper and quieter as they are absorbed and dampened.

Mod-Sound(Airborne): I removed the top cover piece which greatly reduces and lessens the amount of sound wave interference that bounces around down within the inside of the case. From the mounting plate upwards, the top portion of the switch and keycap now sit at the highest point on the keyboard. You can hear a little spring ping if you listen closely along with slightly rougher switch noise, but the sound can now directly travel outwards and away from the keys and case, unobstructed. Each key press has a greater distinction and clarity to it, thus making it a much quieter and smoother typing experience.

Mod-Visual(Keycaps): So after spending some time with this little keyboard, I was able to remap dedicated arrow keys to be on the right side while keeping the main modifier keys on the left side (I also remapped caps lock key to be a backup fn key, as this setup is geared more towards coding on the mac). I purchased some extra keycap colors although it would be nice to find some properly wide arrow keys in the future!

Finale: It’s a pretty good MK overall esp for just a stock unit. I wish I could get one of the ones like Taeha makes as they sound soo super smooth and clean and crisp — mechanical contact points only. He’s able to lube all of the sub-components like springs, switches, stabilizers and basically eliminate or reduce most of the noises, thus creating some purrre asmr — buttery sounding — Holy Pandas! (: