Rust Systems Programming

A comprehensive skill for building high-performance, memory-safe systems software using Rust. This skill covers ownership, borrowing, concurrency, async programming, unsafe code, FFI, and performance optimization for systems-level development.

When to Use This Skill

Use this skill when:

Building systems software requiring memory safety without garbage collection
Developing high-performance applications with zero-cost abstractions
Writing concurrent or parallel programs with data race prevention
Creating async/await applications for I/O-bound workloads (web servers, databases)
Working with low-level code, FFI, or hardware interfaces
Replacing C/C++ code with safer alternatives
Building command-line tools, network services, or embedded systems
Optimizing performance-critical sections of applications
Creating libraries that guarantee memory safety at compile time
Developing WebAssembly modules for near-native performance

Core Concepts

The Ownership Model

Rust's ownership system is the foundation of its memory safety guarantees:

Ownership Rules:

Each value in Rust has exactly one owner
When the owner goes out of scope, the value is dropped
Ownership can be transferred (moved) to new owners

Move Semantics:

struct MyStruct { s: u32 }

fn main() {
    let mut x = MyStruct{ s: 5u32 };
    let y = x;  // Ownership moved from x to y
    // x.s = 6;  // ERROR: x is no longer valid
    // println!("{}", x.s);  // ERROR: cannot use x after move
}

When a type doesn't implement Copy, assignment moves ownership rather than copying. This prevents double-free errors and use-after-move bugs at compile time.

For Copy Types:

let x = 5;  // i32 implements Copy
let y = x;  // x is copied, not moved
println!("{}", x);  // OK: x is still valid

Borrowing and References

Borrowing allows temporary access to data without taking ownership:

Immutable Borrowing:

fn main() {
    let s1 = String::from("hello");

    let len = calculate_length(&s1);  // Borrow s1 immutably

    println!("The length of '{}' is {}.", s1, len);  // s1 still valid
}

fn calculate_length(s: &String) -> usize {
    s.len()  // Can read but not modify
}

Mutable Borrowing:

Rust enforces exclusive mutable access to prevent data races:

fn main() {
    let mut value = 3;
    let borrow = &mut value;  // Mutable borrow
    *borrow += 1;
    println!("{}", borrow);  // 4

    // value is accessible again after borrow ends
}

Borrowing Rules:

You can have either one mutable reference OR any number of immutable references
References must always be valid (no dangling pointers)
Mutable and immutable borrows cannot coexist

Common Borrowing Error:

fn main() {
    let mut value = 3;
    // Create a mutable borrow of `value`.
    let borrow = &mut value;
    let _sum = value + 1; // ERROR: cannot use `value` because
                          //        it was mutably borrowed
    println!("{}", borrow);
}

Lifetimes

Lifetimes ensure references are always valid:

fn longest<'a>(x: &'a str, y: &'a str) -> &'a str {
    if x.len() > y.len() {
        x
    } else {
        y
    }
}

The 'a lifetime parameter tells the compiler that the returned reference will be valid as long as both input references are valid.

Ownership Patterns for Sharing

Rc (Reference Counted) for Single-threaded Shared Ownership:

use std::cell::RefCell;
use std::rc::Rc;

struct MyStruct { s: u32 }

fn main() {
    let mut x = Rc::new(RefCell::new(MyStruct{ s: 5u32 }));
    let y = x.clone();  // Increment reference count
    x.borrow_mut().s = 6;  // Interior mutability via RefCell
    println!("{}", x.borrow().s);
}

Rc<T> provides shared ownership with reference counting. RefCell<T> enables interior mutability, enforcing borrow rules at runtime rather than compile time.

Arc (Atomic Reference Counted) for Thread-safe Sharing:

use std::sync::Arc;
use std::thread;

struct FancyNum {
    num: u8,
}

fn main() {
    let fancy_ref1 = Arc::new(FancyNum { num: 5 });
    let fancy_ref2 = fancy_ref1.clone();

    let x = thread::spawn(move || {
        // `fancy_ref1` can be moved and has a `'static` lifetime
        println!("child thread: {}", fancy_ref1.num);
    });

    x.join().expect("child thread should finish");
    println!("main thread: {}", fancy_ref2.num);
}

Arc<T> is the thread-safe version of Rc<T>, using atomic operations for reference counting.

Box for Heap Allocation

Box<T>

Box<T> is an owning pointer that allocates T on the heap. Useful for:

Recursive types with known size
Large values that should not be copied on the stack
Trait objects with dynamic dispatch

Concurrency Patterns

Threads and Message Passing

Rust prevents data races at compile time through its ownership system:

use std::thread;
use std::sync::mpsc;

fn main() {
    let (tx, rx) = mpsc::channel();

    thread::spawn(move || {
        tx.send("Hello from thread").unwrap();
    });

    let message = rx.recv().unwrap();
    println!("{}", message);
}

Shared State with Mutex

use std::sync::{Arc, Mutex};
use std::thread;

fn main() {
    let counter = Arc::new(Mutex::new(0));
    let mut handles = vec![];

    for _ in 0..10 {
        let counter = Arc::clone(&counter);
        let handle = thread::spawn(move || {
            let mut num = counter.lock().unwrap();
            *num += 1;
        });
        handles.push(handle);
    }

    for handle in handles {
        handle.join().unwrap();
    }

    println!("Result: {}", *counter.lock().unwrap());
}

Key Concurrency Types:

Mutex<T>: Mutual exclusion lock for shared mutable state
RwLock<T>: Reader-writer lock allowing multiple readers or one writer
Arc<T>: Atomic reference counting for thread-safe sharing
mpsc: Multi-producer, single-consumer channels

Data Race Prevention

ThreadSanitizer Example:

static mut A: usize = 0;

fn main() {
    let t = std::thread::spawn(|| {
        unsafe { A += 1 };
    });
    unsafe { A += 1 };

    t.join().unwrap();
}

This code has a data race. ThreadSanitizer (enabled with RUSTFLAGS=-Zsanitizer=thread) detects concurrent access to static mutable data:

WARNING: ThreadSanitizer: data race (pid=10574)
  Read of size 8 at 0x5632dfe3d030 by thread T1:
  Previous write of size 8 at 0x5632dfe3d030 by main thread:

Async Programming

Async/Await Basics

Async programming in Rust allows concurrent I/O without blocking threads:

async fn foo(n: usize) {
    if n > 0 {
        Box::pin(foo(n - 1)).await;
    }
}

Recursive async functions require Box::pin() to give the future a known size.

Async Closures

Async Closure with Move Semantics:

fn force_fnonce<T: async FnOnce()>(t: T) -> T { t }

let x = String::new();
let c = force_fnonce(async move || {
    println!("{x}");
});

When constrained to AsyncFnOnce, the closure captures by move to ensure proper ownership.

Async Closure Borrowing:

let x = &1i32; // Lifetime '1
let c = async move || {
    println!("{:?}", *x);
    // Even though the closure moves x, we're only capturing *x,
    // so the inner coroutine can reborrow the data for its original lifetime.
};

Mutable Borrowing in Async Closures:

let mut x = 1i32;
let c = async || {
    x = 1;
    // The parent borrows `x` mutably.
    // When we call `c()`, we implicitly autoref for `AsyncFnMut::async_call_mut`.
    // The inner coroutine captures with the lifetime of the coroutine-closure.
};

Common Async Errors

E0373: Async block capturing short-lived variable:

use std::future::Future;

async fn f() {
    let v = vec![1, 2, 3i32];
    spawn(async { //~ ERROR E0373
        println!("{:?}", v)  // v might go out of scope before async block runs
    });
}

fn spawn<F: Future + Send + 'static>(future: F) {
    unimplemented!()
}

Solution: Move the variable into the async block:

spawn(async move {
    println!("{:?}", v)
})

Unsafe Code and FFI

When to Use Unsafe

Rust's unsafe keyword allows operations that the compiler cannot verify:

Dereferencing raw pointers
Calling unsafe functions or methods
Accessing or modifying mutable static variables
Implementing unsafe traits
Accessing fields of unions

Unsafe Dereference Example:

macro_rules! unsafe_deref {
    () => {
        *(&() as *const ())
    };
}

Memory Safety with Unsafe

Unsafe Sync Implementation:

use std::cell::Cell;

struct NotThreadSafe<T> {
    value: Cell<T>,
}

unsafe impl<T> Sync for NotThreadSafe<T> {}

static A: NotThreadSafe<usize> = NotThreadSafe { value : Cell::new(1) };
static B: &'static NotThreadSafe<usize> = &A; // ok!

This is unsafe because you must manually ensure thread safety. Cell is not Sync by default, so this implementation requires careful reasoning.

FFI (Foreign Function Interface)

Calling C Functions from Rust:

use std::mem;

#[link(name = "foo")]
extern "C" {
    fn do_twice(f: unsafe extern "C" fn(i32) -> i32, arg: i32) -> i32;
}

unsafe extern "C" fn add_one(x: i32) -> i32 {
    x + 1
}

unsafe extern "C" fn add_two(x: i64) -> i64 {
    x + 2
}

fn main() {
    let answer = unsafe { do_twice(add_one, 5) };
    println!("The answer is: {}", answer);

    // Type-mismatched call (unsafe):
    println!("With CFI enabled, you should not see the next answer");
    let f: unsafe extern "C" fn(i32) -> i32 = unsafe {
        mem::transmute::<*const u8, unsafe extern "C" fn(i32) -> i32>(add_two as *const u8)
    };
    let next_answer = unsafe { do_twice(f, 5) };
    println!("The next answer is: {}", next_answer);
}

Control Flow Integrity (CFI): With CFI enabled, the type-mismatched transmute causes program termination, preventing control flow hijacking.

Inline Assembly for System Calls:

static UNMAP_BASE: usize;
const MEM_RELEASE: usize;
static VirtualFree: usize;
const OffPtr: usize;
const OffFn: usize;

core::arch::asm!("
    push {free_type}
    push {free_size}
    push {base}

    mov eax, fs:[30h]
    mov eax, [eax+8h]
    add eax, {off_fn}
    mov [eax-{off_fn}+{off_ptr}], eax

    push eax

    jmp {virtual_free}
    ",
    off_ptr = const OffPtr,
    off_fn  = const OffFn,

    free_size = const 0,
    free_type = const MEM_RELEASE,

    virtual_free = sym VirtualFree,

    base = sym UNMAP_BASE,
    options(noreturn),
);

This demonstrates direct system calls using inline assembly for Windows memory deallocation.

Error Handling

Result and Option Types

Rust uses Result<T, E> and Option<T> for error handling:

fn divide(a: i32, b: i32) -> Result<i32, String> {
    if b == 0 {
        Err("Division by zero".to_string())
    } else {
        Ok(a / b)
    }
}

fn main() {
    match divide(10, 2) {
        Ok(result) => println!("Result: {}", result),
        Err(e) => println!("Error: {}", e),
    }
}

The ? Operator

fn process_file(path: &str) -> Result<String, std::io::Error> {
    let content = std::fs::read_to_string(path)?;
    Ok(content.to_uppercase())
}

The ? operator propagates errors up the call stack, similar to exceptions but explicit in the type signature.

Custom Error Types

use std::fmt;

#[derive(Debug)]
enum AppError {
    IoError(std::io::Error),
    ParseError(std::num::ParseIntError),
    Custom(String),
}

impl fmt::Display for AppError {
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        match self {
            AppError::IoError(e) => write!(f, "IO error: {}", e),
            AppError::ParseError(e) => write!(f, "Parse error: {}", e),
            AppError::Custom(msg) => write!(f, "Error: {}", msg),
        }
    }
}

impl std::error::Error for AppError {}

Memory Safety and Sanitizers

AddressSanitizer

Detecting Stack Buffer Overflow:

fn main() {
    let xs = [0, 1, 2, 3];
    let _y = unsafe { *xs.as_ptr().offset(4) };
}

Build with AddressSanitizer:

$ export RUSTFLAGS=-Zsanitizer=address RUSTDOCFLAGS=-Zsanitizer=address
$ cargo run -Zbuild-std --target x86_64-unknown-linux-gnu

Output:

==37882==ERROR: AddressSanitizer: stack-buffer-overflow on address 0x7ffe400e6250
READ of size 4 at 0x7ffe400e6250 thread T0
    #0 0x5609a841fb1f in example::main::h628ffc6626ed85b2 /.../src/main.rs:3:23

Address 0x7ffe400e6250 is located in stack of thread T0 at offset 48 in frame
  This frame has 1 object(s):
    [32, 48) 'xs' (line 2) <== Memory access at offset 48 overflows this variable

Detecting Heap Buffer Overflow:

fn main() {
    let xs = vec![0, 1, 2, 3];
    let _y = unsafe { *xs.as_ptr().offset(4) };
}

Detecting Use-After-Scope:

static mut P: *mut usize = std::ptr::null_mut();

fn main() {
    unsafe {
        {
            let mut x = 0;
            P = &mut x;
        }
        std::ptr::write_volatile(P, 123);  // P points to dropped variable
    }
}

AddressSanitizer output:

==39249==ERROR: AddressSanitizer: stack-use-after-scope on address 0x7ffc7ed3e1a0
WRITE of size 8 at 0x7ffc7ed3e1a0 thread T0

Performance Optimization

Zero-Cost Abstractions

Rust provides high-level abstractions without runtime overhead:

// Iterator chains are optimized to simple loops
let sum: i32 = (1..100)
    .filter(|x| x % 2 == 0)
    .map(|x| x * x)
    .sum();

The compiler optimizes this to a tight loop equivalent to manual iteration.

Inlining and Monomorphization

Generic functions are monomorphized (specialized) for each concrete type:

#[inline]
fn add<T: std::ops::Add<Output = T>>(a: T, b: T) -> T {
    a + b
}

let x = add(5, 3);      // Specialized for i32
let y = add(5.0, 3.0);  // Specialized for f64

Smart Pointer Overhead

Different smart pointers have different costs:

Box<T>: Single heap allocation, no overhead
Rc<T>: Reference counting, small overhead per clone/drop
Arc<T>: Atomic reference counting, higher overhead for thread safety
Mutex<T>: Lock acquisition overhead
RefCell<T>: Runtime borrow checking overhead

Avoiding Allocations

// Bad: Allocates a new String
fn greet_bad(name: &str) -> String {
    format!("Hello, {}", name)
}

// Good: Returns a reference, no allocation
fn greet_good(name: &str) -> impl std::fmt::Display + '_ {
    format_args!("Hello, {}", name)
}

Common Patterns and Idioms

Builder Pattern

struct Config {
    host: String,
    port: u16,
    timeout: u64,
}

impl Config {
    fn builder() -> ConfigBuilder {
        ConfigBuilder::default()
    }
}

#[derive(Default)]
struct ConfigBuilder {
    host: Option<String>,
    port: Option<u16>,
    timeout: Option<u64>,
}

impl ConfigBuilder {
    fn host(mut self, host: impl Into<String>) -> Self {
        self.host = Some(host.into());
        self
    }

    fn port(mut self, port: u16) -> Self {
        self.port = Some(port);
        self
    }

    fn build(self) -> Config {
        Config {
            host: self.host.unwrap_or_else(|| "localhost".to_string()),
            port: self.port.unwrap_or(8080),
            timeout: self.timeout.unwrap_or(30),
        }
    }
}

Newtype Pattern

struct UserId(u64);
struct PostId(u64);

fn get_user(id: UserId) -> User { /* ... */ }

// This won't compile: type safety!
// get_user(PostId(42));

RAII (Resource Acquisition Is Initialization)

struct FileGuard {
    file: std::fs::File,
}

impl FileGuard {
    fn new(path: &str) -> std::io::Result<Self> {
        Ok(FileGuard {
            file: std::fs::File::create(path)?,
        })
    }
}

impl Drop for FileGuard {
    fn drop(&mut self) {
        println!("File closed automatically");
    }
}

Closure Patterns and Errors

E0500: Closure Borrowing Conflict

Problem:

fn you_know_nothing(jon_snow: &mut i32) {
    let nights_watch = &jon_snow;
    let starks = || {
        *jon_snow = 3; // error: closure requires unique access to `jon_snow`
                       //        but it is already borrowed
    };
    println!("{}", nights_watch);
}

Solution:

fn you_know_nothing(jon_snow: &mut i32) {
    let nights_watch = &jon_snow;
    println!("{}", nights_watch);  // Use the borrow first
    // Borrow ends here (non-lexical lifetimes)
    let starks = || {
        *jon_snow = 3;  // Now OK
    };
}

E0502: Mutable and Immutable Borrows

Problem:

fn bar(x: &mut i32) {}
fn foo(a: &mut i32) {
    let y = &a;
    bar(a);  // Error: cannot borrow as mutable while borrowed as immutable
    println!("{}", y);
}

Solution:

fn bar(x: &mut i32) {}
fn foo(a: &mut i32) {
    bar(a);  // Mutable borrow first
    let y = &a; // Immutable borrow after mutable borrow ends
    println!("{}", y);
}

E0505: Move of Borrowed Value

Problem:

struct Value {}

fn borrow(val: &Value) {}
fn eat(val: Value) {}

fn main() {
    let x = Value{};
    let _ref_to_val: &Value = &x;
    eat(x);  // Error: cannot move x while borrowed
    borrow(_ref_to_val);
}

E0524: Concurrent Mutable Borrows in Closures

Problem:

fn set(x: &mut isize) {
    *x += 4;
}

fn dragoooon(x: &mut isize) {
    let mut c1 = || set(x);
    let mut c2 = || set(x); // error: two closures trying to borrow mutably
    c2();
    c1();
}

Solution 1: Sequential Execution (Non-Lexical Lifetimes)

fn set(x: &mut isize) {
    *x += 4;
}

fn dragoooon(x: &mut isize) {
    {
        let mut c1 = || set(&mut *x);
        c1();
    } // `c1` has been dropped here so we're free to use `x` again!
    let mut c2 = || set(&mut *x);
    c2();
}

Solution 2: Rc + RefCell for Shared Mutable Access

use std::rc::Rc;
use std::cell::RefCell;

fn set(x: &mut isize) {
    *x += 4;
}

fn dragoooon(x: &mut isize) {
    let x = Rc::new(RefCell::new(x));
    let y = Rc::clone(&x);
    let mut c1 = || { let mut x2 = x.borrow_mut(); set(&mut x2); };
    let mut c2 = || { let mut x2 = y.borrow_mut(); set(&mut x2); };

    c2();
    c1();
}

E0507: Moving Borrowed Content

Problem:

use std::cell::RefCell;

struct TheDarkKnight;

impl TheDarkKnight {
    fn nothing_is_true(self) {}
}

fn main() {
    let x = RefCell::new(TheDarkKnight);
    x.borrow().nothing_is_true(); // Error: cannot move out of borrowed content
}

Solution 1: Take a Reference

impl TheDarkKnight {
    fn nothing_is_true(&self) {}  // Change to &self
}

fn main() {
    let x = RefCell::new(TheDarkKnight);
    x.borrow().nothing_is_true(); // OK
}

Solution 2: Reclaim Ownership

fn main() {
    let x = RefCell::new(TheDarkKnight);
    let x = x.into_inner(); // Reclaim ownership
    x.nothing_is_true(); // OK
}

Production Patterns

Structured Logging

use tracing::{info, warn, error, instrument};

#[instrument]
async fn process_request(id: u64) -> Result<(), Error> {
    info!(request_id = id, "Processing request");

    match do_work(id).await {
        Ok(_) => {
            info!("Request completed successfully");
            Ok(())
        }
        Err(e) => {
            error!(error = ?e, "Request failed");
            Err(e)
        }
    }
}

Configuration Management

use serde::Deserialize;

#[derive(Deserialize)]
struct AppConfig {
    database_url: String,
    server_port: u16,
    log_level: String,
}

fn load_config() -> Result<AppConfig, config::ConfigError> {
    config::Config::builder()
        .add_source(config::File::with_name("config"))
        .add_source(config::Environment::with_prefix("APP"))
        .build()?
        .try_deserialize()
}

Graceful Shutdown

use tokio::signal;

async fn shutdown_signal() {
    let ctrl_c = async {
        signal::ctrl_c()
            .await
            .expect("failed to install Ctrl+C handler");
    };

    #[cfg(unix)]
    let terminate = async {
        signal::unix::signal(signal::unix::SignalKind::terminate())
            .expect("failed to install signal handler")
            .recv()
            .await;
    };

    #[cfg(not(unix))]
    let terminate = std::future::pending::<()>();

    tokio::select! {
        _ = ctrl_c => {},
        _ = terminate => {},
    }

    println!("Shutdown signal received, starting graceful shutdown");
}

Connection Pooling

use deadpool_postgres::{Config, Pool, Runtime};
use tokio_postgres::NoTls;

async fn create_pool() -> Pool {
    let mut cfg = Config::new();
    cfg.host = Some("localhost".to_string());
    cfg.dbname = Some("mydb".to_string());

    cfg.create_pool(Some(Runtime::Tokio1), NoTls)
        .expect("Failed to create pool")
}

async fn query_user(pool: &Pool, id: i64) -> Result<User, Error> {
    let client = pool.get().await?;
    let row = client
        .query_one("SELECT * FROM users WHERE id = $1", &[&id])
        .await?;
    Ok(User::from_row(row))
}

Best Practices

API Design

Use References by Default: Accept &T instead of T unless you need ownership
Return Owned Types: Return String instead of &str for simpler APIs
Implement Standard Traits: Debug, Clone, PartialEq, Send, Sync
Use Into<T> for Flexibility: fn set_name(&mut self, name: impl Into<String>)
Leverage the Type System: Use newtypes, enums, and Result for correctness

Testing Patterns

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_addition() {
        assert_eq!(add(2, 2), 4);
    }

    #[test]
    #[should_panic(expected = "division by zero")]
    fn test_divide_by_zero() {
        divide(10, 0);
    }

    #[tokio::test]
    async fn test_async_function() {
        let result = fetch_data().await;
        assert!(result.is_ok());
    }
}

Documentation

/// Calculates the sum of two numbers.
///
/// # Examples
///
/// ```
/// let result = add(2, 2);
/// assert_eq!(result, 4);
/// ```
///
/// # Panics
///
/// This function will panic if the result overflows.
pub fn add(a: i32, b: i32) -> i32 {
    a.checked_add(b).expect("overflow in add")
}

Dependency Management

[dependencies]
tokio = { version = "1", features = ["full"] }
serde = { version = "1", features = ["derive"] }
tracing = "0.1"

[dev-dependencies]
criterion = "0.5"

[profile.release]
lto = true
codegen-units = 1

Troubleshooting

Common Borrow Checker Issues

Issue: "Cannot borrow as mutable because it is also borrowed as immutable" Solution: Ensure mutable and immutable borrows don't overlap in scope

Issue: "Cannot move out of borrowed content" Solution: Clone the value, take a reference, or use Rc<RefCell<T>>

Issue: "Lifetime issues with references" Solution: Add explicit lifetime annotations or restructure code to avoid complex lifetimes

Performance Issues

Issue: Slow compilation times Solution: Use incremental compilation, reduce generic instantiations, use cargo check

Issue: Runtime performance slower than expected Solution: Profile with perf, enable LTO, check for unnecessary allocations/clones

Issue: High memory usage Solution: Use references instead of clones, consider streaming data, profile with Valgrind

Concurrency Issues

Issue: Deadlocks with multiple mutexes Solution: Always acquire locks in the same order, use try_lock with timeout

Issue: Data races in unsafe code Solution: Run with ThreadSanitizer, carefully review unsafe blocks

Issue: Async tasks not making progress Solution: Check for blocking operations in async code, use spawn_blocking for CPU-bound work

Quick Reference

Smart Pointer Cheat Sheet

Box<T>     - Heap allocation, single owner
Rc<T>      - Reference counted, single-threaded
Arc<T>     - Atomic reference counted, thread-safe
Mutex<T>   - Mutual exclusion lock
RwLock<T>  - Reader-writer lock
RefCell<T> - Runtime borrow checking
Cell<T>    - Interior mutability for Copy types

Common Trait Implementations

#[derive(Debug, Clone, PartialEq, Eq, Hash)]
struct MyStruct {
    field: String,
}

impl Default for MyStruct {
    fn default() -> Self {
        Self {
            field: String::new(),
        }
    }
}

impl std::fmt::Display for MyStruct {
    fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
        write!(f, "MyStruct({})", self.field)
    }
}

Async Runtime Comparison

Tokio:    Full-featured, most popular, excellent ecosystem
async-std: Mirrors std library API, simpler for beginners
smol:     Lightweight, minimal dependencies

Resources

The Rust Book: https://doc.rust-lang.org/book/
Rust by Example: https://doc.rust-lang.org/rust-by-example/
The Rustonomicon (unsafe Rust): https://doc.rust-lang.org/nomicon/
Async Book: https://rust-lang.github.io/async-book/
Rust Performance Book: https://nnethercote.github.io/perf-book/
Crate Documentation: https://docs.rs/

Skill Version: 1.0.0 Last Updated: October 2025 Skill Category: Systems Programming, Performance, Memory Safety Context7 Integration: Rust documentation and error code examples