Podman

w0d0

Logistics

Announcements

  • Welcome to Systems in C

    • You first semester in a compiled language - more on that soon.

  • Action Items:

    • Access the course webpage
    • Join the Discord!
      • Working on this.
    • Spin up a GitHub acct if you don’t have one.

Homework

  • The first homework, “Enigma”, is ready after this class.
    • Due Thursday, 03 Sep at midnight AOE
    • The infamous CS1 hw, now in a good language!
    • Mostly just get the toolchain working.

Enigma

→

Today

  • Introduce the course
    • The C Programming Language
    • Cryptography + Cryptocurrency
  • Introduce the technologies
    • podman
    • vim
    • gcc
    • git

Pre-reqs

  • These things:
    • podman
    • vim
    • gcc
    • git
  • Are commands at the command line
  • If you are afraid of the command line, drop this class.

It me

  • Name
    • Calvin (Deutschbein)
  • Call me
    • (Professor) Calvin
  • Pronouns
    • pronouns or big frowns - me
    • they/them

The Great Work

  • Thesis Title
    • Mining Secure Behavior of Hardware Designs
  • Plain English
    • Just as there are bugs in code that makes software, modern hardware is also written in code and therefore may contain bugs. I find these bugs.

Learning Objectives

How To Do It

  • After ~10 years of systems research I’ve convinced myself only two things that really matter:

    • Pointers, and
    • Recursion

  • Recursion isn’t too bad…

>>> exp = lambda b, n : b if n == 1 else b*exp(b,n-1)
>>> exp(2,8)
256

Pointers

  • Pointers are a beast, but without them nothing makes sense!
>>> x = 1
>>> def addx():
...     x += 1
...
>>> addx()
UnboundLocalError:
<blah blah blah error messages>
>>> x = [1]
>>> def addx():
...     x[0] += 1
...
>>> addx()
>>> x
[2]

Insight

  • Definition:
    • Pointers: Variables that store memory addresses.
    • Recursion: Functions calling themselves to solve sub-problems.
  • Importance:
    • Core to efficient algorithms and memory management in low-level programming.

Python and Pointers

  • Why not .py (.js, .java, .cs, .cpp, etc)?
    • These languages abstract memory.
    • This abstraction makes computation unclear
    • This lack of clarity:
      • Adversely impacts education
      • Complicates low-level design
      • Leads to low performance
    • Good languages (except Java) but not for us

C and Pointers

  • Why C?
    • Everyone hates C because it makes us think…
    • And thinking is hard.
    • But there exist problems (security) for which careful thought is necessary.
    • Learning C makes us better Blub programmers.
  • I claim: You needn’t use C in real life.

Kennedy is my middle name

Not an endorsement; I voted for ******* **** ****** ********!

C is for Clear

  • C does not obscure what a computer is doing.
  • This lack of obscurity allows for greater precision.
  • This greater precision can be used to achieve security goals.
  • Advantages of C:
    • Fine-grained control over memory.
    • Essential for systems programming.

This class

  • I am a systems security researcher and I discerned two core gaps in Willamette CS education:
    • Systems, and
    • Security
  • Teaching systems alone can be boring or tedious.
  • Teaching security alone can be defuse and ethereal.

Learning Objectives

  • Differentiate confidentiality, integrity, and availability.
  • Apply modular arithmetic to computational systems.
  • Apply recursion and a pointers to programming problems.
  • Articulate the impossibility of distributed consensus.
  • Understand a low-level systems language.

Sketch

flowchart LR
  A( ) --> B(SHA)
  A --> C(RSA)
  B --> E(Block Chain)
  C --> F(Merkle Tree)
  B --> F
  E --> G(Nakamoto'08)
  F --> G
  G --> I((Capstone))

flowchart LR
  A( ) --> B(SHA)
  A --> C(RSA)
  B --> E(Block Chain)
  C --> F(Merkle Tree)
  B --> F
  E --> G(Nakamoto'08)
  F --> G
  G --> I((Capstone))

SHA

  • SHA Basics:
    • Cryptographic hash function family.
    • Input data into fixed-size hash values.
  • Use Cases:
    • Data integrity.
  • Learning Objectives:
    • Reason about bits and bytes.

RSA

  • RSA Basics:
    • Public-key cryptography algorithm.
    • Based on prime factorization being hard.
  • Use Cases:
    • Data confidentiality.
  • Learning Objectives:
    • Reason about numerical computing.

RSA + SHA = BTC

  • Facilitating Blockchain:
    • SHA & RSA in play
      • RSA: Keep wallet keys confidential
      • SHA: Maintain wallet content integrity
  • But we need some techniques to implement…
    • Pointers & Recursion

Blockchain

  • Blockchain:
    • Blockchain structure resembles linked lists.
      • Linked lists use pointers
      • Pointers are used recursively
    • Pointers are validated with SHA
      • That is, they have integrity

Merkle Trees

  • Merkle Trees:
    • Tree structure using SHA for data integrity.
    • Hashes stored as nodes; pointers link them.
    • Leaf nodes are RSA signatures!
  • Merkle Trees are balanced
    • We get to see a way to implement complex data structures in a low level language in a way that solves a real problem.

Final

  • Implement Full-Custom BTC
    • Under absolutely no circumstances use it
      • It’ll be insecure for reasons out-of-scope

Today

  • ✓ Introduce the course
    • ✓ The C Programming Language
    • ✓ Cryptography + Cryptocurrency
  • Introduce the technologies
    • podman
    • vim
    • gcc
    • git

podman

What is Podman?

  • Definition:
    • Platform to develop, ship, and run applications in isolated containers.
  • Benefits:
    • Portability, consistency across environments.
  • Editorializing:
    • Probably all code should be written in a container now that they exist.

In practice:

  • Windows only
    • Run wsl --install -d ubuntu command once
  • Windows+Mac:
    • Use Podman Installer once
    • Launch Podman Desktop each time you code
    • Use podman commands
  • Linux:
    • Have fun with your package manager.

Circa ~2013

  • The problem:
    • We want to run code.
    • We want the code to create programs
    • We want the programs to work on more than one computer.
  • The solution:
    • Realistically, the Java Virtual Machine
    • But then you have to write Java (this is pre-Scala, pre-Kotlin)

Circa ~2008

  • The setting:
    • Cloud is taking off, mainly Amazon + Google
    • They already have to run everything in virtual machine.
    • They needed a lightweight solution.
  • The solution:
    • Docker is written in the Google C-like, Go
    • It is a lightweight pretend computer.

Docker→Podman

  • Docker is closed-source (bad)
  • Docker is critical infastructre in most code bases
  • The Redhat team (from CAROLINA) stepped into solve the problem
  • They make Podman - completely compatible, completely open-source
  • We use Podman in this class

Podman

  • Podman solves the problem of writing code on one computer and running it another.
  • Simply specify a Podman image
    • a description of a imaginary computer
  • Podman will pretend to be that computer
  • We can run code “inside” Podman
    • Code in the container can’t tell it’s running inside Podman

Podman & Python

  • Windows, Mac, Linux:
    • Win tends to refer to Python3 as “python”
    • Mac tends to refer to Python2 as “python”
    • Linux, because its good, makes you specify:
user@DESKTOP-THMS2PJ:~$ python
Command 'python' not found, did you mean:
  command 'python3' from deb python3
  command 'python' from deb python-is-python3

Bugs

  • This can lead to bugs pretty quickly
    • Say I write and test a .py file on Mac
    • I email it to my prof for help
    • SyntaxError on my prof’s Linux device:
user@DESKTOP-THMS2PJ:~$ python2 -c "print 'hello world'"
hello world
user@DESKTOP-THMS2PJ:~$ python3 -c "print 'hello world'"
  File "<string>", line 1
    print 'hello world'
    ^^^^^^^^^^^^^^^^^^^
SyntaxError: Missing parentheses in call to 'print'. Did you mean print(...)?

Versioning

  • Python2/3 is just an extreme example
    • Most of code reuse has versioning problems
    • Many CS1 headaches are Win/Mac python’ing
    • But really: everything breaks all the time
  • Solution: Just run in podman:
C:\Users\cd-desk>podman run -it python:alpine python
Python 3.13.1 (main, Dec  4 2024, 20:31:31) [GCC 13.2.1 20240309] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> print("hello world")
hello world
>>> exit()

Images

  • Something we podman run -it is an image
    • It is a template that correspond to many possible pretend computers
    • I think of them like object oriented class
    • They define something that could be

Containers

  • Each time we run said command, we have a different container
    • They have different file systems
    • Things we install don’t persist to future run instances.
    • I think of them like objects of a class
    • Like ‘1’ instead of ‘integer’
    • They actually existing, running code.

Today

  • ✓ Introduce the course
    • ✓ The C Programming Language
    • ✓ Cryptography + Cryptocurrency
  • Introduce the technologies
    • ✓ podman
    • vim
    • gcc
    • git

vim

IDEs

  • At some point in time, code started being written in IDEs.
    • Integrated design environments.
    • To me, the first one was Eclipse (for Java)
  • VS Code is the ascendent IDE, and is everywhere.
    • Google Cloud has browser based VS Code clone
  • VS Code has some nice things - including good git and docker (less so podman) integration.

Integration

  • VS Code has some nice things
    • Works with git/docker (less so podman).
    • You know what else can do that?
    • The command line
    • Which you don’t have to install
  • IDEs are ~new, the was command line utilities
    • Vi(m)
    • Emacs
    • Nano

Why Vim?

  • While IDEs are ascendent…
    • The are not installed on every system
    • They can be hard to use in cloud
    • They can start costing money at any time
    • They seem slower and to produce worse code
  • Like Python, IDEs are easy!
    • This is good for new coders
    • It is a design decision thereafter

Vim

  • Vim isn’t too bad
    • It can basically be used as a keyboard navigated text editor.
    • It has really good keyboard shortcuts and modes that we learn over time.
    • It seems to be the pound-for-pound productivity champion of all time.
  • It is hard to find people who have learned vim who don’t prefer it.

Podvim

  • We can install and run vim inside a container (which happens to run Debian & ∴ apt)
  • Even directly from Python…
>>> import os
>>> os.system("apt update") # update the registry of packages
<blah blah blah>
>>> os.system("apt install vim -y") # -y saves us from having to say "yes" to any prompts
<blah blah blah>
>>> os.system("vim") # a vim window will open
  • This is vim inside python inside podman
    • Don’t do this, it’s a proof-of-concept

Vim

  • It looks like this:
~               VIM - Vi IMproved
~                  version 9.1
~            by Bram Moolenaar et al.
~   Vim is open source and freely distributable
~
~          Become a registered Vim user!
~  type  :help register<Enter>   for information
~
~  type  :q<Enter>              to exit
~  type  :help<Enter>  or  <F1> for on-line help
~  type  :help version9<Enter>  for version info

Usage

  • I usually use vim as vim filename.py
    • When vim opens I type i (for “insert”)
    • Then I write code.
    • ESC → :x → ENTER to save+quit
  • Learn vim incrementally by using it for this class - for code, for notes, for whatever.

Today

  • ✓ Introduce the course
    • ✓ The C Programming Language
    • ✓ Cryptography + Cryptocurrency
  • Introduce the technologies
    • ✓ podman
    • ✓ vim
    • gcc
    • git

gcc

GCC

  • Developed for the C language to make an open source operating system in 1987.
  • Mainly around today for the C language to make an open source operating system (Linux)
  • (The C language was invented to write an operating system).
  • python is written in GCC C
C:\Users\cd-desk>podman run -it python
Python 3.13.1 (main, Dec  4 2024, 20:40:27) [GCC 12.2.0] on linux

Great freeware

  • Many legendary programs created using GCC:
    • HTTPD & NGINX, which serve a majority of webtraffic
    • Every major Linux and FreeBSD distribution
    • Major browsers like Firefox and Chromium
      • Firefox is on clang now but that’s new
    • GNU Bash and Coreutils, like cat, ls, rm
    • Bitcoin Core

Compilation

  • Like python or podman or vim, GCC is a command - gcc
  • gcc takes at least one argument: a filename, usually of a C file
  • GCC takes this C file and makes an executable
    • a program, sort of
  • Executables run as command with ./ prefix
    • This differs from python which runs a script without creating a corresponding program.

.py vs .c

flowchart LR
  A(Python) --> B[fname.py file]
  B --> C[python fname.py]
  C --> D{hello world}
  E(C) --> F[fname.c  file]
  F --> G[gcc fname.c -o ename]
  G --> H[./ename]
  H --> I{hello world}

flowchart LR
  A(Python) --> B[fname.py file]
  B --> C[python fname.py]
  C --> D{hello world}
  E(C) --> F[fname.c  file]
  F --> G[gcc fname.c -o ename]
  G --> H[./ename]
  H --> I{hello world}

  • You only have to compile once to have the executable forever.
  • Most programs are executables, not scripts.

Today

  • ✓ Introduce the course
    • ✓ The C Programming Language
    • ✓ Cryptography + Cryptocurrency
  • Introduce the technologies
    • ✓ podman
    • ✓ vim
    • ✓ gcc
    • git

git

Git

  • Git was invented… to store the C language source code for an open source operating system (Linux)
  • It is de facto the only way in the universe to store code other than “on my HDD/SSD”
  • It is also a very good way to
      1. Keep track of podman images
      2. Keep track of podman containers
      3. Keep track of code used inside containers

As a command

  • Git corresponds, like the others, to a command: git
    • It is common now to use other techniques, but the command remains extremely stable
  • Quoth GitHub, the first and greatest of the collaboration tools:
If you want a lot of control and flexibility, you can use the command line.

As a technology

  • So basically, you have things called repositories or “repos”
flowchart LR
  A(My ️‍🔥 code repo)

flowchart LR
  A(My ️‍🔥 code repo)

As a technology

  • Then you realize you wrote an infinite loop so you update it
flowchart LR
  A(My ️‍🔥 code repo v0 ) --> B(My ️‍🔥 code repo v1)

flowchart LR
  A(My ️‍🔥 code repo v0 ) --> B(My ️‍🔥 code repo v1)

As a technology

  • Then you come to class and realize your code is on your gaming rig in your apartment.
flowchart LR
  A(<s>My ️‍🔥 code repo v0</s>) --> B(<s>My ️‍🔥 code repo v1</s>)

flowchart LR
  A(<s>My ️‍🔥 code repo v0</s>) --> B(<s>My ️‍🔥 code repo v1</s>)

As a technology

  • So you save your code on GitHub
    • GH = GitHub, GR = Gaming Rig
flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GH.‍🔥 v0)

flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GH.‍🔥 v0)

As a technology

  • But you realize you didn’t sanitize your inputs so you update again.
flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)

flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)

As a technology

  • Then you have class again and grab the GitHub version onto your LT = Laptop
flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)

flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)

As a technology

  • You notice it has the bug so you fix it again on your laptop in almost the same way
flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)
  E --> F(LT.‍🔥 v1)

flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)
  E --> F(LT.‍🔥 v1)

As a technology

  • And save that back to GitHub then head back home to play Nethack on your 12000USD Gaming PC
flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)
  E --> F(LT.‍🔥 v1)
  F --> G(GH.‍🔥 v1)

flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)
  E --> F(LT.‍🔥 v1)
  F --> G(GH.‍🔥 v1)

As a technology

  • You realize you also added some ASCII art and try to send that to GitHub
flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)
  E --> F(LT.‍🔥 v1)
  F --> G(GH.‍🔥 v1)
  G --> H(GH.‍🔥 v2)
  C --> H

flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)
  E --> F(LT.‍🔥 v1)
  F --> G(GH.‍🔥 v1)
  G --> H(GH.‍🔥 v2)
  C --> H

As a technology

  • Two arrows into one thing is a merge conflict and out-of-scope for now.
flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)
  E --> F(LT.‍🔥 v1)
  F --> G(GH.‍🔥 v1)
  G --> H{💥}
  C --> H

flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)
  E --> F(LT.‍🔥 v1)
  F --> G(GH.‍🔥 v1)
  G --> H{💥}
  C --> H

As a technology

  • Basically versions of code can live in more than one place.
    • Ah, versions, our old friend…
flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)
  E --> F(LT.‍🔥 v1)
  F --> G(GH.‍🔥 v1)
  G --> H{💥}
  C --> H

flowchart LR
  A(GR.️‍🔥 v0) --> B(GR.️‍🔥 v1)
  B --> C(GR.️‍🔥 v2)
  B --> D(GH.‍🔥 v0)
  D --> E(LT.‍🔥 v0)
  E --> F(LT.‍🔥 v1)
  F --> G(GH.‍🔥 v1)
  G --> H{💥}
  C --> H

In Brief

  • Generally create repositories + tokens in browser
  • Use command line and use ‘git clone ’ once.
  • Authenticate via access token from browser
  • Then use ‘git add’, ‘git commit’, ‘git push’ to save work
  • Use ‘git pull’ to grab saved work
  • Use a “.gitignore” file so only code (NOT executables) live on GitHub

Today

  • ✓ Introduce the course
    • ✓ The C Programming Language
    • ✓ Cryptography + Cryptocurrency
  • ✓ Introduce the technologies
    • ✓ podman
    • ✓ vim
    • ✓ gcc
    • ✓ git