Add Rust storage engine foundation

This commit is contained in:
2026-04-02 17:00:58 +08:00
parent 1eadc7b75c
commit 926a7e6366
48 changed files with 12189 additions and 3 deletions

3
.gitignore vendored
View File

@@ -30,6 +30,9 @@ dist/
myfsio_core/target/ myfsio_core/target/
myfsio_core/Cargo.lock myfsio_core/Cargo.lock
# Rust engine build artifacts
myfsio-engine/target/
# Local runtime artifacts # Local runtime artifacts
logs/ logs/
*.log *.log

View File

@@ -1,6 +1,6 @@
from __future__ import annotations from __future__ import annotations
APP_VERSION = "0.4.2" APP_VERSION = "0.4.3"
def get_version() -> str: def get_version() -> str:

1807
myfsio-engine/Cargo.lock generated Normal file

File diff suppressed because it is too large Load Diff

43
myfsio-engine/Cargo.toml Normal file
View File

@@ -0,0 +1,43 @@
[workspace]
resolver = "2"
members = [
"crates/myfsio-common",
"crates/myfsio-auth",
"crates/myfsio-crypto",
"crates/myfsio-storage",
"crates/myfsio-xml",
"crates/myfsio-server",
]
[workspace.dependencies]
tokio = { version = "1", features = ["full"] }
axum = { version = "0.8" }
tower = { version = "0.5" }
tower-http = { version = "0.6", features = ["cors", "trace"] }
hyper = { version = "1" }
bytes = "1"
serde = { version = "1", features = ["derive"] }
serde_json = "1"
quick-xml = { version = "0.37", features = ["serialize"] }
hmac = "0.12"
sha2 = "0.10"
md-5 = "0.10"
hex = "0.4"
aes = "0.8"
aes-gcm = "0.10"
cbc = { version = "0.1", features = ["alloc"] }
hkdf = "0.12"
uuid = { version = "1", features = ["v4"] }
parking_lot = "0.12"
lru = "0.14"
percent-encoding = "2"
regex = "1"
unicode-normalization = "0.1"
tracing = "0.1"
tracing-subscriber = "0.3"
thiserror = "2"
chrono = { version = "0.4", features = ["serde"] }
base64 = "0.22"
tokio-util = { version = "0.7", features = ["io"] }
futures = "0.3"
dashmap = "6"

View File

@@ -0,0 +1,25 @@
[package]
name = "myfsio-auth"
version = "0.1.0"
edition = "2021"
[dependencies]
myfsio-common = { path = "../myfsio-common" }
hmac = { workspace = true }
sha2 = { workspace = true }
hex = { workspace = true }
aes = { workspace = true }
cbc = { workspace = true }
base64 = { workspace = true }
pbkdf2 = "0.12"
lru = { workspace = true }
parking_lot = { workspace = true }
percent-encoding = { workspace = true }
serde = { workspace = true }
serde_json = { workspace = true }
thiserror = { workspace = true }
chrono = { workspace = true }
tracing = { workspace = true }
[dev-dependencies]
tempfile = "3"

View File

@@ -0,0 +1,80 @@
use aes::cipher::{block_padding::Pkcs7, BlockDecryptMut, KeyIvInit};
use base64::{engine::general_purpose::URL_SAFE, Engine};
use hmac::{Hmac, Mac};
use sha2::Sha256;
type Aes128CbcDec = cbc::Decryptor<aes::Aes128>;
type HmacSha256 = Hmac<Sha256>;
pub fn derive_fernet_key(secret: &str) -> String {
let mut derived = [0u8; 32];
pbkdf2::pbkdf2_hmac::<Sha256>(
secret.as_bytes(),
b"myfsio-iam-encryption",
100_000,
&mut derived,
);
URL_SAFE.encode(derived)
}
pub fn decrypt(key_b64: &str, token: &str) -> Result<Vec<u8>, &'static str> {
let key_bytes = URL_SAFE
.decode(key_b64)
.map_err(|_| "invalid fernet key base64")?;
if key_bytes.len() != 32 {
return Err("fernet key must be 32 bytes");
}
let signing_key = &key_bytes[..16];
let encryption_key = &key_bytes[16..];
let token_bytes = URL_SAFE
.decode(token)
.map_err(|_| "invalid fernet token base64")?;
if token_bytes.len() < 57 {
return Err("fernet token too short");
}
if token_bytes[0] != 0x80 {
return Err("invalid fernet version");
}
let hmac_offset = token_bytes.len() - 32;
let payload = &token_bytes[..hmac_offset];
let expected_hmac = &token_bytes[hmac_offset..];
let mut mac =
HmacSha256::new_from_slice(signing_key).map_err(|_| "hmac key error")?;
mac.update(payload);
mac.verify_slice(expected_hmac)
.map_err(|_| "HMAC verification failed")?;
let iv = &token_bytes[9..25];
let ciphertext = &token_bytes[25..hmac_offset];
let plaintext = Aes128CbcDec::new(encryption_key.into(), iv.into())
.decrypt_padded_vec_mut::<Pkcs7>(ciphertext)
.map_err(|_| "AES-CBC decryption failed")?;
Ok(plaintext)
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_derive_fernet_key_format() {
let key = derive_fernet_key("test-secret");
let decoded = URL_SAFE.decode(&key).unwrap();
assert_eq!(decoded.len(), 32);
}
#[test]
fn test_roundtrip_with_python_compat() {
let key = derive_fernet_key("dev-secret-key");
let decoded = URL_SAFE.decode(&key).unwrap();
assert_eq!(decoded.len(), 32);
}
}

View File

@@ -0,0 +1,499 @@
use chrono::{DateTime, Utc};
use myfsio_common::types::Principal;
use parking_lot::RwLock;
use serde::{Deserialize, Serialize};
use std::collections::HashMap;
use std::path::PathBuf;
use std::sync::Arc;
use std::time::{Instant, SystemTime};
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct IamConfig {
#[serde(default = "default_version")]
pub version: u32,
#[serde(default)]
pub users: Vec<IamUser>,
}
fn default_version() -> u32 {
2
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct IamUser {
pub user_id: String,
pub display_name: String,
#[serde(default = "default_enabled")]
pub enabled: bool,
#[serde(default)]
pub expires_at: Option<String>,
#[serde(default)]
pub access_keys: Vec<AccessKey>,
#[serde(default)]
pub policies: Vec<IamPolicy>,
}
fn default_enabled() -> bool {
true
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct AccessKey {
pub access_key: String,
pub secret_key: String,
#[serde(default = "default_status")]
pub status: String,
#[serde(default)]
pub created_at: Option<String>,
}
fn default_status() -> String {
"active".to_string()
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct IamPolicy {
pub bucket: String,
pub actions: Vec<String>,
#[serde(default = "default_prefix")]
pub prefix: String,
}
fn default_prefix() -> String {
"*".to_string()
}
struct IamState {
key_secrets: HashMap<String, String>,
key_index: HashMap<String, String>,
key_status: HashMap<String, String>,
user_records: HashMap<String, IamUser>,
file_mtime: Option<SystemTime>,
last_check: Instant,
}
pub struct IamService {
config_path: PathBuf,
state: Arc<RwLock<IamState>>,
check_interval: std::time::Duration,
fernet_key: Option<String>,
}
impl IamService {
pub fn new(config_path: PathBuf) -> Self {
Self::new_with_secret(config_path, None)
}
pub fn new_with_secret(config_path: PathBuf, secret_key: Option<String>) -> Self {
let fernet_key = secret_key.map(|s| crate::fernet::derive_fernet_key(&s));
let service = Self {
config_path,
state: Arc::new(RwLock::new(IamState {
key_secrets: HashMap::new(),
key_index: HashMap::new(),
key_status: HashMap::new(),
user_records: HashMap::new(),
file_mtime: None,
last_check: Instant::now(),
})),
check_interval: std::time::Duration::from_secs(2),
fernet_key,
};
service.reload();
service
}
fn reload_if_needed(&self) {
{
let state = self.state.read();
if state.last_check.elapsed() < self.check_interval {
return;
}
}
let current_mtime = std::fs::metadata(&self.config_path)
.and_then(|m| m.modified())
.ok();
let needs_reload = {
let state = self.state.read();
match (&state.file_mtime, &current_mtime) {
(None, Some(_)) => true,
(Some(old), Some(new)) => old != new,
(Some(_), None) => true,
(None, None) => state.key_secrets.is_empty(),
}
};
if needs_reload {
self.reload();
}
self.state.write().last_check = Instant::now();
}
fn reload(&self) {
let content = match std::fs::read_to_string(&self.config_path) {
Ok(c) => c,
Err(e) => {
tracing::warn!("Failed to read IAM config {}: {}", self.config_path.display(), e);
return;
}
};
let raw = if content.starts_with("MYFSIO_IAM_ENC:") {
let encrypted_token = &content["MYFSIO_IAM_ENC:".len()..];
match &self.fernet_key {
Some(key) => match crate::fernet::decrypt(key, encrypted_token.trim()) {
Ok(plaintext) => match String::from_utf8(plaintext) {
Ok(s) => s,
Err(e) => {
tracing::error!("Decrypted IAM config is not valid UTF-8: {}", e);
return;
}
},
Err(e) => {
tracing::error!("Failed to decrypt IAM config: {}. SECRET_KEY may have changed.", e);
return;
}
},
None => {
tracing::error!("IAM config is encrypted but no SECRET_KEY configured");
return;
}
}
} else {
content
};
let config: IamConfig = match serde_json::from_str(&raw) {
Ok(c) => c,
Err(e) => {
tracing::error!("Failed to parse IAM config: {}", e);
return;
}
};
let mut key_secrets = HashMap::new();
let mut key_index = HashMap::new();
let mut key_status = HashMap::new();
let mut user_records = HashMap::new();
for user in &config.users {
user_records.insert(user.user_id.clone(), user.clone());
for ak in &user.access_keys {
key_secrets.insert(ak.access_key.clone(), ak.secret_key.clone());
key_index.insert(ak.access_key.clone(), user.user_id.clone());
key_status.insert(ak.access_key.clone(), ak.status.clone());
}
}
let file_mtime = std::fs::metadata(&self.config_path)
.and_then(|m| m.modified())
.ok();
let mut state = self.state.write();
state.key_secrets = key_secrets;
state.key_index = key_index;
state.key_status = key_status;
state.user_records = user_records;
state.file_mtime = file_mtime;
state.last_check = Instant::now();
tracing::info!("IAM config reloaded: {} users, {} keys",
config.users.len(),
state.key_secrets.len());
}
pub fn get_secret_key(&self, access_key: &str) -> Option<String> {
self.reload_if_needed();
let state = self.state.read();
let status = state.key_status.get(access_key)?;
if status != "active" {
return None;
}
let user_id = state.key_index.get(access_key)?;
let user = state.user_records.get(user_id)?;
if !user.enabled {
return None;
}
if let Some(ref expires_at) = user.expires_at {
if let Ok(exp) = expires_at.parse::<DateTime<Utc>>() {
if Utc::now() > exp {
return None;
}
}
}
state.key_secrets.get(access_key).cloned()
}
pub fn get_principal(&self, access_key: &str) -> Option<Principal> {
self.reload_if_needed();
let state = self.state.read();
let status = state.key_status.get(access_key)?;
if status != "active" {
return None;
}
let user_id = state.key_index.get(access_key)?;
let user = state.user_records.get(user_id)?;
if !user.enabled {
return None;
}
if let Some(ref expires_at) = user.expires_at {
if let Ok(exp) = expires_at.parse::<DateTime<Utc>>() {
if Utc::now() > exp {
return None;
}
}
}
let is_admin = user.policies.iter().any(|p| {
p.bucket == "*" && p.actions.iter().any(|a| a == "*")
});
Some(Principal::new(
access_key.to_string(),
user.user_id.clone(),
user.display_name.clone(),
is_admin,
))
}
pub fn authenticate(&self, access_key: &str, secret_key: &str) -> Option<Principal> {
let stored_secret = self.get_secret_key(access_key)?;
if !crate::sigv4::constant_time_compare(&stored_secret, secret_key) {
return None;
}
self.get_principal(access_key)
}
pub async fn list_users(&self) -> Vec<serde_json::Value> {
self.reload_if_needed();
let state = self.state.read();
state
.user_records
.values()
.map(|u| {
serde_json::json!({
"user_id": u.user_id,
"display_name": u.display_name,
"enabled": u.enabled,
"access_keys": u.access_keys.iter().map(|k| {
serde_json::json!({
"access_key": k.access_key,
"status": k.status,
"created_at": k.created_at,
})
}).collect::<Vec<_>>(),
"policy_count": u.policies.len(),
})
})
.collect()
}
pub async fn get_user(&self, identifier: &str) -> Option<serde_json::Value> {
self.reload_if_needed();
let state = self.state.read();
let user = state
.user_records
.get(identifier)
.or_else(|| {
state.key_index.get(identifier).and_then(|uid| state.user_records.get(uid))
})?;
Some(serde_json::json!({
"user_id": user.user_id,
"display_name": user.display_name,
"enabled": user.enabled,
"expires_at": user.expires_at,
"access_keys": user.access_keys.iter().map(|k| {
serde_json::json!({
"access_key": k.access_key,
"status": k.status,
"created_at": k.created_at,
})
}).collect::<Vec<_>>(),
"policies": user.policies,
}))
}
pub async fn set_user_enabled(&self, identifier: &str, enabled: bool) -> Result<(), String> {
let content = std::fs::read_to_string(&self.config_path)
.map_err(|e| format!("Failed to read IAM config: {}", e))?;
let mut config: IamConfig = serde_json::from_str(&content)
.map_err(|e| format!("Failed to parse IAM config: {}", e))?;
let user = config
.users
.iter_mut()
.find(|u| {
u.user_id == identifier
|| u.access_keys.iter().any(|k| k.access_key == identifier)
})
.ok_or_else(|| "User not found".to_string())?;
user.enabled = enabled;
let json = serde_json::to_string_pretty(&config)
.map_err(|e| format!("Failed to serialize IAM config: {}", e))?;
std::fs::write(&self.config_path, json)
.map_err(|e| format!("Failed to write IAM config: {}", e))?;
self.reload();
Ok(())
}
}
#[cfg(test)]
mod tests {
use super::*;
use std::io::Write;
fn test_iam_json() -> String {
serde_json::json!({
"version": 2,
"users": [{
"user_id": "u-test1234",
"display_name": "admin",
"enabled": true,
"access_keys": [{
"access_key": "AKIAIOSFODNN7EXAMPLE",
"secret_key": "wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY",
"status": "active",
"created_at": "2024-01-01T00:00:00Z"
}],
"policies": [{
"bucket": "*",
"actions": ["*"],
"prefix": "*"
}]
}]
})
.to_string()
}
#[test]
fn test_load_and_lookup() {
let mut tmp = tempfile::NamedTempFile::new().unwrap();
tmp.write_all(test_iam_json().as_bytes()).unwrap();
tmp.flush().unwrap();
let svc = IamService::new(tmp.path().to_path_buf());
let secret = svc.get_secret_key("AKIAIOSFODNN7EXAMPLE");
assert_eq!(
secret.unwrap(),
"wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY"
);
}
#[test]
fn test_get_principal() {
let mut tmp = tempfile::NamedTempFile::new().unwrap();
tmp.write_all(test_iam_json().as_bytes()).unwrap();
tmp.flush().unwrap();
let svc = IamService::new(tmp.path().to_path_buf());
let principal = svc.get_principal("AKIAIOSFODNN7EXAMPLE").unwrap();
assert_eq!(principal.display_name, "admin");
assert_eq!(principal.user_id, "u-test1234");
assert!(principal.is_admin);
}
#[test]
fn test_authenticate_success() {
let mut tmp = tempfile::NamedTempFile::new().unwrap();
tmp.write_all(test_iam_json().as_bytes()).unwrap();
tmp.flush().unwrap();
let svc = IamService::new(tmp.path().to_path_buf());
let principal = svc
.authenticate(
"AKIAIOSFODNN7EXAMPLE",
"wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY",
)
.unwrap();
assert_eq!(principal.display_name, "admin");
}
#[test]
fn test_authenticate_wrong_secret() {
let mut tmp = tempfile::NamedTempFile::new().unwrap();
tmp.write_all(test_iam_json().as_bytes()).unwrap();
tmp.flush().unwrap();
let svc = IamService::new(tmp.path().to_path_buf());
assert!(svc.authenticate("AKIAIOSFODNN7EXAMPLE", "wrongsecret").is_none());
}
#[test]
fn test_unknown_key_returns_none() {
let mut tmp = tempfile::NamedTempFile::new().unwrap();
tmp.write_all(test_iam_json().as_bytes()).unwrap();
tmp.flush().unwrap();
let svc = IamService::new(tmp.path().to_path_buf());
assert!(svc.get_secret_key("NONEXISTENTKEY").is_none());
assert!(svc.get_principal("NONEXISTENTKEY").is_none());
}
#[test]
fn test_disabled_user() {
let json = serde_json::json!({
"version": 2,
"users": [{
"user_id": "u-disabled",
"display_name": "disabled-user",
"enabled": false,
"access_keys": [{
"access_key": "DISABLED_KEY",
"secret_key": "secret123",
"status": "active"
}],
"policies": []
}]
})
.to_string();
let mut tmp = tempfile::NamedTempFile::new().unwrap();
tmp.write_all(json.as_bytes()).unwrap();
tmp.flush().unwrap();
let svc = IamService::new(tmp.path().to_path_buf());
assert!(svc.get_secret_key("DISABLED_KEY").is_none());
}
#[test]
fn test_inactive_key() {
let json = serde_json::json!({
"version": 2,
"users": [{
"user_id": "u-test",
"display_name": "test",
"enabled": true,
"access_keys": [{
"access_key": "INACTIVE_KEY",
"secret_key": "secret123",
"status": "inactive"
}],
"policies": []
}]
})
.to_string();
let mut tmp = tempfile::NamedTempFile::new().unwrap();
tmp.write_all(json.as_bytes()).unwrap();
tmp.flush().unwrap();
let svc = IamService::new(tmp.path().to_path_buf());
assert!(svc.get_secret_key("INACTIVE_KEY").is_none());
}
}

View File

@@ -0,0 +1,4 @@
pub mod sigv4;
pub mod principal;
pub mod iam;
mod fernet;

View File

@@ -0,0 +1 @@
pub use myfsio_common::types::Principal;

View File

@@ -0,0 +1,258 @@
use hmac::{Hmac, Mac};
use lru::LruCache;
use parking_lot::Mutex;
use percent_encoding::{percent_encode, AsciiSet, NON_ALPHANUMERIC};
use sha2::{Digest, Sha256};
use std::num::NonZeroUsize;
use std::sync::LazyLock;
use std::time::Instant;
type HmacSha256 = Hmac<Sha256>;
struct CacheEntry {
key: Vec<u8>,
created: Instant,
}
static SIGNING_KEY_CACHE: LazyLock<Mutex<LruCache<(String, String, String, String), CacheEntry>>> =
LazyLock::new(|| Mutex::new(LruCache::new(NonZeroUsize::new(256).unwrap())));
const CACHE_TTL_SECS: u64 = 60;
const AWS_ENCODE_SET: &AsciiSet = &NON_ALPHANUMERIC
.remove(b'-')
.remove(b'_')
.remove(b'.')
.remove(b'~');
fn hmac_sha256(key: &[u8], msg: &[u8]) -> Vec<u8> {
let mut mac = HmacSha256::new_from_slice(key).expect("HMAC key length is always valid");
mac.update(msg);
mac.finalize().into_bytes().to_vec()
}
fn sha256_hex(data: &[u8]) -> String {
let mut hasher = Sha256::new();
hasher.update(data);
hex::encode(hasher.finalize())
}
fn aws_uri_encode(input: &str) -> String {
percent_encode(input.as_bytes(), AWS_ENCODE_SET).to_string()
}
pub fn derive_signing_key_cached(
secret_key: &str,
date_stamp: &str,
region: &str,
service: &str,
) -> Vec<u8> {
let cache_key = (
secret_key.to_owned(),
date_stamp.to_owned(),
region.to_owned(),
service.to_owned(),
);
{
let mut cache = SIGNING_KEY_CACHE.lock();
if let Some(entry) = cache.get(&cache_key) {
if entry.created.elapsed().as_secs() < CACHE_TTL_SECS {
return entry.key.clone();
}
cache.pop(&cache_key);
}
}
let k_date = hmac_sha256(format!("AWS4{}", secret_key).as_bytes(), date_stamp.as_bytes());
let k_region = hmac_sha256(&k_date, region.as_bytes());
let k_service = hmac_sha256(&k_region, service.as_bytes());
let k_signing = hmac_sha256(&k_service, b"aws4_request");
{
let mut cache = SIGNING_KEY_CACHE.lock();
cache.put(
cache_key,
CacheEntry {
key: k_signing.clone(),
created: Instant::now(),
},
);
}
k_signing
}
fn constant_time_compare_inner(a: &[u8], b: &[u8]) -> bool {
if a.len() != b.len() {
return false;
}
let mut result: u8 = 0;
for (x, y) in a.iter().zip(b.iter()) {
result |= x ^ y;
}
result == 0
}
pub fn verify_sigv4_signature(
method: &str,
canonical_uri: &str,
query_params: &[(String, String)],
signed_headers_str: &str,
header_values: &[(String, String)],
payload_hash: &str,
amz_date: &str,
date_stamp: &str,
region: &str,
service: &str,
secret_key: &str,
provided_signature: &str,
) -> bool {
let mut sorted_params = query_params.to_vec();
sorted_params.sort_by(|a, b| a.0.cmp(&b.0).then_with(|| a.1.cmp(&b.1)));
let canonical_query_string = sorted_params
.iter()
.map(|(k, v)| format!("{}={}", aws_uri_encode(k), aws_uri_encode(v)))
.collect::<Vec<_>>()
.join("&");
let mut canonical_headers = String::new();
for (name, value) in header_values {
let lower_name = name.to_lowercase();
let normalized = value.split_whitespace().collect::<Vec<_>>().join(" ");
let final_value = if lower_name == "expect" && normalized.is_empty() {
"100-continue"
} else {
&normalized
};
canonical_headers.push_str(&lower_name);
canonical_headers.push(':');
canonical_headers.push_str(final_value);
canonical_headers.push('\n');
}
let canonical_request = format!(
"{}\n{}\n{}\n{}\n{}\n{}",
method, canonical_uri, canonical_query_string, canonical_headers, signed_headers_str,
payload_hash
);
let credential_scope = format!("{}/{}/{}/aws4_request", date_stamp, region, service);
let cr_hash = sha256_hex(canonical_request.as_bytes());
let string_to_sign = format!(
"AWS4-HMAC-SHA256\n{}\n{}\n{}",
amz_date, credential_scope, cr_hash
);
let signing_key = derive_signing_key_cached(secret_key, date_stamp, region, service);
let calculated = hmac_sha256(&signing_key, string_to_sign.as_bytes());
let calculated_hex = hex::encode(&calculated);
constant_time_compare_inner(calculated_hex.as_bytes(), provided_signature.as_bytes())
}
pub fn derive_signing_key(
secret_key: &str,
date_stamp: &str,
region: &str,
service: &str,
) -> Vec<u8> {
derive_signing_key_cached(secret_key, date_stamp, region, service)
}
pub fn compute_signature(signing_key: &[u8], string_to_sign: &str) -> String {
let sig = hmac_sha256(signing_key, string_to_sign.as_bytes());
hex::encode(sig)
}
pub fn build_string_to_sign(
amz_date: &str,
credential_scope: &str,
canonical_request: &str,
) -> String {
let cr_hash = sha256_hex(canonical_request.as_bytes());
format!(
"AWS4-HMAC-SHA256\n{}\n{}\n{}",
amz_date, credential_scope, cr_hash
)
}
pub fn constant_time_compare(a: &str, b: &str) -> bool {
constant_time_compare_inner(a.as_bytes(), b.as_bytes())
}
pub fn clear_signing_key_cache() {
SIGNING_KEY_CACHE.lock().clear();
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_derive_signing_key() {
let key = derive_signing_key("wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY", "20130524", "us-east-1", "s3");
assert_eq!(key.len(), 32);
}
#[test]
fn test_derive_signing_key_cached() {
let key1 = derive_signing_key("secret", "20240101", "us-east-1", "s3");
let key2 = derive_signing_key("secret", "20240101", "us-east-1", "s3");
assert_eq!(key1, key2);
}
#[test]
fn test_constant_time_compare() {
assert!(constant_time_compare("abc", "abc"));
assert!(!constant_time_compare("abc", "abd"));
assert!(!constant_time_compare("abc", "abcd"));
}
#[test]
fn test_build_string_to_sign() {
let result = build_string_to_sign("20130524T000000Z", "20130524/us-east-1/s3/aws4_request", "GET\n/\n\nhost:example.com\n\nhost\nUNSIGNED-PAYLOAD");
assert!(result.starts_with("AWS4-HMAC-SHA256\n"));
assert!(result.contains("20130524T000000Z"));
}
#[test]
fn test_aws_uri_encode() {
assert_eq!(aws_uri_encode("hello world"), "hello%20world");
assert_eq!(aws_uri_encode("test-file_name.txt"), "test-file_name.txt");
assert_eq!(aws_uri_encode("a/b"), "a%2Fb");
}
#[test]
fn test_verify_sigv4_roundtrip() {
let secret = "wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY";
let date_stamp = "20130524";
let region = "us-east-1";
let service = "s3";
let amz_date = "20130524T000000Z";
let signing_key = derive_signing_key(secret, date_stamp, region, service);
let canonical_request = "GET\n/\n\nhost:examplebucket.s3.amazonaws.com\n\nhost\nUNSIGNED-PAYLOAD";
let string_to_sign = build_string_to_sign(amz_date, &format!("{}/{}/{}/aws4_request", date_stamp, region, service), canonical_request);
let signature = compute_signature(&signing_key, &string_to_sign);
let result = verify_sigv4_signature(
"GET",
"/",
&[],
"host",
&[("host".to_string(), "examplebucket.s3.amazonaws.com".to_string())],
"UNSIGNED-PAYLOAD",
amz_date,
date_stamp,
region,
service,
secret,
&signature,
);
assert!(result);
}
}

View File

@@ -0,0 +1,11 @@
[package]
name = "myfsio-common"
version = "0.1.0"
edition = "2021"
[dependencies]
thiserror = { workspace = true }
serde = { workspace = true }
serde_json = { workspace = true }
chrono = { workspace = true }
uuid = { workspace = true }

View File

@@ -0,0 +1,20 @@
pub const SYSTEM_ROOT: &str = ".myfsio.sys";
pub const SYSTEM_BUCKETS_DIR: &str = "buckets";
pub const SYSTEM_MULTIPART_DIR: &str = "multipart";
pub const BUCKET_META_DIR: &str = "meta";
pub const BUCKET_VERSIONS_DIR: &str = "versions";
pub const BUCKET_CONFIG_FILE: &str = ".bucket.json";
pub const STATS_FILE: &str = "stats.json";
pub const ETAG_INDEX_FILE: &str = "etag_index.json";
pub const INDEX_FILE: &str = "_index.json";
pub const MANIFEST_FILE: &str = "manifest.json";
pub const INTERNAL_FOLDERS: &[&str] = &[".meta", ".versions", ".multipart"];
pub const DEFAULT_REGION: &str = "us-east-1";
pub const AWS_SERVICE: &str = "s3";
pub const DEFAULT_MAX_KEYS: usize = 1000;
pub const DEFAULT_OBJECT_KEY_MAX_BYTES: usize = 1024;
pub const DEFAULT_CHUNK_SIZE: usize = 65536;
pub const STREAM_CHUNK_SIZE: usize = 1_048_576;

View File

@@ -0,0 +1,221 @@
use std::fmt;
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum S3ErrorCode {
AccessDenied,
BucketAlreadyExists,
BucketNotEmpty,
EntityTooLarge,
InternalError,
InvalidAccessKeyId,
InvalidArgument,
InvalidBucketName,
InvalidKey,
InvalidRange,
InvalidRequest,
MalformedXML,
MethodNotAllowed,
NoSuchBucket,
NoSuchKey,
NoSuchUpload,
NoSuchVersion,
NoSuchTagSet,
PreconditionFailed,
NotModified,
QuotaExceeded,
SignatureDoesNotMatch,
SlowDown,
}
impl S3ErrorCode {
pub fn http_status(&self) -> u16 {
match self {
Self::AccessDenied => 403,
Self::BucketAlreadyExists => 409,
Self::BucketNotEmpty => 409,
Self::EntityTooLarge => 413,
Self::InternalError => 500,
Self::InvalidAccessKeyId => 403,
Self::InvalidArgument => 400,
Self::InvalidBucketName => 400,
Self::InvalidKey => 400,
Self::InvalidRange => 416,
Self::InvalidRequest => 400,
Self::MalformedXML => 400,
Self::MethodNotAllowed => 405,
Self::NoSuchBucket => 404,
Self::NoSuchKey => 404,
Self::NoSuchUpload => 404,
Self::NoSuchVersion => 404,
Self::NoSuchTagSet => 404,
Self::PreconditionFailed => 412,
Self::NotModified => 304,
Self::QuotaExceeded => 403,
Self::SignatureDoesNotMatch => 403,
Self::SlowDown => 429,
}
}
pub fn as_str(&self) -> &'static str {
match self {
Self::AccessDenied => "AccessDenied",
Self::BucketAlreadyExists => "BucketAlreadyExists",
Self::BucketNotEmpty => "BucketNotEmpty",
Self::EntityTooLarge => "EntityTooLarge",
Self::InternalError => "InternalError",
Self::InvalidAccessKeyId => "InvalidAccessKeyId",
Self::InvalidArgument => "InvalidArgument",
Self::InvalidBucketName => "InvalidBucketName",
Self::InvalidKey => "InvalidKey",
Self::InvalidRange => "InvalidRange",
Self::InvalidRequest => "InvalidRequest",
Self::MalformedXML => "MalformedXML",
Self::MethodNotAllowed => "MethodNotAllowed",
Self::NoSuchBucket => "NoSuchBucket",
Self::NoSuchKey => "NoSuchKey",
Self::NoSuchUpload => "NoSuchUpload",
Self::NoSuchVersion => "NoSuchVersion",
Self::NoSuchTagSet => "NoSuchTagSet",
Self::PreconditionFailed => "PreconditionFailed",
Self::NotModified => "NotModified",
Self::QuotaExceeded => "QuotaExceeded",
Self::SignatureDoesNotMatch => "SignatureDoesNotMatch",
Self::SlowDown => "SlowDown",
}
}
pub fn default_message(&self) -> &'static str {
match self {
Self::AccessDenied => "Access Denied",
Self::BucketAlreadyExists => "The requested bucket name is not available",
Self::BucketNotEmpty => "The bucket you tried to delete is not empty",
Self::EntityTooLarge => "Your proposed upload exceeds the maximum allowed size",
Self::InternalError => "We encountered an internal error. Please try again.",
Self::InvalidAccessKeyId => "The access key ID you provided does not exist",
Self::InvalidArgument => "Invalid argument",
Self::InvalidBucketName => "The specified bucket is not valid",
Self::InvalidKey => "The specified key is not valid",
Self::InvalidRange => "The requested range is not satisfiable",
Self::InvalidRequest => "Invalid request",
Self::MalformedXML => "The XML you provided was not well-formed",
Self::MethodNotAllowed => "The specified method is not allowed against this resource",
Self::NoSuchBucket => "The specified bucket does not exist",
Self::NoSuchKey => "The specified key does not exist",
Self::NoSuchUpload => "The specified multipart upload does not exist",
Self::NoSuchVersion => "The specified version does not exist",
Self::NoSuchTagSet => "The TagSet does not exist",
Self::PreconditionFailed => "At least one of the preconditions you specified did not hold",
Self::NotModified => "Not Modified",
Self::QuotaExceeded => "The bucket quota has been exceeded",
Self::SignatureDoesNotMatch => "The request signature we calculated does not match the signature you provided",
Self::SlowDown => "Please reduce your request rate",
}
}
}
impl fmt::Display for S3ErrorCode {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
f.write_str(self.as_str())
}
}
#[derive(Debug, Clone)]
pub struct S3Error {
pub code: S3ErrorCode,
pub message: String,
pub resource: String,
pub request_id: String,
}
impl S3Error {
pub fn new(code: S3ErrorCode, message: impl Into<String>) -> Self {
Self {
code,
message: message.into(),
resource: String::new(),
request_id: String::new(),
}
}
pub fn from_code(code: S3ErrorCode) -> Self {
Self::new(code, code.default_message())
}
pub fn with_resource(mut self, resource: impl Into<String>) -> Self {
self.resource = resource.into();
self
}
pub fn with_request_id(mut self, request_id: impl Into<String>) -> Self {
self.request_id = request_id.into();
self
}
pub fn http_status(&self) -> u16 {
self.code.http_status()
}
pub fn to_xml(&self) -> String {
format!(
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<Error>\
<Code>{}</Code>\
<Message>{}</Message>\
<Resource>{}</Resource>\
<RequestId>{}</RequestId>\
</Error>",
self.code.as_str(),
xml_escape(&self.message),
xml_escape(&self.resource),
xml_escape(&self.request_id),
)
}
}
impl fmt::Display for S3Error {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
write!(f, "{}: {}", self.code, self.message)
}
}
impl std::error::Error for S3Error {}
fn xml_escape(s: &str) -> String {
s.replace('&', "&amp;")
.replace('<', "&lt;")
.replace('>', "&gt;")
.replace('"', "&quot;")
.replace('\'', "&apos;")
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_error_codes() {
assert_eq!(S3ErrorCode::NoSuchKey.http_status(), 404);
assert_eq!(S3ErrorCode::AccessDenied.http_status(), 403);
assert_eq!(S3ErrorCode::NoSuchBucket.as_str(), "NoSuchBucket");
}
#[test]
fn test_error_to_xml() {
let err = S3Error::from_code(S3ErrorCode::NoSuchKey)
.with_resource("/test-bucket/test-key")
.with_request_id("abc123");
let xml = err.to_xml();
assert!(xml.contains("<Code>NoSuchKey</Code>"));
assert!(xml.contains("<Resource>/test-bucket/test-key</Resource>"));
assert!(xml.contains("<RequestId>abc123</RequestId>"));
}
#[test]
fn test_xml_escape() {
let err = S3Error::new(S3ErrorCode::InvalidArgument, "key <test> & \"value\"")
.with_resource("/bucket/key&amp");
let xml = err.to_xml();
assert!(xml.contains("&lt;test&gt;"));
assert!(xml.contains("&amp;"));
}
}

View File

@@ -0,0 +1,3 @@
pub mod constants;
pub mod error;
pub mod types;

View File

@@ -0,0 +1,172 @@
use std::collections::HashMap;
use chrono::{DateTime, Utc};
use serde::{Deserialize, Serialize};
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ObjectMeta {
pub key: String,
pub size: u64,
pub last_modified: DateTime<Utc>,
pub etag: Option<String>,
pub content_type: Option<String>,
pub storage_class: Option<String>,
pub metadata: HashMap<String, String>,
}
impl ObjectMeta {
pub fn new(key: String, size: u64, last_modified: DateTime<Utc>) -> Self {
Self {
key,
size,
last_modified,
etag: None,
content_type: None,
storage_class: Some("STANDARD".to_string()),
metadata: HashMap::new(),
}
}
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct BucketMeta {
pub name: String,
pub creation_date: DateTime<Utc>,
}
#[derive(Debug, Clone, Default)]
pub struct BucketStats {
pub objects: u64,
pub bytes: u64,
pub version_count: u64,
pub version_bytes: u64,
}
impl BucketStats {
pub fn total_objects(&self) -> u64 {
self.objects + self.version_count
}
pub fn total_bytes(&self) -> u64 {
self.bytes + self.version_bytes
}
}
#[derive(Debug, Clone)]
pub struct ListObjectsResult {
pub objects: Vec<ObjectMeta>,
pub is_truncated: bool,
pub next_continuation_token: Option<String>,
}
#[derive(Debug, Clone)]
pub struct ShallowListResult {
pub objects: Vec<ObjectMeta>,
pub common_prefixes: Vec<String>,
pub is_truncated: bool,
pub next_continuation_token: Option<String>,
}
#[derive(Debug, Clone, Default)]
pub struct ListParams {
pub max_keys: usize,
pub continuation_token: Option<String>,
pub prefix: Option<String>,
pub start_after: Option<String>,
}
#[derive(Debug, Clone, Default)]
pub struct ShallowListParams {
pub prefix: String,
pub delimiter: String,
pub max_keys: usize,
pub continuation_token: Option<String>,
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct PartMeta {
pub part_number: u32,
pub etag: String,
pub size: u64,
pub last_modified: Option<DateTime<Utc>>,
}
#[derive(Debug, Clone)]
pub struct PartInfo {
pub part_number: u32,
pub etag: String,
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct MultipartUploadInfo {
pub upload_id: String,
pub key: String,
pub initiated: DateTime<Utc>,
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct VersionInfo {
pub version_id: String,
pub key: String,
pub size: u64,
pub last_modified: DateTime<Utc>,
pub etag: Option<String>,
pub is_latest: bool,
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Tag {
pub key: String,
pub value: String,
}
#[derive(Debug, Clone, Default, Serialize, Deserialize)]
pub struct BucketConfig {
#[serde(default)]
pub versioning_enabled: bool,
#[serde(default)]
pub tags: Vec<Tag>,
#[serde(default)]
pub cors: Option<serde_json::Value>,
#[serde(default)]
pub encryption: Option<serde_json::Value>,
#[serde(default)]
pub lifecycle: Option<serde_json::Value>,
#[serde(default)]
pub website: Option<serde_json::Value>,
#[serde(default)]
pub quota: Option<QuotaConfig>,
#[serde(default)]
pub acl: Option<serde_json::Value>,
#[serde(default)]
pub notification: Option<serde_json::Value>,
#[serde(default)]
pub logging: Option<serde_json::Value>,
#[serde(default)]
pub object_lock: Option<serde_json::Value>,
}
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct QuotaConfig {
pub max_bytes: Option<u64>,
pub max_objects: Option<u64>,
}
#[derive(Debug, Clone)]
pub struct Principal {
pub access_key: String,
pub user_id: String,
pub display_name: String,
pub is_admin: bool,
}
impl Principal {
pub fn new(access_key: String, user_id: String, display_name: String, is_admin: bool) -> Self {
Self {
access_key,
user_id,
display_name,
is_admin,
}
}
}

View File

@@ -0,0 +1,24 @@
[package]
name = "myfsio-crypto"
version = "0.1.0"
edition = "2021"
[dependencies]
myfsio-common = { path = "../myfsio-common" }
md-5 = { workspace = true }
sha2 = { workspace = true }
hex = { workspace = true }
aes-gcm = { workspace = true }
hkdf = { workspace = true }
thiserror = { workspace = true }
tokio = { workspace = true }
serde = { workspace = true }
serde_json = { workspace = true }
uuid = { workspace = true }
chrono = { workspace = true }
base64 = { workspace = true }
rand = "0.8"
[dev-dependencies]
tokio = { workspace = true, features = ["macros", "rt-multi-thread"] }
tempfile = "3"

View File

@@ -0,0 +1,238 @@
use aes_gcm::aead::Aead;
use aes_gcm::{Aes256Gcm, KeyInit, Nonce};
use hkdf::Hkdf;
use sha2::Sha256;
use std::fs::File;
use std::io::{Read, Seek, SeekFrom, Write};
use std::path::Path;
use thiserror::Error;
const DEFAULT_CHUNK_SIZE: usize = 65536;
const HEADER_SIZE: usize = 4;
#[derive(Debug, Error)]
pub enum CryptoError {
#[error("IO error: {0}")]
Io(#[from] std::io::Error),
#[error("Invalid key size: expected 32 bytes, got {0}")]
InvalidKeySize(usize),
#[error("Invalid nonce size: expected 12 bytes, got {0}")]
InvalidNonceSize(usize),
#[error("Encryption failed: {0}")]
EncryptionFailed(String),
#[error("Decryption failed at chunk {0}")]
DecryptionFailed(u32),
#[error("HKDF expand failed: {0}")]
HkdfFailed(String),
}
fn read_exact_chunk(reader: &mut impl Read, buf: &mut [u8]) -> std::io::Result<usize> {
let mut filled = 0;
while filled < buf.len() {
match reader.read(&mut buf[filled..]) {
Ok(0) => break,
Ok(n) => filled += n,
Err(ref e) if e.kind() == std::io::ErrorKind::Interrupted => continue,
Err(e) => return Err(e),
}
}
Ok(filled)
}
fn derive_chunk_nonce(base_nonce: &[u8], chunk_index: u32) -> Result<[u8; 12], CryptoError> {
let hkdf = Hkdf::<Sha256>::new(Some(base_nonce), b"chunk_nonce");
let mut okm = [0u8; 12];
hkdf.expand(&chunk_index.to_be_bytes(), &mut okm)
.map_err(|e| CryptoError::HkdfFailed(e.to_string()))?;
Ok(okm)
}
pub fn encrypt_stream_chunked(
input_path: &Path,
output_path: &Path,
key: &[u8],
base_nonce: &[u8],
chunk_size: Option<usize>,
) -> Result<u32, CryptoError> {
if key.len() != 32 {
return Err(CryptoError::InvalidKeySize(key.len()));
}
if base_nonce.len() != 12 {
return Err(CryptoError::InvalidNonceSize(base_nonce.len()));
}
let chunk_size = chunk_size.unwrap_or(DEFAULT_CHUNK_SIZE);
let key_arr: [u8; 32] = key.try_into().unwrap();
let nonce_arr: [u8; 12] = base_nonce.try_into().unwrap();
let cipher = Aes256Gcm::new(&key_arr.into());
let mut infile = File::open(input_path)?;
let mut outfile = File::create(output_path)?;
outfile.write_all(&[0u8; 4])?;
let mut buf = vec![0u8; chunk_size];
let mut chunk_index: u32 = 0;
loop {
let n = read_exact_chunk(&mut infile, &mut buf)?;
if n == 0 {
break;
}
let nonce_bytes = derive_chunk_nonce(&nonce_arr, chunk_index)?;
let nonce = Nonce::from_slice(&nonce_bytes);
let encrypted = cipher
.encrypt(nonce, &buf[..n])
.map_err(|e| CryptoError::EncryptionFailed(e.to_string()))?;
let size = encrypted.len() as u32;
outfile.write_all(&size.to_be_bytes())?;
outfile.write_all(&encrypted)?;
chunk_index += 1;
}
outfile.seek(SeekFrom::Start(0))?;
outfile.write_all(&chunk_index.to_be_bytes())?;
Ok(chunk_index)
}
pub fn decrypt_stream_chunked(
input_path: &Path,
output_path: &Path,
key: &[u8],
base_nonce: &[u8],
) -> Result<u32, CryptoError> {
if key.len() != 32 {
return Err(CryptoError::InvalidKeySize(key.len()));
}
if base_nonce.len() != 12 {
return Err(CryptoError::InvalidNonceSize(base_nonce.len()));
}
let key_arr: [u8; 32] = key.try_into().unwrap();
let nonce_arr: [u8; 12] = base_nonce.try_into().unwrap();
let cipher = Aes256Gcm::new(&key_arr.into());
let mut infile = File::open(input_path)?;
let mut outfile = File::create(output_path)?;
let mut header = [0u8; HEADER_SIZE];
infile.read_exact(&mut header)?;
let chunk_count = u32::from_be_bytes(header);
let mut size_buf = [0u8; HEADER_SIZE];
for chunk_index in 0..chunk_count {
infile.read_exact(&mut size_buf)?;
let chunk_size = u32::from_be_bytes(size_buf) as usize;
let mut encrypted = vec![0u8; chunk_size];
infile.read_exact(&mut encrypted)?;
let nonce_bytes = derive_chunk_nonce(&nonce_arr, chunk_index)?;
let nonce = Nonce::from_slice(&nonce_bytes);
let decrypted = cipher
.decrypt(nonce, encrypted.as_ref())
.map_err(|_| CryptoError::DecryptionFailed(chunk_index))?;
outfile.write_all(&decrypted)?;
}
Ok(chunk_count)
}
pub async fn encrypt_stream_chunked_async(
input_path: &Path,
output_path: &Path,
key: &[u8],
base_nonce: &[u8],
chunk_size: Option<usize>,
) -> Result<u32, CryptoError> {
let input_path = input_path.to_owned();
let output_path = output_path.to_owned();
let key = key.to_vec();
let base_nonce = base_nonce.to_vec();
tokio::task::spawn_blocking(move || {
encrypt_stream_chunked(&input_path, &output_path, &key, &base_nonce, chunk_size)
})
.await
.map_err(|e| CryptoError::Io(std::io::Error::new(std::io::ErrorKind::Other, e)))?
}
pub async fn decrypt_stream_chunked_async(
input_path: &Path,
output_path: &Path,
key: &[u8],
base_nonce: &[u8],
) -> Result<u32, CryptoError> {
let input_path = input_path.to_owned();
let output_path = output_path.to_owned();
let key = key.to_vec();
let base_nonce = base_nonce.to_vec();
tokio::task::spawn_blocking(move || {
decrypt_stream_chunked(&input_path, &output_path, &key, &base_nonce)
})
.await
.map_err(|e| CryptoError::Io(std::io::Error::new(std::io::ErrorKind::Other, e)))?
}
#[cfg(test)]
mod tests {
use super::*;
use std::io::Write as IoWrite;
#[test]
fn test_encrypt_decrypt_roundtrip() {
let dir = tempfile::tempdir().unwrap();
let input = dir.path().join("input.bin");
let encrypted = dir.path().join("encrypted.bin");
let decrypted = dir.path().join("decrypted.bin");
let data = b"Hello, this is a test of AES-256-GCM chunked encryption!";
std::fs::File::create(&input).unwrap().write_all(data).unwrap();
let key = [0x42u8; 32];
let nonce = [0x01u8; 12];
let chunks = encrypt_stream_chunked(&input, &encrypted, &key, &nonce, Some(16)).unwrap();
assert!(chunks > 0);
let chunks2 = decrypt_stream_chunked(&encrypted, &decrypted, &key, &nonce).unwrap();
assert_eq!(chunks, chunks2);
let result = std::fs::read(&decrypted).unwrap();
assert_eq!(result, data);
}
#[test]
fn test_invalid_key_size() {
let dir = tempfile::tempdir().unwrap();
let input = dir.path().join("input.bin");
std::fs::File::create(&input).unwrap().write_all(b"test").unwrap();
let result = encrypt_stream_chunked(&input, &dir.path().join("out"), &[0u8; 16], &[0u8; 12], None);
assert!(matches!(result, Err(CryptoError::InvalidKeySize(16))));
}
#[test]
fn test_wrong_key_fails_decrypt() {
let dir = tempfile::tempdir().unwrap();
let input = dir.path().join("input.bin");
let encrypted = dir.path().join("encrypted.bin");
let decrypted = dir.path().join("decrypted.bin");
std::fs::File::create(&input).unwrap().write_all(b"secret data").unwrap();
let key = [0x42u8; 32];
let nonce = [0x01u8; 12];
encrypt_stream_chunked(&input, &encrypted, &key, &nonce, None).unwrap();
let wrong_key = [0x43u8; 32];
let result = decrypt_stream_chunked(&encrypted, &decrypted, &wrong_key, &nonce);
assert!(matches!(result, Err(CryptoError::DecryptionFailed(_))));
}
}

View File

@@ -0,0 +1,375 @@
use base64::engine::general_purpose::STANDARD as B64;
use base64::Engine;
use rand::RngCore;
use std::collections::HashMap;
use std::path::Path;
use crate::aes_gcm::{
encrypt_stream_chunked, decrypt_stream_chunked, CryptoError,
};
use crate::kms::KmsService;
#[derive(Debug, Clone, PartialEq)]
pub enum SseAlgorithm {
Aes256,
AwsKms,
CustomerProvided,
}
impl SseAlgorithm {
pub fn as_str(&self) -> &'static str {
match self {
SseAlgorithm::Aes256 => "AES256",
SseAlgorithm::AwsKms => "aws:kms",
SseAlgorithm::CustomerProvided => "AES256",
}
}
}
#[derive(Debug, Clone)]
pub struct EncryptionContext {
pub algorithm: SseAlgorithm,
pub kms_key_id: Option<String>,
pub customer_key: Option<Vec<u8>>,
}
#[derive(Debug, Clone)]
pub struct EncryptionMetadata {
pub algorithm: String,
pub nonce: String,
pub encrypted_data_key: Option<String>,
pub kms_key_id: Option<String>,
}
impl EncryptionMetadata {
pub fn to_metadata_map(&self) -> HashMap<String, String> {
let mut map = HashMap::new();
map.insert(
"x-amz-server-side-encryption".to_string(),
self.algorithm.clone(),
);
map.insert("x-amz-encryption-nonce".to_string(), self.nonce.clone());
if let Some(ref dk) = self.encrypted_data_key {
map.insert("x-amz-encrypted-data-key".to_string(), dk.clone());
}
if let Some(ref kid) = self.kms_key_id {
map.insert("x-amz-encryption-key-id".to_string(), kid.clone());
}
map
}
pub fn from_metadata(meta: &HashMap<String, String>) -> Option<Self> {
let algorithm = meta.get("x-amz-server-side-encryption")?;
let nonce = meta.get("x-amz-encryption-nonce")?;
Some(Self {
algorithm: algorithm.clone(),
nonce: nonce.clone(),
encrypted_data_key: meta.get("x-amz-encrypted-data-key").cloned(),
kms_key_id: meta.get("x-amz-encryption-key-id").cloned(),
})
}
pub fn is_encrypted(meta: &HashMap<String, String>) -> bool {
meta.contains_key("x-amz-server-side-encryption")
}
pub fn clean_metadata(meta: &mut HashMap<String, String>) {
meta.remove("x-amz-server-side-encryption");
meta.remove("x-amz-encryption-nonce");
meta.remove("x-amz-encrypted-data-key");
meta.remove("x-amz-encryption-key-id");
}
}
pub struct EncryptionService {
master_key: [u8; 32],
kms: Option<std::sync::Arc<KmsService>>,
}
impl EncryptionService {
pub fn new(master_key: [u8; 32], kms: Option<std::sync::Arc<KmsService>>) -> Self {
Self { master_key, kms }
}
pub fn generate_data_key(&self) -> ([u8; 32], [u8; 12]) {
let mut data_key = [0u8; 32];
let mut nonce = [0u8; 12];
rand::thread_rng().fill_bytes(&mut data_key);
rand::thread_rng().fill_bytes(&mut nonce);
(data_key, nonce)
}
pub fn wrap_data_key(&self, data_key: &[u8; 32]) -> Result<String, CryptoError> {
use aes_gcm::aead::Aead;
use aes_gcm::{Aes256Gcm, KeyInit, Nonce};
let cipher = Aes256Gcm::new((&self.master_key).into());
let mut nonce_bytes = [0u8; 12];
rand::thread_rng().fill_bytes(&mut nonce_bytes);
let nonce = Nonce::from_slice(&nonce_bytes);
let encrypted = cipher
.encrypt(nonce, data_key.as_slice())
.map_err(|e| CryptoError::EncryptionFailed(e.to_string()))?;
let mut combined = Vec::with_capacity(12 + encrypted.len());
combined.extend_from_slice(&nonce_bytes);
combined.extend_from_slice(&encrypted);
Ok(B64.encode(&combined))
}
pub fn unwrap_data_key(&self, wrapped_b64: &str) -> Result<[u8; 32], CryptoError> {
use aes_gcm::aead::Aead;
use aes_gcm::{Aes256Gcm, KeyInit, Nonce};
let combined = B64.decode(wrapped_b64).map_err(|e| {
CryptoError::EncryptionFailed(format!("Bad wrapped key encoding: {}", e))
})?;
if combined.len() < 12 {
return Err(CryptoError::EncryptionFailed(
"Wrapped key too short".to_string(),
));
}
let (nonce_bytes, ciphertext) = combined.split_at(12);
let cipher = Aes256Gcm::new((&self.master_key).into());
let nonce = Nonce::from_slice(nonce_bytes);
let plaintext = cipher
.decrypt(nonce, ciphertext)
.map_err(|_| CryptoError::DecryptionFailed(0))?;
if plaintext.len() != 32 {
return Err(CryptoError::InvalidKeySize(plaintext.len()));
}
let mut key = [0u8; 32];
key.copy_from_slice(&plaintext);
Ok(key)
}
pub async fn encrypt_object(
&self,
input_path: &Path,
output_path: &Path,
ctx: &EncryptionContext,
) -> Result<EncryptionMetadata, CryptoError> {
let (data_key, nonce) = self.generate_data_key();
let (encrypted_data_key, kms_key_id) = match ctx.algorithm {
SseAlgorithm::Aes256 => {
let wrapped = self.wrap_data_key(&data_key)?;
(Some(wrapped), None)
}
SseAlgorithm::AwsKms => {
let kms = self
.kms
.as_ref()
.ok_or_else(|| CryptoError::EncryptionFailed("KMS not available".into()))?;
let kid = ctx
.kms_key_id
.as_ref()
.ok_or_else(|| CryptoError::EncryptionFailed("No KMS key ID".into()))?;
let ciphertext = kms.encrypt_data(kid, &data_key).await?;
(Some(B64.encode(&ciphertext)), Some(kid.clone()))
}
SseAlgorithm::CustomerProvided => {
(None, None)
}
};
let actual_key = if ctx.algorithm == SseAlgorithm::CustomerProvided {
let ck = ctx.customer_key.as_ref().ok_or_else(|| {
CryptoError::EncryptionFailed("No customer key provided".into())
})?;
if ck.len() != 32 {
return Err(CryptoError::InvalidKeySize(ck.len()));
}
let mut k = [0u8; 32];
k.copy_from_slice(ck);
k
} else {
data_key
};
let ip = input_path.to_owned();
let op = output_path.to_owned();
let ak = actual_key;
let n = nonce;
tokio::task::spawn_blocking(move || {
encrypt_stream_chunked(&ip, &op, &ak, &n, None)
})
.await
.map_err(|e| CryptoError::Io(std::io::Error::new(std::io::ErrorKind::Other, e)))??;
Ok(EncryptionMetadata {
algorithm: ctx.algorithm.as_str().to_string(),
nonce: B64.encode(nonce),
encrypted_data_key,
kms_key_id,
})
}
pub async fn decrypt_object(
&self,
input_path: &Path,
output_path: &Path,
enc_meta: &EncryptionMetadata,
customer_key: Option<&[u8]>,
) -> Result<(), CryptoError> {
let nonce_bytes = B64.decode(&enc_meta.nonce).map_err(|e| {
CryptoError::EncryptionFailed(format!("Bad nonce encoding: {}", e))
})?;
if nonce_bytes.len() != 12 {
return Err(CryptoError::InvalidNonceSize(nonce_bytes.len()));
}
let data_key: [u8; 32] = if let Some(ck) = customer_key {
if ck.len() != 32 {
return Err(CryptoError::InvalidKeySize(ck.len()));
}
let mut k = [0u8; 32];
k.copy_from_slice(ck);
k
} else if enc_meta.algorithm == "aws:kms" {
let kms = self
.kms
.as_ref()
.ok_or_else(|| CryptoError::EncryptionFailed("KMS not available".into()))?;
let kid = enc_meta
.kms_key_id
.as_ref()
.ok_or_else(|| CryptoError::EncryptionFailed("No KMS key ID in metadata".into()))?;
let encrypted_dk = enc_meta.encrypted_data_key.as_ref().ok_or_else(|| {
CryptoError::EncryptionFailed("No encrypted data key in metadata".into())
})?;
let ct = B64.decode(encrypted_dk).map_err(|e| {
CryptoError::EncryptionFailed(format!("Bad data key encoding: {}", e))
})?;
let dk = kms.decrypt_data(kid, &ct).await?;
if dk.len() != 32 {
return Err(CryptoError::InvalidKeySize(dk.len()));
}
let mut k = [0u8; 32];
k.copy_from_slice(&dk);
k
} else {
let wrapped = enc_meta.encrypted_data_key.as_ref().ok_or_else(|| {
CryptoError::EncryptionFailed("No encrypted data key in metadata".into())
})?;
self.unwrap_data_key(wrapped)?
};
let ip = input_path.to_owned();
let op = output_path.to_owned();
let nb: [u8; 12] = nonce_bytes.try_into().unwrap();
tokio::task::spawn_blocking(move || {
decrypt_stream_chunked(&ip, &op, &data_key, &nb)
})
.await
.map_err(|e| CryptoError::Io(std::io::Error::new(std::io::ErrorKind::Other, e)))??;
Ok(())
}
}
#[cfg(test)]
mod tests {
use super::*;
use std::io::Write;
fn test_master_key() -> [u8; 32] {
[0x42u8; 32]
}
#[test]
fn test_wrap_unwrap_data_key() {
let svc = EncryptionService::new(test_master_key(), None);
let dk = [0xAAu8; 32];
let wrapped = svc.wrap_data_key(&dk).unwrap();
let unwrapped = svc.unwrap_data_key(&wrapped).unwrap();
assert_eq!(dk, unwrapped);
}
#[tokio::test]
async fn test_encrypt_decrypt_object_sse_s3() {
let dir = tempfile::tempdir().unwrap();
let input = dir.path().join("plain.bin");
let encrypted = dir.path().join("enc.bin");
let decrypted = dir.path().join("dec.bin");
let data = b"SSE-S3 encrypted content for testing!";
std::fs::File::create(&input).unwrap().write_all(data).unwrap();
let svc = EncryptionService::new(test_master_key(), None);
let ctx = EncryptionContext {
algorithm: SseAlgorithm::Aes256,
kms_key_id: None,
customer_key: None,
};
let meta = svc.encrypt_object(&input, &encrypted, &ctx).await.unwrap();
assert_eq!(meta.algorithm, "AES256");
assert!(meta.encrypted_data_key.is_some());
svc.decrypt_object(&encrypted, &decrypted, &meta, None)
.await
.unwrap();
let result = std::fs::read(&decrypted).unwrap();
assert_eq!(result, data);
}
#[tokio::test]
async fn test_encrypt_decrypt_object_sse_c() {
let dir = tempfile::tempdir().unwrap();
let input = dir.path().join("plain.bin");
let encrypted = dir.path().join("enc.bin");
let decrypted = dir.path().join("dec.bin");
let data = b"SSE-C encrypted content!";
std::fs::File::create(&input).unwrap().write_all(data).unwrap();
let customer_key = [0xBBu8; 32];
let svc = EncryptionService::new(test_master_key(), None);
let ctx = EncryptionContext {
algorithm: SseAlgorithm::CustomerProvided,
kms_key_id: None,
customer_key: Some(customer_key.to_vec()),
};
let meta = svc.encrypt_object(&input, &encrypted, &ctx).await.unwrap();
assert!(meta.encrypted_data_key.is_none());
svc.decrypt_object(&encrypted, &decrypted, &meta, Some(&customer_key))
.await
.unwrap();
let result = std::fs::read(&decrypted).unwrap();
assert_eq!(result, data);
}
#[test]
fn test_encryption_metadata_roundtrip() {
let meta = EncryptionMetadata {
algorithm: "AES256".to_string(),
nonce: "dGVzdG5vbmNlMTI=".to_string(),
encrypted_data_key: Some("c29tZWtleQ==".to_string()),
kms_key_id: None,
};
let map = meta.to_metadata_map();
let restored = EncryptionMetadata::from_metadata(&map).unwrap();
assert_eq!(restored.algorithm, "AES256");
assert_eq!(restored.nonce, meta.nonce);
assert_eq!(restored.encrypted_data_key, meta.encrypted_data_key);
}
#[test]
fn test_is_encrypted() {
let mut meta = HashMap::new();
assert!(!EncryptionMetadata::is_encrypted(&meta));
meta.insert("x-amz-server-side-encryption".to_string(), "AES256".to_string());
assert!(EncryptionMetadata::is_encrypted(&meta));
}
}

View File

@@ -0,0 +1,132 @@
use md5::{Digest, Md5};
use sha2::Sha256;
use std::io::Read;
use std::path::Path;
const CHUNK_SIZE: usize = 65536;
pub fn md5_file(path: &Path) -> std::io::Result<String> {
let mut file = std::fs::File::open(path)?;
let mut hasher = Md5::new();
let mut buf = vec![0u8; CHUNK_SIZE];
loop {
let n = file.read(&mut buf)?;
if n == 0 {
break;
}
hasher.update(&buf[..n]);
}
Ok(format!("{:x}", hasher.finalize()))
}
pub fn md5_bytes(data: &[u8]) -> String {
let mut hasher = Md5::new();
hasher.update(data);
format!("{:x}", hasher.finalize())
}
pub fn sha256_file(path: &Path) -> std::io::Result<String> {
let mut file = std::fs::File::open(path)?;
let mut hasher = Sha256::new();
let mut buf = vec![0u8; CHUNK_SIZE];
loop {
let n = file.read(&mut buf)?;
if n == 0 {
break;
}
hasher.update(&buf[..n]);
}
Ok(format!("{:x}", hasher.finalize()))
}
pub fn sha256_bytes(data: &[u8]) -> String {
let mut hasher = Sha256::new();
hasher.update(data);
format!("{:x}", hasher.finalize())
}
pub fn md5_sha256_file(path: &Path) -> std::io::Result<(String, String)> {
let mut file = std::fs::File::open(path)?;
let mut md5_hasher = Md5::new();
let mut sha_hasher = Sha256::new();
let mut buf = vec![0u8; CHUNK_SIZE];
loop {
let n = file.read(&mut buf)?;
if n == 0 {
break;
}
md5_hasher.update(&buf[..n]);
sha_hasher.update(&buf[..n]);
}
Ok((
format!("{:x}", md5_hasher.finalize()),
format!("{:x}", sha_hasher.finalize()),
))
}
pub async fn md5_file_async(path: &Path) -> std::io::Result<String> {
let path = path.to_owned();
tokio::task::spawn_blocking(move || md5_file(&path))
.await
.map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?
}
pub async fn sha256_file_async(path: &Path) -> std::io::Result<String> {
let path = path.to_owned();
tokio::task::spawn_blocking(move || sha256_file(&path))
.await
.map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?
}
pub async fn md5_sha256_file_async(path: &Path) -> std::io::Result<(String, String)> {
let path = path.to_owned();
tokio::task::spawn_blocking(move || md5_sha256_file(&path))
.await
.map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?
}
#[cfg(test)]
mod tests {
use super::*;
use std::io::Write;
#[test]
fn test_md5_bytes() {
assert_eq!(md5_bytes(b""), "d41d8cd98f00b204e9800998ecf8427e");
assert_eq!(md5_bytes(b"hello"), "5d41402abc4b2a76b9719d911017c592");
}
#[test]
fn test_sha256_bytes() {
let hash = sha256_bytes(b"hello");
assert_eq!(hash, "2cf24dba5fb0a30e26e83b2ac5b9e29e1b161e5c1fa7425e73043362938b9824");
}
#[test]
fn test_md5_file() {
let mut tmp = tempfile::NamedTempFile::new().unwrap();
tmp.write_all(b"hello").unwrap();
tmp.flush().unwrap();
let hash = md5_file(tmp.path()).unwrap();
assert_eq!(hash, "5d41402abc4b2a76b9719d911017c592");
}
#[test]
fn test_md5_sha256_file() {
let mut tmp = tempfile::NamedTempFile::new().unwrap();
tmp.write_all(b"hello").unwrap();
tmp.flush().unwrap();
let (md5, sha) = md5_sha256_file(tmp.path()).unwrap();
assert_eq!(md5, "5d41402abc4b2a76b9719d911017c592");
assert_eq!(sha, "2cf24dba5fb0a30e26e83b2ac5b9e29e1b161e5c1fa7425e73043362938b9824");
}
#[tokio::test]
async fn test_md5_file_async() {
let mut tmp = tempfile::NamedTempFile::new().unwrap();
tmp.write_all(b"hello").unwrap();
tmp.flush().unwrap();
let hash = md5_file_async(tmp.path()).await.unwrap();
assert_eq!(hash, "5d41402abc4b2a76b9719d911017c592");
}
}

View File

@@ -0,0 +1,453 @@
use aes_gcm::aead::Aead;
use aes_gcm::{Aes256Gcm, KeyInit, Nonce};
use base64::engine::general_purpose::STANDARD as B64;
use base64::Engine;
use chrono::{DateTime, Utc};
use rand::RngCore;
use serde::{Deserialize, Serialize};
use std::path::{Path, PathBuf};
use std::sync::Arc;
use tokio::sync::RwLock;
use crate::aes_gcm::CryptoError;
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct KmsKey {
#[serde(rename = "KeyId")]
pub key_id: String,
#[serde(rename = "Arn")]
pub arn: String,
#[serde(rename = "Description")]
pub description: String,
#[serde(rename = "CreationDate")]
pub creation_date: DateTime<Utc>,
#[serde(rename = "Enabled")]
pub enabled: bool,
#[serde(rename = "KeyState")]
pub key_state: String,
#[serde(rename = "KeyUsage")]
pub key_usage: String,
#[serde(rename = "KeySpec")]
pub key_spec: String,
#[serde(rename = "EncryptedKeyMaterial")]
pub encrypted_key_material: String,
}
#[derive(Debug, Clone, Serialize, Deserialize)]
struct KmsStore {
keys: Vec<KmsKey>,
}
pub struct KmsService {
keys_path: PathBuf,
master_key: Arc<RwLock<[u8; 32]>>,
keys: Arc<RwLock<Vec<KmsKey>>>,
}
impl KmsService {
pub async fn new(keys_dir: &Path) -> Result<Self, CryptoError> {
std::fs::create_dir_all(keys_dir).map_err(CryptoError::Io)?;
let keys_path = keys_dir.join("kms_keys.json");
let master_key = Self::load_or_create_master_key(&keys_dir.join("kms_master.key"))?;
let keys = if keys_path.exists() {
let data = std::fs::read_to_string(&keys_path).map_err(CryptoError::Io)?;
let store: KmsStore = serde_json::from_str(&data)
.map_err(|e| CryptoError::EncryptionFailed(format!("Bad KMS store: {}", e)))?;
store.keys
} else {
Vec::new()
};
Ok(Self {
keys_path,
master_key: Arc::new(RwLock::new(master_key)),
keys: Arc::new(RwLock::new(keys)),
})
}
fn load_or_create_master_key(path: &Path) -> Result<[u8; 32], CryptoError> {
if path.exists() {
let encoded = std::fs::read_to_string(path).map_err(CryptoError::Io)?;
let decoded = B64.decode(encoded.trim()).map_err(|e| {
CryptoError::EncryptionFailed(format!("Bad master key encoding: {}", e))
})?;
if decoded.len() != 32 {
return Err(CryptoError::InvalidKeySize(decoded.len()));
}
let mut key = [0u8; 32];
key.copy_from_slice(&decoded);
Ok(key)
} else {
let mut key = [0u8; 32];
rand::thread_rng().fill_bytes(&mut key);
let encoded = B64.encode(key);
std::fs::write(path, &encoded).map_err(CryptoError::Io)?;
Ok(key)
}
}
fn encrypt_key_material(
master_key: &[u8; 32],
plaintext_key: &[u8],
) -> Result<String, CryptoError> {
let cipher = Aes256Gcm::new(master_key.into());
let mut nonce_bytes = [0u8; 12];
rand::thread_rng().fill_bytes(&mut nonce_bytes);
let nonce = Nonce::from_slice(&nonce_bytes);
let ciphertext = cipher
.encrypt(nonce, plaintext_key)
.map_err(|e| CryptoError::EncryptionFailed(e.to_string()))?;
let mut combined = Vec::with_capacity(12 + ciphertext.len());
combined.extend_from_slice(&nonce_bytes);
combined.extend_from_slice(&ciphertext);
Ok(B64.encode(&combined))
}
fn decrypt_key_material(
master_key: &[u8; 32],
encrypted_b64: &str,
) -> Result<Vec<u8>, CryptoError> {
let combined = B64.decode(encrypted_b64).map_err(|e| {
CryptoError::EncryptionFailed(format!("Bad key material encoding: {}", e))
})?;
if combined.len() < 12 {
return Err(CryptoError::EncryptionFailed(
"Encrypted key material too short".to_string(),
));
}
let (nonce_bytes, ciphertext) = combined.split_at(12);
let cipher = Aes256Gcm::new(master_key.into());
let nonce = Nonce::from_slice(nonce_bytes);
cipher
.decrypt(nonce, ciphertext)
.map_err(|_| CryptoError::DecryptionFailed(0))
}
async fn save(&self) -> Result<(), CryptoError> {
let keys = self.keys.read().await;
let store = KmsStore {
keys: keys.clone(),
};
let json = serde_json::to_string_pretty(&store)
.map_err(|e| CryptoError::EncryptionFailed(e.to_string()))?;
std::fs::write(&self.keys_path, json).map_err(CryptoError::Io)?;
Ok(())
}
pub async fn create_key(&self, description: &str) -> Result<KmsKey, CryptoError> {
let key_id = uuid::Uuid::new_v4().to_string();
let arn = format!("arn:aws:kms:local:000000000000:key/{}", key_id);
let mut plaintext_key = [0u8; 32];
rand::thread_rng().fill_bytes(&mut plaintext_key);
let master = self.master_key.read().await;
let encrypted = Self::encrypt_key_material(&master, &plaintext_key)?;
let kms_key = KmsKey {
key_id: key_id.clone(),
arn,
description: description.to_string(),
creation_date: Utc::now(),
enabled: true,
key_state: "Enabled".to_string(),
key_usage: "ENCRYPT_DECRYPT".to_string(),
key_spec: "SYMMETRIC_DEFAULT".to_string(),
encrypted_key_material: encrypted,
};
self.keys.write().await.push(kms_key.clone());
self.save().await?;
Ok(kms_key)
}
pub async fn list_keys(&self) -> Vec<KmsKey> {
self.keys.read().await.clone()
}
pub async fn get_key(&self, key_id: &str) -> Option<KmsKey> {
let keys = self.keys.read().await;
keys.iter()
.find(|k| k.key_id == key_id || k.arn == key_id)
.cloned()
}
pub async fn delete_key(&self, key_id: &str) -> Result<bool, CryptoError> {
let mut keys = self.keys.write().await;
let len_before = keys.len();
keys.retain(|k| k.key_id != key_id && k.arn != key_id);
let removed = keys.len() < len_before;
drop(keys);
if removed {
self.save().await?;
}
Ok(removed)
}
pub async fn enable_key(&self, key_id: &str) -> Result<bool, CryptoError> {
let mut keys = self.keys.write().await;
if let Some(key) = keys.iter_mut().find(|k| k.key_id == key_id) {
key.enabled = true;
key.key_state = "Enabled".to_string();
drop(keys);
self.save().await?;
Ok(true)
} else {
Ok(false)
}
}
pub async fn disable_key(&self, key_id: &str) -> Result<bool, CryptoError> {
let mut keys = self.keys.write().await;
if let Some(key) = keys.iter_mut().find(|k| k.key_id == key_id) {
key.enabled = false;
key.key_state = "Disabled".to_string();
drop(keys);
self.save().await?;
Ok(true)
} else {
Ok(false)
}
}
pub async fn decrypt_data_key(&self, key_id: &str) -> Result<Vec<u8>, CryptoError> {
let keys = self.keys.read().await;
let key = keys
.iter()
.find(|k| k.key_id == key_id || k.arn == key_id)
.ok_or_else(|| CryptoError::EncryptionFailed("KMS key not found".to_string()))?;
if !key.enabled {
return Err(CryptoError::EncryptionFailed(
"KMS key is disabled".to_string(),
));
}
let master = self.master_key.read().await;
Self::decrypt_key_material(&master, &key.encrypted_key_material)
}
pub async fn encrypt_data(
&self,
key_id: &str,
plaintext: &[u8],
) -> Result<Vec<u8>, CryptoError> {
let data_key = self.decrypt_data_key(key_id).await?;
if data_key.len() != 32 {
return Err(CryptoError::InvalidKeySize(data_key.len()));
}
let key_arr: [u8; 32] = data_key.try_into().unwrap();
let cipher = Aes256Gcm::new(&key_arr.into());
let mut nonce_bytes = [0u8; 12];
rand::thread_rng().fill_bytes(&mut nonce_bytes);
let nonce = Nonce::from_slice(&nonce_bytes);
let ciphertext = cipher
.encrypt(nonce, plaintext)
.map_err(|e| CryptoError::EncryptionFailed(e.to_string()))?;
let mut result = Vec::with_capacity(12 + ciphertext.len());
result.extend_from_slice(&nonce_bytes);
result.extend_from_slice(&ciphertext);
Ok(result)
}
pub async fn decrypt_data(
&self,
key_id: &str,
ciphertext: &[u8],
) -> Result<Vec<u8>, CryptoError> {
if ciphertext.len() < 12 {
return Err(CryptoError::EncryptionFailed(
"Ciphertext too short".to_string(),
));
}
let data_key = self.decrypt_data_key(key_id).await?;
if data_key.len() != 32 {
return Err(CryptoError::InvalidKeySize(data_key.len()));
}
let key_arr: [u8; 32] = data_key.try_into().unwrap();
let (nonce_bytes, ct) = ciphertext.split_at(12);
let cipher = Aes256Gcm::new(&key_arr.into());
let nonce = Nonce::from_slice(nonce_bytes);
cipher
.decrypt(nonce, ct)
.map_err(|_| CryptoError::DecryptionFailed(0))
}
pub async fn generate_data_key(
&self,
key_id: &str,
num_bytes: usize,
) -> Result<(Vec<u8>, Vec<u8>), CryptoError> {
let kms_key = self.decrypt_data_key(key_id).await?;
if kms_key.len() != 32 {
return Err(CryptoError::InvalidKeySize(kms_key.len()));
}
let mut plaintext_key = vec![0u8; num_bytes];
rand::thread_rng().fill_bytes(&mut plaintext_key);
let key_arr: [u8; 32] = kms_key.try_into().unwrap();
let cipher = Aes256Gcm::new(&key_arr.into());
let mut nonce_bytes = [0u8; 12];
rand::thread_rng().fill_bytes(&mut nonce_bytes);
let nonce = Nonce::from_slice(&nonce_bytes);
let encrypted = cipher
.encrypt(nonce, plaintext_key.as_slice())
.map_err(|e| CryptoError::EncryptionFailed(e.to_string()))?;
let mut wrapped = Vec::with_capacity(12 + encrypted.len());
wrapped.extend_from_slice(&nonce_bytes);
wrapped.extend_from_slice(&encrypted);
Ok((plaintext_key, wrapped))
}
}
pub async fn load_or_create_master_key(keys_dir: &Path) -> Result<[u8; 32], CryptoError> {
std::fs::create_dir_all(keys_dir).map_err(CryptoError::Io)?;
let path = keys_dir.join("master.key");
if path.exists() {
let encoded = std::fs::read_to_string(&path).map_err(CryptoError::Io)?;
let decoded = B64.decode(encoded.trim()).map_err(|e| {
CryptoError::EncryptionFailed(format!("Bad master key encoding: {}", e))
})?;
if decoded.len() != 32 {
return Err(CryptoError::InvalidKeySize(decoded.len()));
}
let mut key = [0u8; 32];
key.copy_from_slice(&decoded);
Ok(key)
} else {
let mut key = [0u8; 32];
rand::thread_rng().fill_bytes(&mut key);
let encoded = B64.encode(key);
std::fs::write(&path, &encoded).map_err(CryptoError::Io)?;
Ok(key)
}
}
#[cfg(test)]
mod tests {
use super::*;
#[tokio::test]
async fn test_create_and_list_keys() {
let dir = tempfile::tempdir().unwrap();
let kms = KmsService::new(dir.path()).await.unwrap();
let key = kms.create_key("test key").await.unwrap();
assert!(key.enabled);
assert_eq!(key.description, "test key");
assert!(key.key_id.len() > 0);
let keys = kms.list_keys().await;
assert_eq!(keys.len(), 1);
assert_eq!(keys[0].key_id, key.key_id);
}
#[tokio::test]
async fn test_enable_disable_key() {
let dir = tempfile::tempdir().unwrap();
let kms = KmsService::new(dir.path()).await.unwrap();
let key = kms.create_key("toggle").await.unwrap();
assert!(key.enabled);
kms.disable_key(&key.key_id).await.unwrap();
let k = kms.get_key(&key.key_id).await.unwrap();
assert!(!k.enabled);
kms.enable_key(&key.key_id).await.unwrap();
let k = kms.get_key(&key.key_id).await.unwrap();
assert!(k.enabled);
}
#[tokio::test]
async fn test_delete_key() {
let dir = tempfile::tempdir().unwrap();
let kms = KmsService::new(dir.path()).await.unwrap();
let key = kms.create_key("doomed").await.unwrap();
assert!(kms.delete_key(&key.key_id).await.unwrap());
assert!(kms.get_key(&key.key_id).await.is_none());
assert_eq!(kms.list_keys().await.len(), 0);
}
#[tokio::test]
async fn test_encrypt_decrypt_data() {
let dir = tempfile::tempdir().unwrap();
let kms = KmsService::new(dir.path()).await.unwrap();
let key = kms.create_key("enc-key").await.unwrap();
let plaintext = b"Hello, KMS!";
let ciphertext = kms.encrypt_data(&key.key_id, plaintext).await.unwrap();
assert_ne!(&ciphertext, plaintext);
let decrypted = kms.decrypt_data(&key.key_id, &ciphertext).await.unwrap();
assert_eq!(decrypted, plaintext);
}
#[tokio::test]
async fn test_generate_data_key() {
let dir = tempfile::tempdir().unwrap();
let kms = KmsService::new(dir.path()).await.unwrap();
let key = kms.create_key("data-key-gen").await.unwrap();
let (plaintext, wrapped) = kms.generate_data_key(&key.key_id, 32).await.unwrap();
assert_eq!(plaintext.len(), 32);
assert!(wrapped.len() > 32);
}
#[tokio::test]
async fn test_disabled_key_cannot_encrypt() {
let dir = tempfile::tempdir().unwrap();
let kms = KmsService::new(dir.path()).await.unwrap();
let key = kms.create_key("disabled").await.unwrap();
kms.disable_key(&key.key_id).await.unwrap();
let result = kms.encrypt_data(&key.key_id, b"test").await;
assert!(result.is_err());
}
#[tokio::test]
async fn test_persistence_across_reload() {
let dir = tempfile::tempdir().unwrap();
let key_id = {
let kms = KmsService::new(dir.path()).await.unwrap();
let key = kms.create_key("persistent").await.unwrap();
key.key_id
};
let kms2 = KmsService::new(dir.path()).await.unwrap();
let key = kms2.get_key(&key_id).await;
assert!(key.is_some());
assert_eq!(key.unwrap().description, "persistent");
}
#[tokio::test]
async fn test_master_key_roundtrip() {
let dir = tempfile::tempdir().unwrap();
let key1 = load_or_create_master_key(dir.path()).await.unwrap();
let key2 = load_or_create_master_key(dir.path()).await.unwrap();
assert_eq!(key1, key2);
}
}

View File

@@ -0,0 +1,4 @@
pub mod hashing;
pub mod aes_gcm;
pub mod kms;
pub mod encryption;

View File

@@ -0,0 +1,33 @@
[package]
name = "myfsio-server"
version = "0.1.0"
edition = "2021"
[dependencies]
myfsio-common = { path = "../myfsio-common" }
myfsio-auth = { path = "../myfsio-auth" }
myfsio-crypto = { path = "../myfsio-crypto" }
myfsio-storage = { path = "../myfsio-storage" }
myfsio-xml = { path = "../myfsio-xml" }
base64 = { workspace = true }
axum = { workspace = true }
tokio = { workspace = true }
tower = { workspace = true }
tower-http = { workspace = true }
hyper = { workspace = true }
bytes = { workspace = true }
serde = { workspace = true }
serde_json = { workspace = true }
tracing = { workspace = true }
tracing-subscriber = { workspace = true }
tokio-util = { workspace = true }
chrono = { workspace = true }
uuid = { workspace = true }
futures = { workspace = true }
http-body-util = "0.1"
percent-encoding = { workspace = true }
quick-xml = { workspace = true }
[dev-dependencies]
tempfile = "3"
tower = { workspace = true, features = ["util"] }

View File

@@ -0,0 +1,111 @@
use std::net::SocketAddr;
use std::path::PathBuf;
#[derive(Debug, Clone)]
pub struct ServerConfig {
pub bind_addr: SocketAddr,
pub storage_root: PathBuf,
pub region: String,
pub iam_config_path: PathBuf,
pub sigv4_timestamp_tolerance_secs: u64,
pub presigned_url_min_expiry: u64,
pub presigned_url_max_expiry: u64,
pub secret_key: Option<String>,
pub encryption_enabled: bool,
pub kms_enabled: bool,
pub gc_enabled: bool,
pub integrity_enabled: bool,
pub metrics_enabled: bool,
pub lifecycle_enabled: bool,
}
impl ServerConfig {
pub fn from_env() -> Self {
let host = std::env::var("HOST").unwrap_or_else(|_| "127.0.0.1".to_string());
let port: u16 = std::env::var("PORT")
.unwrap_or_else(|_| "5000".to_string())
.parse()
.unwrap_or(5000);
let storage_root = std::env::var("STORAGE_ROOT")
.unwrap_or_else(|_| "./data".to_string());
let region = std::env::var("AWS_REGION")
.unwrap_or_else(|_| "us-east-1".to_string());
let storage_path = PathBuf::from(&storage_root);
let iam_config_path = std::env::var("IAM_CONFIG")
.map(PathBuf::from)
.unwrap_or_else(|_| {
storage_path.join(".myfsio.sys").join("config").join("iam.json")
});
let sigv4_timestamp_tolerance_secs: u64 = std::env::var("SIGV4_TIMESTAMP_TOLERANCE_SECONDS")
.unwrap_or_else(|_| "900".to_string())
.parse()
.unwrap_or(900);
let presigned_url_min_expiry: u64 = std::env::var("PRESIGNED_URL_MIN_EXPIRY_SECONDS")
.unwrap_or_else(|_| "1".to_string())
.parse()
.unwrap_or(1);
let presigned_url_max_expiry: u64 = std::env::var("PRESIGNED_URL_MAX_EXPIRY_SECONDS")
.unwrap_or_else(|_| "604800".to_string())
.parse()
.unwrap_or(604800);
let secret_key = {
let env_key = std::env::var("SECRET_KEY").ok();
match env_key {
Some(k) if !k.is_empty() && k != "dev-secret-key" => Some(k),
_ => {
let secret_file = storage_path
.join(".myfsio.sys")
.join("config")
.join(".secret");
std::fs::read_to_string(&secret_file).ok().map(|s| s.trim().to_string())
}
}
};
let encryption_enabled = std::env::var("ENCRYPTION_ENABLED")
.unwrap_or_else(|_| "false".to_string())
.to_lowercase() == "true";
let kms_enabled = std::env::var("KMS_ENABLED")
.unwrap_or_else(|_| "false".to_string())
.to_lowercase() == "true";
let gc_enabled = std::env::var("GC_ENABLED")
.unwrap_or_else(|_| "false".to_string())
.to_lowercase() == "true";
let integrity_enabled = std::env::var("INTEGRITY_ENABLED")
.unwrap_or_else(|_| "false".to_string())
.to_lowercase() == "true";
let metrics_enabled = std::env::var("OPERATION_METRICS_ENABLED")
.unwrap_or_else(|_| "false".to_string())
.to_lowercase() == "true";
let lifecycle_enabled = std::env::var("LIFECYCLE_ENABLED")
.unwrap_or_else(|_| "false".to_string())
.to_lowercase() == "true";
Self {
bind_addr: SocketAddr::new(host.parse().unwrap(), port),
storage_root: storage_path,
region,
iam_config_path,
sigv4_timestamp_tolerance_secs,
presigned_url_min_expiry,
presigned_url_max_expiry,
secret_key,
encryption_enabled,
kms_enabled,
gc_enabled,
integrity_enabled,
metrics_enabled,
lifecycle_enabled,
}
}
}

View File

@@ -0,0 +1,623 @@
use axum::body::Body;
use axum::http::StatusCode;
use axum::response::{IntoResponse, Response};
use myfsio_common::error::{S3Error, S3ErrorCode};
use myfsio_storage::traits::StorageEngine;
use crate::state::AppState;
fn xml_response(status: StatusCode, xml: String) -> Response {
(status, [("content-type", "application/xml")], xml).into_response()
}
fn storage_err(err: myfsio_storage::error::StorageError) -> Response {
let s3err = S3Error::from(err);
let status = StatusCode::from_u16(s3err.http_status()).unwrap_or(StatusCode::INTERNAL_SERVER_ERROR);
(status, [("content-type", "application/xml")], s3err.to_xml()).into_response()
}
pub async fn get_versioning(state: &AppState, bucket: &str) -> Response {
match state.storage.is_versioning_enabled(bucket).await {
Ok(enabled) => {
let status_str = if enabled { "Enabled" } else { "Suspended" };
let xml = format!(
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<VersioningConfiguration xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\">\
<Status>{}</Status>\
</VersioningConfiguration>",
status_str
);
xml_response(StatusCode::OK, xml)
}
Err(e) => storage_err(e),
}
}
pub async fn put_versioning(state: &AppState, bucket: &str, body: Body) -> Response {
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(collected) => collected.to_bytes(),
Err(_) => {
return xml_response(
StatusCode::BAD_REQUEST,
S3Error::from_code(S3ErrorCode::MalformedXML).to_xml(),
);
}
};
let xml_str = String::from_utf8_lossy(&body_bytes);
let enabled = xml_str.contains("<Status>Enabled</Status>");
match state.storage.set_versioning(bucket, enabled).await {
Ok(()) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
pub async fn get_tagging(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(config) => {
let mut xml = String::from(
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<Tagging xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\"><TagSet>"
);
for tag in &config.tags {
xml.push_str(&format!(
"<Tag><Key>{}</Key><Value>{}</Value></Tag>",
tag.key, tag.value
));
}
xml.push_str("</TagSet></Tagging>");
xml_response(StatusCode::OK, xml)
}
Err(e) => storage_err(e),
}
}
pub async fn put_tagging(state: &AppState, bucket: &str, body: Body) -> Response {
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(collected) => collected.to_bytes(),
Err(_) => {
return xml_response(
StatusCode::BAD_REQUEST,
S3Error::from_code(S3ErrorCode::MalformedXML).to_xml(),
);
}
};
let xml_str = String::from_utf8_lossy(&body_bytes);
let tags = parse_tagging_xml(&xml_str);
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.tags = tags;
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn delete_tagging(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.tags.clear();
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::NO_CONTENT.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn get_cors(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(config) => {
if let Some(cors) = &config.cors {
xml_response(StatusCode::OK, cors.to_string())
} else {
xml_response(
StatusCode::NOT_FOUND,
S3Error::new(S3ErrorCode::NoSuchKey, "The CORS configuration does not exist").to_xml(),
)
}
}
Err(e) => storage_err(e),
}
}
pub async fn put_cors(state: &AppState, bucket: &str, body: Body) -> Response {
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(collected) => collected.to_bytes(),
Err(_) => return StatusCode::BAD_REQUEST.into_response(),
};
let body_str = String::from_utf8_lossy(&body_bytes);
let value = serde_json::Value::String(body_str.to_string());
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.cors = Some(value);
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn delete_cors(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.cors = None;
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::NO_CONTENT.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn get_location(state: &AppState, _bucket: &str) -> Response {
let xml = format!(
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<LocationConstraint xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\">{}</LocationConstraint>",
state.config.region
);
xml_response(StatusCode::OK, xml)
}
pub async fn get_encryption(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(config) => {
if let Some(enc) = &config.encryption {
xml_response(StatusCode::OK, enc.to_string())
} else {
xml_response(
StatusCode::NOT_FOUND,
S3Error::new(
S3ErrorCode::InvalidRequest,
"The server side encryption configuration was not found",
).to_xml(),
)
}
}
Err(e) => storage_err(e),
}
}
pub async fn put_encryption(state: &AppState, bucket: &str, body: Body) -> Response {
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(collected) => collected.to_bytes(),
Err(_) => return StatusCode::BAD_REQUEST.into_response(),
};
let value = serde_json::Value::String(String::from_utf8_lossy(&body_bytes).to_string());
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.encryption = Some(value);
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn delete_encryption(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.encryption = None;
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::NO_CONTENT.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn get_lifecycle(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(config) => {
if let Some(lc) = &config.lifecycle {
xml_response(StatusCode::OK, lc.to_string())
} else {
xml_response(
StatusCode::NOT_FOUND,
S3Error::new(S3ErrorCode::NoSuchKey, "The lifecycle configuration does not exist").to_xml(),
)
}
}
Err(e) => storage_err(e),
}
}
pub async fn put_lifecycle(state: &AppState, bucket: &str, body: Body) -> Response {
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(collected) => collected.to_bytes(),
Err(_) => return StatusCode::BAD_REQUEST.into_response(),
};
let value = serde_json::Value::String(String::from_utf8_lossy(&body_bytes).to_string());
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.lifecycle = Some(value);
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn delete_lifecycle(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.lifecycle = None;
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::NO_CONTENT.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn get_acl(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(config) => {
if let Some(acl) = &config.acl {
xml_response(StatusCode::OK, acl.to_string())
} else {
let xml = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<AccessControlPolicy xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\">\
<Owner><ID>myfsio</ID><DisplayName>myfsio</DisplayName></Owner>\
<AccessControlList>\
<Grant><Grantee xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\" xsi:type=\"CanonicalUser\">\
<ID>myfsio</ID><DisplayName>myfsio</DisplayName></Grantee>\
<Permission>FULL_CONTROL</Permission></Grant>\
</AccessControlList></AccessControlPolicy>";
xml_response(StatusCode::OK, xml.to_string())
}
}
Err(e) => storage_err(e),
}
}
pub async fn put_acl(state: &AppState, bucket: &str, body: Body) -> Response {
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(collected) => collected.to_bytes(),
Err(_) => return StatusCode::BAD_REQUEST.into_response(),
};
let value = serde_json::Value::String(String::from_utf8_lossy(&body_bytes).to_string());
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.acl = Some(value);
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn get_website(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(config) => {
if let Some(ws) = &config.website {
xml_response(StatusCode::OK, ws.to_string())
} else {
xml_response(
StatusCode::NOT_FOUND,
S3Error::new(S3ErrorCode::NoSuchKey, "The website configuration does not exist").to_xml(),
)
}
}
Err(e) => storage_err(e),
}
}
pub async fn put_website(state: &AppState, bucket: &str, body: Body) -> Response {
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(collected) => collected.to_bytes(),
Err(_) => return StatusCode::BAD_REQUEST.into_response(),
};
let value = serde_json::Value::String(String::from_utf8_lossy(&body_bytes).to_string());
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.website = Some(value);
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn delete_website(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(mut config) => {
config.website = None;
match state.storage.set_bucket_config(bucket, &config).await {
Ok(()) => StatusCode::NO_CONTENT.into_response(),
Err(e) => storage_err(e),
}
}
Err(e) => storage_err(e),
}
}
pub async fn get_object_lock(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(config) => {
if let Some(ol) = &config.object_lock {
xml_response(StatusCode::OK, ol.to_string())
} else {
let xml = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<ObjectLockConfiguration xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\">\
<ObjectLockEnabled>Disabled</ObjectLockEnabled>\
</ObjectLockConfiguration>";
xml_response(StatusCode::OK, xml.to_string())
}
}
Err(e) => storage_err(e),
}
}
pub async fn get_notification(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(config) => {
if let Some(n) = &config.notification {
xml_response(StatusCode::OK, n.to_string())
} else {
let xml = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<NotificationConfiguration xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\">\
</NotificationConfiguration>";
xml_response(StatusCode::OK, xml.to_string())
}
}
Err(e) => storage_err(e),
}
}
pub async fn get_logging(state: &AppState, bucket: &str) -> Response {
match state.storage.get_bucket_config(bucket).await {
Ok(config) => {
if let Some(l) = &config.logging {
xml_response(StatusCode::OK, l.to_string())
} else {
let xml = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<BucketLoggingStatus xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\">\
</BucketLoggingStatus>";
xml_response(StatusCode::OK, xml.to_string())
}
}
Err(e) => storage_err(e),
}
}
pub async fn list_object_versions(state: &AppState, bucket: &str) -> Response {
match state.storage.list_buckets().await {
Ok(buckets) => {
if !buckets.iter().any(|b| b.name == bucket) {
return storage_err(myfsio_storage::error::StorageError::BucketNotFound(
bucket.to_string(),
));
}
}
Err(e) => return storage_err(e),
}
let params = myfsio_common::types::ListParams {
max_keys: 1000,
..Default::default()
};
let objects = match state.storage.list_objects(bucket, &params).await {
Ok(result) => result.objects,
Err(e) => return storage_err(e),
};
let mut xml = String::from(
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<ListVersionsResult xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\">"
);
xml.push_str(&format!("<Name>{}</Name>", bucket));
for obj in &objects {
xml.push_str("<Version>");
xml.push_str(&format!("<Key>{}</Key>", obj.key));
xml.push_str("<VersionId>null</VersionId>");
xml.push_str("<IsLatest>true</IsLatest>");
xml.push_str(&format!(
"<LastModified>{}</LastModified>",
obj.last_modified.to_rfc3339()
));
if let Some(ref etag) = obj.etag {
xml.push_str(&format!("<ETag>\"{}\"</ETag>", etag));
}
xml.push_str(&format!("<Size>{}</Size>", obj.size));
xml.push_str("<StorageClass>STANDARD</StorageClass>");
xml.push_str("</Version>");
}
xml.push_str("</ListVersionsResult>");
xml_response(StatusCode::OK, xml)
}
pub async fn get_object_tagging(state: &AppState, bucket: &str, key: &str) -> Response {
match state.storage.get_object_tags(bucket, key).await {
Ok(tags) => {
let mut xml = String::from(
"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<Tagging xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\"><TagSet>"
);
for tag in &tags {
xml.push_str(&format!(
"<Tag><Key>{}</Key><Value>{}</Value></Tag>",
tag.key, tag.value
));
}
xml.push_str("</TagSet></Tagging>");
xml_response(StatusCode::OK, xml)
}
Err(e) => storage_err(e),
}
}
pub async fn put_object_tagging(state: &AppState, bucket: &str, key: &str, body: Body) -> Response {
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(collected) => collected.to_bytes(),
Err(_) => {
return xml_response(
StatusCode::BAD_REQUEST,
S3Error::from_code(S3ErrorCode::MalformedXML).to_xml(),
);
}
};
let xml_str = String::from_utf8_lossy(&body_bytes);
let tags = parse_tagging_xml(&xml_str);
match state.storage.set_object_tags(bucket, key, &tags).await {
Ok(()) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
pub async fn delete_object_tagging(state: &AppState, bucket: &str, key: &str) -> Response {
match state.storage.delete_object_tags(bucket, key).await {
Ok(()) => StatusCode::NO_CONTENT.into_response(),
Err(e) => storage_err(e),
}
}
pub async fn get_object_acl(state: &AppState, bucket: &str, key: &str) -> Response {
match state.storage.head_object(bucket, key).await {
Ok(_) => {
let xml = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<AccessControlPolicy xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\">\
<Owner><ID>myfsio</ID><DisplayName>myfsio</DisplayName></Owner>\
<AccessControlList>\
<Grant><Grantee xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\" xsi:type=\"CanonicalUser\">\
<ID>myfsio</ID><DisplayName>myfsio</DisplayName></Grantee>\
<Permission>FULL_CONTROL</Permission></Grant>\
</AccessControlList></AccessControlPolicy>";
xml_response(StatusCode::OK, xml.to_string())
}
Err(e) => storage_err(e),
}
}
pub async fn put_object_acl(state: &AppState, bucket: &str, key: &str, _body: Body) -> Response {
match state.storage.head_object(bucket, key).await {
Ok(_) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
pub async fn get_object_retention(state: &AppState, bucket: &str, key: &str) -> Response {
match state.storage.head_object(bucket, key).await {
Ok(_) => {
xml_response(
StatusCode::NOT_FOUND,
S3Error::new(
S3ErrorCode::InvalidRequest,
"No retention policy configured",
).to_xml(),
)
}
Err(e) => storage_err(e),
}
}
pub async fn put_object_retention(state: &AppState, bucket: &str, key: &str, _body: Body) -> Response {
match state.storage.head_object(bucket, key).await {
Ok(_) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
pub async fn get_object_legal_hold(state: &AppState, bucket: &str, key: &str) -> Response {
match state.storage.head_object(bucket, key).await {
Ok(_) => {
let xml = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\
<LegalHold xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\">\
<Status>OFF</Status></LegalHold>";
xml_response(StatusCode::OK, xml.to_string())
}
Err(e) => storage_err(e),
}
}
pub async fn put_object_legal_hold(state: &AppState, bucket: &str, key: &str, _body: Body) -> Response {
match state.storage.head_object(bucket, key).await {
Ok(_) => StatusCode::OK.into_response(),
Err(e) => storage_err(e),
}
}
fn parse_tagging_xml(xml: &str) -> Vec<myfsio_common::types::Tag> {
let mut tags = Vec::new();
let mut in_tag = false;
let mut current_key = String::new();
let mut current_value = String::new();
let mut current_element = String::new();
let mut reader = quick_xml::Reader::from_str(xml);
let mut buf = Vec::new();
loop {
match reader.read_event_into(&mut buf) {
Ok(quick_xml::events::Event::Start(ref e)) => {
let name = String::from_utf8_lossy(e.name().as_ref()).to_string();
current_element = name.clone();
if name == "Tag" {
in_tag = true;
current_key.clear();
current_value.clear();
}
}
Ok(quick_xml::events::Event::Text(ref e)) => {
if in_tag {
let text = e.unescape().unwrap_or_default().to_string();
match current_element.as_str() {
"Key" => current_key = text,
"Value" => current_value = text,
_ => {}
}
}
}
Ok(quick_xml::events::Event::End(ref e)) => {
let name = String::from_utf8_lossy(e.name().as_ref()).to_string();
if name == "Tag" && in_tag {
if !current_key.is_empty() {
tags.push(myfsio_common::types::Tag {
key: current_key.clone(),
value: current_value.clone(),
});
}
in_tag = false;
}
}
Ok(quick_xml::events::Event::Eof) => break,
Err(_) => break,
_ => {}
}
buf.clear();
}
tags
}

View File

@@ -0,0 +1,278 @@
use axum::body::Body;
use axum::extract::State;
use axum::http::StatusCode;
use axum::response::{IntoResponse, Response};
use base64::engine::general_purpose::STANDARD as B64;
use base64::Engine;
use serde_json::json;
use crate::state::AppState;
fn json_ok(value: serde_json::Value) -> Response {
(
StatusCode::OK,
[("content-type", "application/json")],
value.to_string(),
)
.into_response()
}
fn json_err(status: StatusCode, msg: &str) -> Response {
(
status,
[("content-type", "application/json")],
json!({"error": msg}).to_string(),
)
.into_response()
}
pub async fn list_keys(State(state): State<AppState>) -> Response {
let kms = match &state.kms {
Some(k) => k,
None => return json_err(StatusCode::SERVICE_UNAVAILABLE, "KMS not enabled"),
};
let keys = kms.list_keys().await;
let keys_json: Vec<serde_json::Value> = keys
.iter()
.map(|k| {
json!({
"KeyId": k.key_id,
"Arn": k.arn,
"Description": k.description,
"CreationDate": k.creation_date.to_rfc3339(),
"Enabled": k.enabled,
"KeyState": k.key_state,
"KeyUsage": k.key_usage,
"KeySpec": k.key_spec,
})
})
.collect();
json_ok(json!({"keys": keys_json}))
}
pub async fn create_key(State(state): State<AppState>, body: Body) -> Response {
let kms = match &state.kms {
Some(k) => k,
None => return json_err(StatusCode::SERVICE_UNAVAILABLE, "KMS not enabled"),
};
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(c) => c.to_bytes(),
Err(_) => return json_err(StatusCode::BAD_REQUEST, "Invalid request body"),
};
let description = if body_bytes.is_empty() {
String::new()
} else {
match serde_json::from_slice::<serde_json::Value>(&body_bytes) {
Ok(v) => v
.get("Description")
.or_else(|| v.get("description"))
.and_then(|d| d.as_str())
.unwrap_or("")
.to_string(),
Err(_) => String::new(),
}
};
match kms.create_key(&description).await {
Ok(key) => json_ok(json!({
"KeyId": key.key_id,
"Arn": key.arn,
"Description": key.description,
"CreationDate": key.creation_date.to_rfc3339(),
"Enabled": key.enabled,
"KeyState": key.key_state,
})),
Err(e) => json_err(StatusCode::INTERNAL_SERVER_ERROR, &e.to_string()),
}
}
pub async fn get_key(
State(state): State<AppState>,
axum::extract::Path(key_id): axum::extract::Path<String>,
) -> Response {
let kms = match &state.kms {
Some(k) => k,
None => return json_err(StatusCode::SERVICE_UNAVAILABLE, "KMS not enabled"),
};
match kms.get_key(&key_id).await {
Some(key) => json_ok(json!({
"KeyId": key.key_id,
"Arn": key.arn,
"Description": key.description,
"CreationDate": key.creation_date.to_rfc3339(),
"Enabled": key.enabled,
"KeyState": key.key_state,
"KeyUsage": key.key_usage,
"KeySpec": key.key_spec,
})),
None => json_err(StatusCode::NOT_FOUND, "Key not found"),
}
}
pub async fn delete_key(
State(state): State<AppState>,
axum::extract::Path(key_id): axum::extract::Path<String>,
) -> Response {
let kms = match &state.kms {
Some(k) => k,
None => return json_err(StatusCode::SERVICE_UNAVAILABLE, "KMS not enabled"),
};
match kms.delete_key(&key_id).await {
Ok(true) => StatusCode::NO_CONTENT.into_response(),
Ok(false) => json_err(StatusCode::NOT_FOUND, "Key not found"),
Err(e) => json_err(StatusCode::INTERNAL_SERVER_ERROR, &e.to_string()),
}
}
pub async fn enable_key(
State(state): State<AppState>,
axum::extract::Path(key_id): axum::extract::Path<String>,
) -> Response {
let kms = match &state.kms {
Some(k) => k,
None => return json_err(StatusCode::SERVICE_UNAVAILABLE, "KMS not enabled"),
};
match kms.enable_key(&key_id).await {
Ok(true) => json_ok(json!({"status": "enabled"})),
Ok(false) => json_err(StatusCode::NOT_FOUND, "Key not found"),
Err(e) => json_err(StatusCode::INTERNAL_SERVER_ERROR, &e.to_string()),
}
}
pub async fn disable_key(
State(state): State<AppState>,
axum::extract::Path(key_id): axum::extract::Path<String>,
) -> Response {
let kms = match &state.kms {
Some(k) => k,
None => return json_err(StatusCode::SERVICE_UNAVAILABLE, "KMS not enabled"),
};
match kms.disable_key(&key_id).await {
Ok(true) => json_ok(json!({"status": "disabled"})),
Ok(false) => json_err(StatusCode::NOT_FOUND, "Key not found"),
Err(e) => json_err(StatusCode::INTERNAL_SERVER_ERROR, &e.to_string()),
}
}
pub async fn encrypt(State(state): State<AppState>, body: Body) -> Response {
let kms = match &state.kms {
Some(k) => k,
None => return json_err(StatusCode::SERVICE_UNAVAILABLE, "KMS not enabled"),
};
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(c) => c.to_bytes(),
Err(_) => return json_err(StatusCode::BAD_REQUEST, "Invalid request body"),
};
let req: serde_json::Value = match serde_json::from_slice(&body_bytes) {
Ok(v) => v,
Err(_) => return json_err(StatusCode::BAD_REQUEST, "Invalid JSON"),
};
let key_id = match req.get("KeyId").and_then(|v| v.as_str()) {
Some(k) => k,
None => return json_err(StatusCode::BAD_REQUEST, "Missing KeyId"),
};
let plaintext_b64 = match req.get("Plaintext").and_then(|v| v.as_str()) {
Some(p) => p,
None => return json_err(StatusCode::BAD_REQUEST, "Missing Plaintext"),
};
let plaintext = match B64.decode(plaintext_b64) {
Ok(p) => p,
Err(_) => return json_err(StatusCode::BAD_REQUEST, "Invalid base64 Plaintext"),
};
match kms.encrypt_data(key_id, &plaintext).await {
Ok(ct) => json_ok(json!({
"KeyId": key_id,
"CiphertextBlob": B64.encode(&ct),
})),
Err(e) => json_err(StatusCode::INTERNAL_SERVER_ERROR, &e.to_string()),
}
}
pub async fn decrypt(State(state): State<AppState>, body: Body) -> Response {
let kms = match &state.kms {
Some(k) => k,
None => return json_err(StatusCode::SERVICE_UNAVAILABLE, "KMS not enabled"),
};
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(c) => c.to_bytes(),
Err(_) => return json_err(StatusCode::BAD_REQUEST, "Invalid request body"),
};
let req: serde_json::Value = match serde_json::from_slice(&body_bytes) {
Ok(v) => v,
Err(_) => return json_err(StatusCode::BAD_REQUEST, "Invalid JSON"),
};
let key_id = match req.get("KeyId").and_then(|v| v.as_str()) {
Some(k) => k,
None => return json_err(StatusCode::BAD_REQUEST, "Missing KeyId"),
};
let ct_b64 = match req.get("CiphertextBlob").and_then(|v| v.as_str()) {
Some(c) => c,
None => return json_err(StatusCode::BAD_REQUEST, "Missing CiphertextBlob"),
};
let ciphertext = match B64.decode(ct_b64) {
Ok(c) => c,
Err(_) => return json_err(StatusCode::BAD_REQUEST, "Invalid base64"),
};
match kms.decrypt_data(key_id, &ciphertext).await {
Ok(pt) => json_ok(json!({
"KeyId": key_id,
"Plaintext": B64.encode(&pt),
})),
Err(e) => json_err(StatusCode::INTERNAL_SERVER_ERROR, &e.to_string()),
}
}
pub async fn generate_data_key(State(state): State<AppState>, body: Body) -> Response {
let kms = match &state.kms {
Some(k) => k,
None => return json_err(StatusCode::SERVICE_UNAVAILABLE, "KMS not enabled"),
};
let body_bytes = match http_body_util::BodyExt::collect(body).await {
Ok(c) => c.to_bytes(),
Err(_) => return json_err(StatusCode::BAD_REQUEST, "Invalid request body"),
};
let req: serde_json::Value = match serde_json::from_slice(&body_bytes) {
Ok(v) => v,
Err(_) => return json_err(StatusCode::BAD_REQUEST, "Invalid JSON"),
};
let key_id = match req.get("KeyId").and_then(|v| v.as_str()) {
Some(k) => k,
None => return json_err(StatusCode::BAD_REQUEST, "Missing KeyId"),
};
let num_bytes = req
.get("NumberOfBytes")
.and_then(|v| v.as_u64())
.unwrap_or(32) as usize;
if num_bytes < 1 || num_bytes > 1024 {
return json_err(StatusCode::BAD_REQUEST, "NumberOfBytes must be 1-1024");
}
match kms.generate_data_key(key_id, num_bytes).await {
Ok((plaintext, wrapped)) => json_ok(json!({
"KeyId": key_id,
"Plaintext": B64.encode(&plaintext),
"CiphertextBlob": B64.encode(&wrapped),
})),
Err(e) => json_err(StatusCode::INTERNAL_SERVER_ERROR, &e.to_string()),
}
}

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,49 @@
pub mod config;
pub mod handlers;
pub mod middleware;
pub mod services;
pub mod state;
use axum::Router;
pub const SERVER_HEADER: &str = concat!("MyFSIO-Rust/", env!("CARGO_PKG_VERSION"));
pub fn create_router(state: state::AppState) -> Router {
let mut router = Router::new()
.route("/", axum::routing::get(handlers::list_buckets))
.route(
"/{bucket}",
axum::routing::put(handlers::create_bucket)
.get(handlers::get_bucket)
.delete(handlers::delete_bucket)
.head(handlers::head_bucket)
.post(handlers::post_bucket),
)
.route(
"/{bucket}/{*key}",
axum::routing::put(handlers::put_object)
.get(handlers::get_object)
.delete(handlers::delete_object)
.head(handlers::head_object)
.post(handlers::post_object),
);
if state.config.kms_enabled {
router = router
.route("/kms/keys", axum::routing::get(handlers::kms::list_keys).post(handlers::kms::create_key))
.route("/kms/keys/{key_id}", axum::routing::get(handlers::kms::get_key).delete(handlers::kms::delete_key))
.route("/kms/keys/{key_id}/enable", axum::routing::post(handlers::kms::enable_key))
.route("/kms/keys/{key_id}/disable", axum::routing::post(handlers::kms::disable_key))
.route("/kms/encrypt", axum::routing::post(handlers::kms::encrypt))
.route("/kms/decrypt", axum::routing::post(handlers::kms::decrypt))
.route("/kms/generate-data-key", axum::routing::post(handlers::kms::generate_data_key));
}
router
.layer(axum::middleware::from_fn_with_state(
state.clone(),
middleware::auth_layer,
))
.layer(axum::middleware::from_fn(middleware::server_header))
.with_state(state)
}

View File

@@ -0,0 +1,78 @@
use myfsio_server::config::ServerConfig;
use myfsio_server::state::AppState;
#[tokio::main]
async fn main() {
tracing_subscriber::fmt::init();
let config = ServerConfig::from_env();
let bind_addr = config.bind_addr;
tracing::info!("MyFSIO Rust Engine starting on {}", bind_addr);
tracing::info!("Storage root: {}", config.storage_root.display());
tracing::info!("Region: {}", config.region);
tracing::info!(
"Encryption: {}, KMS: {}, GC: {}, Lifecycle: {}, Integrity: {}, Metrics: {}",
config.encryption_enabled,
config.kms_enabled,
config.gc_enabled,
config.lifecycle_enabled,
config.integrity_enabled,
config.metrics_enabled
);
let state = if config.encryption_enabled || config.kms_enabled {
AppState::new_with_encryption(config.clone()).await
} else {
AppState::new(config.clone())
};
let mut bg_handles: Vec<tokio::task::JoinHandle<()>> = Vec::new();
if let Some(ref gc) = state.gc {
bg_handles.push(gc.clone().start_background());
tracing::info!("GC background service started");
}
if let Some(ref integrity) = state.integrity {
bg_handles.push(integrity.clone().start_background());
tracing::info!("Integrity checker background service started");
}
if let Some(ref metrics) = state.metrics {
bg_handles.push(metrics.clone().start_background());
tracing::info!("Metrics collector background service started");
}
if config.lifecycle_enabled {
let lifecycle = std::sync::Arc::new(
myfsio_server::services::lifecycle::LifecycleService::new(
state.storage.clone(),
myfsio_server::services::lifecycle::LifecycleConfig::default(),
),
);
bg_handles.push(lifecycle.start_background());
tracing::info!("Lifecycle manager background service started");
}
let app = myfsio_server::create_router(state);
let listener = tokio::net::TcpListener::bind(bind_addr).await.unwrap();
tracing::info!("Listening on {}", bind_addr);
axum::serve(listener, app)
.with_graceful_shutdown(shutdown_signal())
.await
.unwrap();
for handle in bg_handles {
handle.abort();
}
}
async fn shutdown_signal() {
tokio::signal::ctrl_c()
.await
.expect("Failed to listen for Ctrl+C");
tracing::info!("Shutdown signal received");
}

View File

@@ -0,0 +1,390 @@
use axum::extract::{Request, State};
use axum::http::StatusCode;
use axum::middleware::Next;
use axum::response::{IntoResponse, Response};
use chrono::{NaiveDateTime, Utc};
use myfsio_auth::sigv4;
use myfsio_common::error::{S3Error, S3ErrorCode};
use myfsio_common::types::Principal;
use crate::state::AppState;
pub async fn auth_layer(
State(state): State<AppState>,
mut req: Request,
next: Next,
) -> Response {
let uri = req.uri().clone();
let path = uri.path();
if path == "/" && req.method() == axum::http::Method::GET {
match try_auth(&state, &req) {
AuthResult::Ok(principal) => {
req.extensions_mut().insert(principal);
}
AuthResult::Denied(err) => return error_response(err),
AuthResult::NoAuth => {
return error_response(
S3Error::from_code(S3ErrorCode::AccessDenied),
);
}
}
return next.run(req).await;
}
match try_auth(&state, &req) {
AuthResult::Ok(principal) => {
req.extensions_mut().insert(principal);
next.run(req).await
}
AuthResult::Denied(err) => error_response(err),
AuthResult::NoAuth => {
error_response(S3Error::from_code(S3ErrorCode::AccessDenied))
}
}
}
enum AuthResult {
Ok(Principal),
Denied(S3Error),
NoAuth,
}
fn try_auth(state: &AppState, req: &Request) -> AuthResult {
if let Some(auth_header) = req.headers().get("authorization") {
if let Ok(auth_str) = auth_header.to_str() {
if auth_str.starts_with("AWS4-HMAC-SHA256 ") {
return verify_sigv4_header(state, req, auth_str);
}
}
}
let query = req.uri().query().unwrap_or("");
if query.contains("X-Amz-Algorithm=AWS4-HMAC-SHA256") {
return verify_sigv4_query(state, req);
}
if let (Some(ak), Some(sk)) = (
req.headers().get("x-access-key").and_then(|v| v.to_str().ok()),
req.headers().get("x-secret-key").and_then(|v| v.to_str().ok()),
) {
return match state.iam.authenticate(ak, sk) {
Some(principal) => AuthResult::Ok(principal),
None => AuthResult::Denied(
S3Error::from_code(S3ErrorCode::SignatureDoesNotMatch),
),
};
}
AuthResult::NoAuth
}
fn verify_sigv4_header(state: &AppState, req: &Request, auth_str: &str) -> AuthResult {
let parts: Vec<&str> = auth_str
.strip_prefix("AWS4-HMAC-SHA256 ")
.unwrap()
.split(", ")
.collect();
if parts.len() != 3 {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::InvalidArgument, "Malformed Authorization header"),
);
}
let credential = parts[0].strip_prefix("Credential=").unwrap_or("");
let signed_headers_str = parts[1].strip_prefix("SignedHeaders=").unwrap_or("");
let provided_signature = parts[2].strip_prefix("Signature=").unwrap_or("");
let cred_parts: Vec<&str> = credential.split('/').collect();
if cred_parts.len() != 5 {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::InvalidArgument, "Malformed credential"),
);
}
let access_key = cred_parts[0];
let date_stamp = cred_parts[1];
let region = cred_parts[2];
let service = cred_parts[3];
let amz_date = req
.headers()
.get("x-amz-date")
.or_else(|| req.headers().get("date"))
.and_then(|v| v.to_str().ok())
.unwrap_or("");
if amz_date.is_empty() {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::AccessDenied, "Missing Date header"),
);
}
if let Some(err) = check_timestamp_freshness(amz_date, state.config.sigv4_timestamp_tolerance_secs) {
return AuthResult::Denied(err);
}
let secret_key = match state.iam.get_secret_key(access_key) {
Some(sk) => sk,
None => {
return AuthResult::Denied(
S3Error::from_code(S3ErrorCode::InvalidAccessKeyId),
);
}
};
let method = req.method().as_str();
let canonical_uri = req.uri().path();
let query_params = parse_query_params(req.uri().query().unwrap_or(""));
let payload_hash = req
.headers()
.get("x-amz-content-sha256")
.and_then(|v| v.to_str().ok())
.unwrap_or("UNSIGNED-PAYLOAD");
let signed_headers: Vec<&str> = signed_headers_str.split(';').collect();
let header_values: Vec<(String, String)> = signed_headers
.iter()
.map(|&name| {
let value = req
.headers()
.get(name)
.and_then(|v| v.to_str().ok())
.unwrap_or("");
(name.to_string(), value.to_string())
})
.collect();
let verified = sigv4::verify_sigv4_signature(
method,
canonical_uri,
&query_params,
signed_headers_str,
&header_values,
payload_hash,
amz_date,
date_stamp,
region,
service,
&secret_key,
provided_signature,
);
if !verified {
return AuthResult::Denied(
S3Error::from_code(S3ErrorCode::SignatureDoesNotMatch),
);
}
match state.iam.get_principal(access_key) {
Some(p) => AuthResult::Ok(p),
None => AuthResult::Denied(
S3Error::from_code(S3ErrorCode::InvalidAccessKeyId),
),
}
}
fn verify_sigv4_query(state: &AppState, req: &Request) -> AuthResult {
let query = req.uri().query().unwrap_or("");
let params = parse_query_params(query);
let param_map: std::collections::HashMap<&str, &str> = params
.iter()
.map(|(k, v)| (k.as_str(), v.as_str()))
.collect();
let credential = match param_map.get("X-Amz-Credential") {
Some(c) => *c,
None => {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::InvalidArgument, "Missing X-Amz-Credential"),
);
}
};
let signed_headers_str = param_map
.get("X-Amz-SignedHeaders")
.copied()
.unwrap_or("host");
let provided_signature = match param_map.get("X-Amz-Signature") {
Some(s) => *s,
None => {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::InvalidArgument, "Missing X-Amz-Signature"),
);
}
};
let amz_date = match param_map.get("X-Amz-Date") {
Some(d) => *d,
None => {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::InvalidArgument, "Missing X-Amz-Date"),
);
}
};
let expires_str = match param_map.get("X-Amz-Expires") {
Some(e) => *e,
None => {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::InvalidArgument, "Missing X-Amz-Expires"),
);
}
};
let cred_parts: Vec<&str> = credential.split('/').collect();
if cred_parts.len() != 5 {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::InvalidArgument, "Malformed credential"),
);
}
let access_key = cred_parts[0];
let date_stamp = cred_parts[1];
let region = cred_parts[2];
let service = cred_parts[3];
let expires: u64 = match expires_str.parse() {
Ok(e) => e,
Err(_) => {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::InvalidArgument, "Invalid X-Amz-Expires"),
);
}
};
if expires < state.config.presigned_url_min_expiry
|| expires > state.config.presigned_url_max_expiry
{
return AuthResult::Denied(
S3Error::new(S3ErrorCode::InvalidArgument, "X-Amz-Expires out of range"),
);
}
if let Ok(request_time) =
NaiveDateTime::parse_from_str(amz_date, "%Y%m%dT%H%M%SZ")
{
let request_utc = request_time.and_utc();
let now = Utc::now();
let elapsed = (now - request_utc).num_seconds();
if elapsed > expires as i64 {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::AccessDenied, "Request has expired"),
);
}
if elapsed < -(state.config.sigv4_timestamp_tolerance_secs as i64) {
return AuthResult::Denied(
S3Error::new(S3ErrorCode::AccessDenied, "Request is too far in the future"),
);
}
}
let secret_key = match state.iam.get_secret_key(access_key) {
Some(sk) => sk,
None => {
return AuthResult::Denied(
S3Error::from_code(S3ErrorCode::InvalidAccessKeyId),
);
}
};
let method = req.method().as_str();
let canonical_uri = req.uri().path();
let query_params_no_sig: Vec<(String, String)> = params
.iter()
.filter(|(k, _)| k != "X-Amz-Signature")
.cloned()
.collect();
let payload_hash = "UNSIGNED-PAYLOAD";
let signed_headers: Vec<&str> = signed_headers_str.split(';').collect();
let header_values: Vec<(String, String)> = signed_headers
.iter()
.map(|&name| {
let value = req
.headers()
.get(name)
.and_then(|v| v.to_str().ok())
.unwrap_or("");
(name.to_string(), value.to_string())
})
.collect();
let verified = sigv4::verify_sigv4_signature(
method,
canonical_uri,
&query_params_no_sig,
signed_headers_str,
&header_values,
payload_hash,
amz_date,
date_stamp,
region,
service,
&secret_key,
provided_signature,
);
if !verified {
return AuthResult::Denied(
S3Error::from_code(S3ErrorCode::SignatureDoesNotMatch),
);
}
match state.iam.get_principal(access_key) {
Some(p) => AuthResult::Ok(p),
None => AuthResult::Denied(
S3Error::from_code(S3ErrorCode::InvalidAccessKeyId),
),
}
}
fn check_timestamp_freshness(amz_date: &str, tolerance_secs: u64) -> Option<S3Error> {
let request_time = NaiveDateTime::parse_from_str(amz_date, "%Y%m%dT%H%M%SZ").ok()?;
let request_utc = request_time.and_utc();
let now = Utc::now();
let diff = (now - request_utc).num_seconds().unsigned_abs();
if diff > tolerance_secs {
return Some(S3Error::new(
S3ErrorCode::AccessDenied,
"Request timestamp too old or too far in the future",
));
}
None
}
fn parse_query_params(query: &str) -> Vec<(String, String)> {
if query.is_empty() {
return Vec::new();
}
query
.split('&')
.filter_map(|pair| {
let mut parts = pair.splitn(2, '=');
let key = parts.next()?;
let value = parts.next().unwrap_or("");
Some((
urlencoding_decode(key),
urlencoding_decode(value),
))
})
.collect()
}
fn urlencoding_decode(s: &str) -> String {
percent_encoding::percent_decode_str(s)
.decode_utf8_lossy()
.into_owned()
}
fn error_response(err: S3Error) -> Response {
let status =
StatusCode::from_u16(err.http_status()).unwrap_or(StatusCode::INTERNAL_SERVER_ERROR);
let body = err.to_xml();
(status, [("content-type", "application/xml")], body).into_response()
}

View File

@@ -0,0 +1,16 @@
mod auth;
pub use auth::auth_layer;
use axum::extract::Request;
use axum::middleware::Next;
use axum::response::Response;
pub async fn server_header(req: Request, next: Next) -> Response {
let mut resp = next.run(req).await;
resp.headers_mut().insert(
"server",
crate::SERVER_HEADER.parse().unwrap(),
);
resp
}

View File

@@ -0,0 +1,263 @@
use serde_json::{json, Value};
use std::path::PathBuf;
use std::sync::Arc;
use std::time::Instant;
use tokio::sync::RwLock;
pub struct GcConfig {
pub interval_hours: f64,
pub temp_file_max_age_hours: f64,
pub multipart_max_age_days: u64,
pub lock_file_max_age_hours: f64,
pub dry_run: bool,
}
impl Default for GcConfig {
fn default() -> Self {
Self {
interval_hours: 6.0,
temp_file_max_age_hours: 24.0,
multipart_max_age_days: 7,
lock_file_max_age_hours: 1.0,
dry_run: false,
}
}
}
pub struct GcService {
storage_root: PathBuf,
config: GcConfig,
running: Arc<RwLock<bool>>,
history: Arc<RwLock<Vec<Value>>>,
history_path: PathBuf,
}
impl GcService {
pub fn new(storage_root: PathBuf, config: GcConfig) -> Self {
let history_path = storage_root
.join(".myfsio.sys")
.join("config")
.join("gc_history.json");
let history = if history_path.exists() {
std::fs::read_to_string(&history_path)
.ok()
.and_then(|s| serde_json::from_str::<Value>(&s).ok())
.and_then(|v| v.get("executions").and_then(|e| e.as_array().cloned()))
.unwrap_or_default()
} else {
Vec::new()
};
Self {
storage_root,
config,
running: Arc::new(RwLock::new(false)),
history: Arc::new(RwLock::new(history)),
history_path,
}
}
pub async fn status(&self) -> Value {
let running = *self.running.read().await;
json!({
"enabled": true,
"running": running,
"interval_hours": self.config.interval_hours,
"temp_file_max_age_hours": self.config.temp_file_max_age_hours,
"multipart_max_age_days": self.config.multipart_max_age_days,
"lock_file_max_age_hours": self.config.lock_file_max_age_hours,
"dry_run": self.config.dry_run,
})
}
pub async fn history(&self) -> Value {
let history = self.history.read().await;
json!({ "executions": *history })
}
pub async fn run_now(&self, dry_run: bool) -> Result<Value, String> {
{
let mut running = self.running.write().await;
if *running {
return Err("GC already running".to_string());
}
*running = true;
}
let start = Instant::now();
let result = self.execute_gc(dry_run || self.config.dry_run).await;
let elapsed = start.elapsed().as_secs_f64();
*self.running.write().await = false;
let mut result_json = result.clone();
if let Some(obj) = result_json.as_object_mut() {
obj.insert("execution_time_seconds".to_string(), json!(elapsed));
}
let record = json!({
"timestamp": chrono::Utc::now().timestamp_millis() as f64 / 1000.0,
"dry_run": dry_run || self.config.dry_run,
"result": result_json,
});
{
let mut history = self.history.write().await;
history.push(record);
if history.len() > 50 {
let excess = history.len() - 50;
history.drain(..excess);
}
}
self.save_history().await;
Ok(result)
}
async fn execute_gc(&self, dry_run: bool) -> Value {
let mut temp_files_deleted = 0u64;
let mut temp_bytes_freed = 0u64;
let mut multipart_uploads_deleted = 0u64;
let mut lock_files_deleted = 0u64;
let mut empty_dirs_removed = 0u64;
let mut errors: Vec<String> = Vec::new();
let now = std::time::SystemTime::now();
let temp_max_age = std::time::Duration::from_secs_f64(self.config.temp_file_max_age_hours * 3600.0);
let multipart_max_age = std::time::Duration::from_secs(self.config.multipart_max_age_days * 86400);
let lock_max_age = std::time::Duration::from_secs_f64(self.config.lock_file_max_age_hours * 3600.0);
let tmp_dir = self.storage_root.join(".myfsio.sys").join("tmp");
if tmp_dir.exists() {
match std::fs::read_dir(&tmp_dir) {
Ok(entries) => {
for entry in entries.flatten() {
if let Ok(metadata) = entry.metadata() {
if let Ok(modified) = metadata.modified() {
if let Ok(age) = now.duration_since(modified) {
if age > temp_max_age {
let size = metadata.len();
if !dry_run {
if let Err(e) = std::fs::remove_file(entry.path()) {
errors.push(format!("Failed to remove temp file: {}", e));
continue;
}
}
temp_files_deleted += 1;
temp_bytes_freed += size;
}
}
}
}
}
}
Err(e) => errors.push(format!("Failed to read tmp dir: {}", e)),
}
}
let multipart_dir = self.storage_root.join(".myfsio.sys").join("multipart");
if multipart_dir.exists() {
if let Ok(bucket_dirs) = std::fs::read_dir(&multipart_dir) {
for bucket_entry in bucket_dirs.flatten() {
if let Ok(uploads) = std::fs::read_dir(bucket_entry.path()) {
for upload in uploads.flatten() {
if let Ok(metadata) = upload.metadata() {
if let Ok(modified) = metadata.modified() {
if let Ok(age) = now.duration_since(modified) {
if age > multipart_max_age {
if !dry_run {
let _ = std::fs::remove_dir_all(upload.path());
}
multipart_uploads_deleted += 1;
}
}
}
}
}
}
}
}
}
let buckets_dir = self.storage_root.join(".myfsio.sys").join("buckets");
if buckets_dir.exists() {
if let Ok(bucket_dirs) = std::fs::read_dir(&buckets_dir) {
for bucket_entry in bucket_dirs.flatten() {
let locks_dir = bucket_entry.path().join("locks");
if locks_dir.exists() {
if let Ok(locks) = std::fs::read_dir(&locks_dir) {
for lock in locks.flatten() {
if let Ok(metadata) = lock.metadata() {
if let Ok(modified) = metadata.modified() {
if let Ok(age) = now.duration_since(modified) {
if age > lock_max_age {
if !dry_run {
let _ = std::fs::remove_file(lock.path());
}
lock_files_deleted += 1;
}
}
}
}
}
}
}
}
}
}
if !dry_run {
for dir in [&tmp_dir, &multipart_dir] {
if dir.exists() {
if let Ok(entries) = std::fs::read_dir(dir) {
for entry in entries.flatten() {
if entry.path().is_dir() {
if let Ok(mut contents) = std::fs::read_dir(entry.path()) {
if contents.next().is_none() {
let _ = std::fs::remove_dir(entry.path());
empty_dirs_removed += 1;
}
}
}
}
}
}
}
}
json!({
"temp_files_deleted": temp_files_deleted,
"temp_bytes_freed": temp_bytes_freed,
"multipart_uploads_deleted": multipart_uploads_deleted,
"lock_files_deleted": lock_files_deleted,
"empty_dirs_removed": empty_dirs_removed,
"errors": errors,
})
}
async fn save_history(&self) {
let history = self.history.read().await;
let data = json!({ "executions": *history });
if let Some(parent) = self.history_path.parent() {
let _ = std::fs::create_dir_all(parent);
}
let _ = std::fs::write(&self.history_path, serde_json::to_string_pretty(&data).unwrap_or_default());
}
pub fn start_background(self: Arc<Self>) -> tokio::task::JoinHandle<()> {
let interval = std::time::Duration::from_secs_f64(self.config.interval_hours * 3600.0);
tokio::spawn(async move {
let mut timer = tokio::time::interval(interval);
timer.tick().await;
loop {
timer.tick().await;
tracing::info!("GC cycle starting");
match self.run_now(false).await {
Ok(result) => tracing::info!("GC cycle complete: {:?}", result),
Err(e) => tracing::warn!("GC cycle failed: {}", e),
}
}
})
}
}

View File

@@ -0,0 +1,204 @@
use myfsio_storage::fs_backend::FsStorageBackend;
use myfsio_storage::traits::StorageEngine;
use serde_json::{json, Value};
use std::path::PathBuf;
use std::sync::Arc;
use std::time::Instant;
use tokio::sync::RwLock;
pub struct IntegrityConfig {
pub interval_hours: f64,
pub batch_size: usize,
pub auto_heal: bool,
pub dry_run: bool,
}
impl Default for IntegrityConfig {
fn default() -> Self {
Self {
interval_hours: 24.0,
batch_size: 1000,
auto_heal: false,
dry_run: false,
}
}
}
pub struct IntegrityService {
storage: Arc<FsStorageBackend>,
config: IntegrityConfig,
running: Arc<RwLock<bool>>,
history: Arc<RwLock<Vec<Value>>>,
history_path: PathBuf,
}
impl IntegrityService {
pub fn new(
storage: Arc<FsStorageBackend>,
storage_root: &std::path::Path,
config: IntegrityConfig,
) -> Self {
let history_path = storage_root
.join(".myfsio.sys")
.join("config")
.join("integrity_history.json");
let history = if history_path.exists() {
std::fs::read_to_string(&history_path)
.ok()
.and_then(|s| serde_json::from_str::<Value>(&s).ok())
.and_then(|v| v.get("executions").and_then(|e| e.as_array().cloned()))
.unwrap_or_default()
} else {
Vec::new()
};
Self {
storage,
config,
running: Arc::new(RwLock::new(false)),
history: Arc::new(RwLock::new(history)),
history_path,
}
}
pub async fn status(&self) -> Value {
let running = *self.running.read().await;
json!({
"enabled": true,
"running": running,
"interval_hours": self.config.interval_hours,
"batch_size": self.config.batch_size,
"auto_heal": self.config.auto_heal,
"dry_run": self.config.dry_run,
})
}
pub async fn history(&self) -> Value {
let history = self.history.read().await;
json!({ "executions": *history })
}
pub async fn run_now(&self, dry_run: bool, auto_heal: bool) -> Result<Value, String> {
{
let mut running = self.running.write().await;
if *running {
return Err("Integrity check already running".to_string());
}
*running = true;
}
let start = Instant::now();
let result = self.check_integrity(dry_run, auto_heal).await;
let elapsed = start.elapsed().as_secs_f64();
*self.running.write().await = false;
let mut result_json = result.clone();
if let Some(obj) = result_json.as_object_mut() {
obj.insert("execution_time_seconds".to_string(), json!(elapsed));
}
let record = json!({
"timestamp": chrono::Utc::now().timestamp_millis() as f64 / 1000.0,
"dry_run": dry_run,
"auto_heal": auto_heal,
"result": result_json,
});
{
let mut history = self.history.write().await;
history.push(record);
if history.len() > 50 {
let excess = history.len() - 50;
history.drain(..excess);
}
}
self.save_history().await;
Ok(result)
}
async fn check_integrity(&self, _dry_run: bool, _auto_heal: bool) -> Value {
let buckets = match self.storage.list_buckets().await {
Ok(b) => b,
Err(e) => return json!({"error": e.to_string()}),
};
let mut objects_scanned = 0u64;
let mut corrupted = 0u64;
let mut phantom_metadata = 0u64;
let mut errors: Vec<String> = Vec::new();
for bucket in &buckets {
let params = myfsio_common::types::ListParams {
max_keys: self.config.batch_size,
..Default::default()
};
let objects = match self.storage.list_objects(&bucket.name, &params).await {
Ok(r) => r.objects,
Err(e) => {
errors.push(format!("{}: {}", bucket.name, e));
continue;
}
};
for obj in &objects {
objects_scanned += 1;
match self.storage.get_object_path(&bucket.name, &obj.key).await {
Ok(path) => {
if !path.exists() {
phantom_metadata += 1;
} else if let Some(ref expected_etag) = obj.etag {
match myfsio_crypto::hashing::md5_file(&path) {
Ok(actual_etag) => {
if &actual_etag != expected_etag {
corrupted += 1;
}
}
Err(e) => errors.push(format!("{}:{}: {}", bucket.name, obj.key, e)),
}
}
}
Err(e) => errors.push(format!("{}:{}: {}", bucket.name, obj.key, e)),
}
}
}
json!({
"objects_scanned": objects_scanned,
"buckets_scanned": buckets.len(),
"corrupted_objects": corrupted,
"phantom_metadata": phantom_metadata,
"errors": errors,
})
}
async fn save_history(&self) {
let history = self.history.read().await;
let data = json!({ "executions": *history });
if let Some(parent) = self.history_path.parent() {
let _ = std::fs::create_dir_all(parent);
}
let _ = std::fs::write(
&self.history_path,
serde_json::to_string_pretty(&data).unwrap_or_default(),
);
}
pub fn start_background(self: Arc<Self>) -> tokio::task::JoinHandle<()> {
let interval = std::time::Duration::from_secs_f64(self.config.interval_hours * 3600.0);
tokio::spawn(async move {
let mut timer = tokio::time::interval(interval);
timer.tick().await;
loop {
timer.tick().await;
tracing::info!("Integrity check starting");
match self.run_now(false, false).await {
Ok(result) => tracing::info!("Integrity check complete: {:?}", result),
Err(e) => tracing::warn!("Integrity check failed: {}", e),
}
}
})
}
}

View File

@@ -0,0 +1,153 @@
use myfsio_storage::fs_backend::FsStorageBackend;
use myfsio_storage::traits::StorageEngine;
use serde_json::{json, Value};
use std::sync::Arc;
use tokio::sync::RwLock;
pub struct LifecycleConfig {
pub interval_seconds: u64,
}
impl Default for LifecycleConfig {
fn default() -> Self {
Self {
interval_seconds: 3600,
}
}
}
pub struct LifecycleService {
storage: Arc<FsStorageBackend>,
config: LifecycleConfig,
running: Arc<RwLock<bool>>,
}
impl LifecycleService {
pub fn new(storage: Arc<FsStorageBackend>, config: LifecycleConfig) -> Self {
Self {
storage,
config,
running: Arc::new(RwLock::new(false)),
}
}
pub async fn run_cycle(&self) -> Result<Value, String> {
{
let mut running = self.running.write().await;
if *running {
return Err("Lifecycle already running".to_string());
}
*running = true;
}
let result = self.evaluate_rules().await;
*self.running.write().await = false;
Ok(result)
}
async fn evaluate_rules(&self) -> Value {
let buckets = match self.storage.list_buckets().await {
Ok(b) => b,
Err(e) => return json!({"error": e.to_string()}),
};
let mut total_expired = 0u64;
let mut total_multipart_aborted = 0u64;
let mut errors: Vec<String> = Vec::new();
for bucket in &buckets {
let config = match self.storage.get_bucket_config(&bucket.name).await {
Ok(c) => c,
Err(_) => continue,
};
let lifecycle = match &config.lifecycle {
Some(lc) => lc,
None => continue,
};
let rules = match lifecycle.as_str().and_then(|s| serde_json::from_str::<Value>(s).ok()) {
Some(v) => v,
None => continue,
};
let rules_arr = match rules.get("Rules").and_then(|r| r.as_array()) {
Some(a) => a.clone(),
None => continue,
};
for rule in &rules_arr {
if rule.get("Status").and_then(|s| s.as_str()) != Some("Enabled") {
continue;
}
let prefix = rule
.get("Filter")
.and_then(|f| f.get("Prefix"))
.and_then(|p| p.as_str())
.or_else(|| rule.get("Prefix").and_then(|p| p.as_str()))
.unwrap_or("");
if let Some(exp) = rule.get("Expiration") {
if let Some(days) = exp.get("Days").and_then(|d| d.as_u64()) {
let cutoff = chrono::Utc::now() - chrono::Duration::days(days as i64);
let params = myfsio_common::types::ListParams {
max_keys: 1000,
prefix: if prefix.is_empty() { None } else { Some(prefix.to_string()) },
..Default::default()
};
if let Ok(result) = self.storage.list_objects(&bucket.name, &params).await {
for obj in &result.objects {
if obj.last_modified < cutoff {
match self.storage.delete_object(&bucket.name, &obj.key).await {
Ok(()) => total_expired += 1,
Err(e) => errors.push(format!("{}:{}: {}", bucket.name, obj.key, e)),
}
}
}
}
}
}
if let Some(abort) = rule.get("AbortIncompleteMultipartUpload") {
if let Some(days) = abort.get("DaysAfterInitiation").and_then(|d| d.as_u64()) {
let cutoff = chrono::Utc::now() - chrono::Duration::days(days as i64);
if let Ok(uploads) = self.storage.list_multipart_uploads(&bucket.name).await {
for upload in &uploads {
if upload.initiated < cutoff {
match self.storage.abort_multipart(&bucket.name, &upload.upload_id).await {
Ok(()) => total_multipart_aborted += 1,
Err(e) => errors.push(format!("abort {}: {}", upload.upload_id, e)),
}
}
}
}
}
}
}
}
json!({
"objects_expired": total_expired,
"multipart_aborted": total_multipart_aborted,
"buckets_evaluated": buckets.len(),
"errors": errors,
})
}
pub fn start_background(self: Arc<Self>) -> tokio::task::JoinHandle<()> {
let interval = std::time::Duration::from_secs(self.config.interval_seconds);
tokio::spawn(async move {
let mut timer = tokio::time::interval(interval);
timer.tick().await;
loop {
timer.tick().await;
tracing::info!("Lifecycle evaluation starting");
match self.run_cycle().await {
Ok(result) => tracing::info!("Lifecycle cycle complete: {:?}", result),
Err(e) => tracing::warn!("Lifecycle cycle failed: {}", e),
}
}
})
}
}

View File

@@ -0,0 +1,219 @@
use serde_json::{json, Value};
use std::collections::HashMap;
use std::path::PathBuf;
use std::sync::Arc;
use std::time::Instant;
use tokio::sync::RwLock;
pub struct MetricsConfig {
pub interval_minutes: u64,
pub retention_hours: u64,
}
impl Default for MetricsConfig {
fn default() -> Self {
Self {
interval_minutes: 5,
retention_hours: 24,
}
}
}
struct MethodStats {
count: u64,
success_count: u64,
error_count: u64,
bytes_in: u64,
bytes_out: u64,
latencies: Vec<f64>,
}
impl MethodStats {
fn new() -> Self {
Self {
count: 0,
success_count: 0,
error_count: 0,
bytes_in: 0,
bytes_out: 0,
latencies: Vec::new(),
}
}
fn to_json(&self) -> Value {
let (min, max, avg, p50, p95, p99) = if self.latencies.is_empty() {
(0.0, 0.0, 0.0, 0.0, 0.0, 0.0)
} else {
let mut sorted = self.latencies.clone();
sorted.sort_by(|a, b| a.partial_cmp(b).unwrap_or(std::cmp::Ordering::Equal));
let len = sorted.len();
let sum: f64 = sorted.iter().sum();
(
sorted[0],
sorted[len - 1],
sum / len as f64,
sorted[len / 2],
sorted[((len as f64 * 0.95) as usize).min(len - 1)],
sorted[((len as f64 * 0.99) as usize).min(len - 1)],
)
};
json!({
"count": self.count,
"success_count": self.success_count,
"error_count": self.error_count,
"bytes_in": self.bytes_in,
"bytes_out": self.bytes_out,
"latency_min_ms": min,
"latency_max_ms": max,
"latency_avg_ms": avg,
"latency_p50_ms": p50,
"latency_p95_ms": p95,
"latency_p99_ms": p99,
})
}
}
struct CurrentWindow {
by_method: HashMap<String, MethodStats>,
by_status_class: HashMap<String, u64>,
start_time: Instant,
}
impl CurrentWindow {
fn new() -> Self {
Self {
by_method: HashMap::new(),
by_status_class: HashMap::new(),
start_time: Instant::now(),
}
}
fn reset(&mut self) {
self.by_method.clear();
self.by_status_class.clear();
self.start_time = Instant::now();
}
}
pub struct MetricsService {
config: MetricsConfig,
current: Arc<RwLock<CurrentWindow>>,
snapshots: Arc<RwLock<Vec<Value>>>,
snapshots_path: PathBuf,
}
impl MetricsService {
pub fn new(storage_root: &std::path::Path, config: MetricsConfig) -> Self {
let snapshots_path = storage_root
.join(".myfsio.sys")
.join("config")
.join("operation_metrics.json");
let snapshots = if snapshots_path.exists() {
std::fs::read_to_string(&snapshots_path)
.ok()
.and_then(|s| serde_json::from_str::<Value>(&s).ok())
.and_then(|v| v.get("snapshots").and_then(|s| s.as_array().cloned()))
.unwrap_or_default()
} else {
Vec::new()
};
Self {
config,
current: Arc::new(RwLock::new(CurrentWindow::new())),
snapshots: Arc::new(RwLock::new(snapshots)),
snapshots_path,
}
}
pub async fn record(&self, method: &str, status: u16, latency_ms: f64, bytes_in: u64, bytes_out: u64) {
let mut window = self.current.write().await;
let stats = window.by_method.entry(method.to_string()).or_insert_with(MethodStats::new);
stats.count += 1;
if status < 400 {
stats.success_count += 1;
} else {
stats.error_count += 1;
}
stats.bytes_in += bytes_in;
stats.bytes_out += bytes_out;
stats.latencies.push(latency_ms);
let class = format!("{}xx", status / 100);
*window.by_status_class.entry(class).or_insert(0) += 1;
}
pub async fn snapshot(&self) -> Value {
let window = self.current.read().await;
let mut by_method = serde_json::Map::new();
for (method, stats) in &window.by_method {
by_method.insert(method.clone(), stats.to_json());
}
let snapshots = self.snapshots.read().await;
json!({
"enabled": true,
"current_window": {
"by_method": by_method,
"by_status_class": window.by_status_class,
"window_start_elapsed_secs": window.start_time.elapsed().as_secs_f64(),
},
"snapshots": *snapshots,
})
}
async fn flush_window(&self) {
let snap = {
let mut window = self.current.write().await;
let mut by_method = serde_json::Map::new();
for (method, stats) in &window.by_method {
by_method.insert(method.clone(), stats.to_json());
}
let snap = json!({
"timestamp": chrono::Utc::now().to_rfc3339(),
"window_seconds": self.config.interval_minutes * 60,
"by_method": by_method,
"by_status_class": window.by_status_class,
});
window.reset();
snap
};
let max_snapshots = (self.config.retention_hours * 60 / self.config.interval_minutes) as usize;
{
let mut snapshots = self.snapshots.write().await;
snapshots.push(snap);
if snapshots.len() > max_snapshots {
let excess = snapshots.len() - max_snapshots;
snapshots.drain(..excess);
}
}
self.save_snapshots().await;
}
async fn save_snapshots(&self) {
let snapshots = self.snapshots.read().await;
let data = json!({ "snapshots": *snapshots });
if let Some(parent) = self.snapshots_path.parent() {
let _ = std::fs::create_dir_all(parent);
}
let _ = std::fs::write(
&self.snapshots_path,
serde_json::to_string_pretty(&data).unwrap_or_default(),
);
}
pub fn start_background(self: Arc<Self>) -> tokio::task::JoinHandle<()> {
let interval = std::time::Duration::from_secs(self.config.interval_minutes * 60);
tokio::spawn(async move {
let mut timer = tokio::time::interval(interval);
timer.tick().await;
loop {
timer.tick().await;
self.flush_window().await;
}
})
}
}

View File

@@ -0,0 +1,4 @@
pub mod gc;
pub mod lifecycle;
pub mod integrity;
pub mod metrics;

View File

@@ -0,0 +1,107 @@
use std::sync::Arc;
use crate::config::ServerConfig;
use crate::services::gc::GcService;
use crate::services::integrity::IntegrityService;
use crate::services::metrics::MetricsService;
use myfsio_auth::iam::IamService;
use myfsio_crypto::encryption::EncryptionService;
use myfsio_crypto::kms::KmsService;
use myfsio_storage::fs_backend::FsStorageBackend;
#[derive(Clone)]
pub struct AppState {
pub config: ServerConfig,
pub storage: Arc<FsStorageBackend>,
pub iam: Arc<IamService>,
pub encryption: Option<Arc<EncryptionService>>,
pub kms: Option<Arc<KmsService>>,
pub gc: Option<Arc<GcService>>,
pub integrity: Option<Arc<IntegrityService>>,
pub metrics: Option<Arc<MetricsService>>,
}
impl AppState {
pub fn new(config: ServerConfig) -> Self {
let storage = Arc::new(FsStorageBackend::new(config.storage_root.clone()));
let iam = Arc::new(IamService::new_with_secret(
config.iam_config_path.clone(),
config.secret_key.clone(),
));
let gc = if config.gc_enabled {
Some(Arc::new(GcService::new(
config.storage_root.clone(),
crate::services::gc::GcConfig::default(),
)))
} else {
None
};
let integrity = if config.integrity_enabled {
Some(Arc::new(IntegrityService::new(
storage.clone(),
&config.storage_root,
crate::services::integrity::IntegrityConfig::default(),
)))
} else {
None
};
let metrics = if config.metrics_enabled {
Some(Arc::new(MetricsService::new(
&config.storage_root,
crate::services::metrics::MetricsConfig::default(),
)))
} else {
None
};
Self {
config,
storage,
iam,
encryption: None,
kms: None,
gc,
integrity,
metrics,
}
}
pub async fn new_with_encryption(config: ServerConfig) -> Self {
let mut state = Self::new(config.clone());
let keys_dir = config.storage_root.join(".myfsio.sys").join("keys");
let kms = if config.kms_enabled {
match KmsService::new(&keys_dir).await {
Ok(k) => Some(Arc::new(k)),
Err(e) => {
tracing::error!("Failed to initialize KMS: {}", e);
None
}
}
} else {
None
};
let encryption = if config.encryption_enabled {
match myfsio_crypto::kms::load_or_create_master_key(&keys_dir).await {
Ok(master_key) => {
Some(Arc::new(EncryptionService::new(master_key, kms.clone())))
}
Err(e) => {
tracing::error!("Failed to initialize encryption: {}", e);
None
}
}
} else {
None
};
state.encryption = encryption;
state.kms = kms;
state
}
}

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,26 @@
[package]
name = "myfsio-storage"
version = "0.1.0"
edition = "2021"
[dependencies]
myfsio-common = { path = "../myfsio-common" }
myfsio-crypto = { path = "../myfsio-crypto" }
serde = { workspace = true }
serde_json = { workspace = true }
tokio = { workspace = true }
dashmap = { workspace = true }
parking_lot = { workspace = true }
uuid = { workspace = true }
chrono = { workspace = true }
thiserror = { workspace = true }
tracing = { workspace = true }
regex = { workspace = true }
unicode-normalization = { workspace = true }
md-5 = { workspace = true }
sha2 = { workspace = true }
hex = { workspace = true }
[dev-dependencies]
tokio = { workspace = true, features = ["macros", "rt-multi-thread"] }
tempfile = "3"

View File

@@ -0,0 +1,59 @@
use myfsio_common::error::{S3Error, S3ErrorCode};
use thiserror::Error;
#[derive(Debug, Error)]
pub enum StorageError {
#[error("Bucket not found: {0}")]
BucketNotFound(String),
#[error("Bucket already exists: {0}")]
BucketAlreadyExists(String),
#[error("Bucket not empty: {0}")]
BucketNotEmpty(String),
#[error("Object not found: {bucket}/{key}")]
ObjectNotFound { bucket: String, key: String },
#[error("Invalid bucket name: {0}")]
InvalidBucketName(String),
#[error("Invalid object key: {0}")]
InvalidObjectKey(String),
#[error("Upload not found: {0}")]
UploadNotFound(String),
#[error("Quota exceeded: {0}")]
QuotaExceeded(String),
#[error("IO error: {0}")]
Io(#[from] std::io::Error),
#[error("JSON error: {0}")]
Json(#[from] serde_json::Error),
#[error("Internal error: {0}")]
Internal(String),
}
impl From<StorageError> for S3Error {
fn from(err: StorageError) -> Self {
match err {
StorageError::BucketNotFound(name) => {
S3Error::from_code(S3ErrorCode::NoSuchBucket).with_resource(format!("/{}", name))
}
StorageError::BucketAlreadyExists(name) => {
S3Error::from_code(S3ErrorCode::BucketAlreadyExists)
.with_resource(format!("/{}", name))
}
StorageError::BucketNotEmpty(name) => {
S3Error::from_code(S3ErrorCode::BucketNotEmpty)
.with_resource(format!("/{}", name))
}
StorageError::ObjectNotFound { bucket, key } => {
S3Error::from_code(S3ErrorCode::NoSuchKey)
.with_resource(format!("/{}/{}", bucket, key))
}
StorageError::InvalidBucketName(msg) => S3Error::new(S3ErrorCode::InvalidBucketName, msg),
StorageError::InvalidObjectKey(msg) => S3Error::new(S3ErrorCode::InvalidKey, msg),
StorageError::UploadNotFound(id) => {
S3Error::new(S3ErrorCode::NoSuchUpload, format!("Upload {} not found", id))
}
StorageError::QuotaExceeded(msg) => S3Error::new(S3ErrorCode::QuotaExceeded, msg),
StorageError::Io(e) => S3Error::new(S3ErrorCode::InternalError, e.to_string()),
StorageError::Json(e) => S3Error::new(S3ErrorCode::InternalError, e.to_string()),
StorageError::Internal(msg) => S3Error::new(S3ErrorCode::InternalError, msg),
}
}
}

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,4 @@
pub mod validation;
pub mod traits;
pub mod error;
pub mod fs_backend;

View File

@@ -0,0 +1,125 @@
use crate::error::StorageError;
use myfsio_common::types::*;
use std::collections::HashMap;
use std::path::PathBuf;
use std::pin::Pin;
use tokio::io::AsyncRead;
pub type StorageResult<T> = Result<T, StorageError>;
pub type AsyncReadStream = Pin<Box<dyn AsyncRead + Send>>;
#[allow(async_fn_in_trait)]
pub trait StorageEngine: Send + Sync {
async fn list_buckets(&self) -> StorageResult<Vec<BucketMeta>>;
async fn create_bucket(&self, name: &str) -> StorageResult<()>;
async fn delete_bucket(&self, name: &str) -> StorageResult<()>;
async fn bucket_exists(&self, name: &str) -> StorageResult<bool>;
async fn bucket_stats(&self, name: &str) -> StorageResult<BucketStats>;
async fn put_object(
&self,
bucket: &str,
key: &str,
stream: AsyncReadStream,
metadata: Option<HashMap<String, String>>,
) -> StorageResult<ObjectMeta>;
async fn get_object(&self, bucket: &str, key: &str) -> StorageResult<(ObjectMeta, AsyncReadStream)>;
async fn get_object_path(&self, bucket: &str, key: &str) -> StorageResult<PathBuf>;
async fn head_object(&self, bucket: &str, key: &str) -> StorageResult<ObjectMeta>;
async fn delete_object(&self, bucket: &str, key: &str) -> StorageResult<()>;
async fn copy_object(
&self,
src_bucket: &str,
src_key: &str,
dst_bucket: &str,
dst_key: &str,
) -> StorageResult<ObjectMeta>;
async fn get_object_metadata(
&self,
bucket: &str,
key: &str,
) -> StorageResult<HashMap<String, String>>;
async fn put_object_metadata(
&self,
bucket: &str,
key: &str,
metadata: &HashMap<String, String>,
) -> StorageResult<()>;
async fn list_objects(&self, bucket: &str, params: &ListParams) -> StorageResult<ListObjectsResult>;
async fn list_objects_shallow(
&self,
bucket: &str,
params: &ShallowListParams,
) -> StorageResult<ShallowListResult>;
async fn initiate_multipart(
&self,
bucket: &str,
key: &str,
metadata: Option<HashMap<String, String>>,
) -> StorageResult<String>;
async fn upload_part(
&self,
bucket: &str,
upload_id: &str,
part_number: u32,
stream: AsyncReadStream,
) -> StorageResult<String>;
async fn complete_multipart(
&self,
bucket: &str,
upload_id: &str,
parts: &[PartInfo],
) -> StorageResult<ObjectMeta>;
async fn abort_multipart(&self, bucket: &str, upload_id: &str) -> StorageResult<()>;
async fn list_parts(&self, bucket: &str, upload_id: &str) -> StorageResult<Vec<PartMeta>>;
async fn list_multipart_uploads(
&self,
bucket: &str,
) -> StorageResult<Vec<MultipartUploadInfo>>;
async fn get_bucket_config(&self, bucket: &str) -> StorageResult<BucketConfig>;
async fn set_bucket_config(&self, bucket: &str, config: &BucketConfig) -> StorageResult<()>;
async fn is_versioning_enabled(&self, bucket: &str) -> StorageResult<bool>;
async fn set_versioning(&self, bucket: &str, enabled: bool) -> StorageResult<()>;
async fn list_object_versions(
&self,
bucket: &str,
key: &str,
) -> StorageResult<Vec<VersionInfo>>;
async fn get_object_tags(
&self,
bucket: &str,
key: &str,
) -> StorageResult<Vec<Tag>>;
async fn set_object_tags(
&self,
bucket: &str,
key: &str,
tags: &[Tag],
) -> StorageResult<()>;
async fn delete_object_tags(
&self,
bucket: &str,
key: &str,
) -> StorageResult<()>;
}

View File

@@ -0,0 +1,194 @@
use std::sync::LazyLock;
use unicode_normalization::UnicodeNormalization;
const WINDOWS_RESERVED: &[&str] = &[
"CON", "PRN", "AUX", "NUL", "COM0", "COM1", "COM2", "COM3", "COM4", "COM5", "COM6", "COM7",
"COM8", "COM9", "LPT0", "LPT1", "LPT2", "LPT3", "LPT4", "LPT5", "LPT6", "LPT7", "LPT8",
"LPT9",
];
const WINDOWS_ILLEGAL_CHARS: &[char] = &['<', '>', ':', '"', '/', '\\', '|', '?', '*'];
const INTERNAL_FOLDERS: &[&str] = &[".meta", ".versions", ".multipart"];
const SYSTEM_ROOT: &str = ".myfsio.sys";
static IP_REGEX: LazyLock<regex::Regex> =
LazyLock::new(|| regex::Regex::new(r"^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}$").unwrap());
pub fn validate_object_key(
object_key: &str,
max_length_bytes: usize,
is_windows: bool,
reserved_prefixes: Option<&[&str]>,
) -> Option<String> {
if object_key.is_empty() {
return Some("Object key required".to_string());
}
if object_key.contains('\0') {
return Some("Object key contains null bytes".to_string());
}
let normalized: String = object_key.nfc().collect();
if normalized.len() > max_length_bytes {
return Some(format!(
"Object key exceeds maximum length of {} bytes",
max_length_bytes
));
}
if normalized.starts_with('/') || normalized.starts_with('\\') {
return Some("Object key cannot start with a slash".to_string());
}
let parts: Vec<&str> = if cfg!(windows) || is_windows {
normalized.split(['/', '\\']).collect()
} else {
normalized.split('/').collect()
};
for part in &parts {
if part.is_empty() {
continue;
}
if *part == ".." {
return Some("Object key contains parent directory references".to_string());
}
if *part == "." {
return Some("Object key contains invalid segments".to_string());
}
if part.chars().any(|c| (c as u32) < 32) {
return Some("Object key contains control characters".to_string());
}
if is_windows {
if part.chars().any(|c| WINDOWS_ILLEGAL_CHARS.contains(&c)) {
return Some(
"Object key contains characters not supported on Windows filesystems"
.to_string(),
);
}
if part.ends_with(' ') || part.ends_with('.') {
return Some(
"Object key segments cannot end with spaces or periods on Windows".to_string(),
);
}
let trimmed = part.trim_end_matches(['.', ' ']).to_uppercase();
if WINDOWS_RESERVED.contains(&trimmed.as_str()) {
return Some(format!("Invalid filename segment: {}", part));
}
}
}
let non_empty_parts: Vec<&str> = parts.iter().filter(|p| !p.is_empty()).copied().collect();
if let Some(top) = non_empty_parts.first() {
if INTERNAL_FOLDERS.contains(top) || *top == SYSTEM_ROOT {
return Some("Object key uses a reserved prefix".to_string());
}
if let Some(prefixes) = reserved_prefixes {
for prefix in prefixes {
if *top == *prefix {
return Some("Object key uses a reserved prefix".to_string());
}
}
}
}
None
}
pub fn validate_bucket_name(bucket_name: &str) -> Option<String> {
let len = bucket_name.len();
if len < 3 || len > 63 {
return Some("Bucket name must be between 3 and 63 characters".to_string());
}
let bytes = bucket_name.as_bytes();
if !bytes[0].is_ascii_lowercase() && !bytes[0].is_ascii_digit() {
return Some(
"Bucket name must start and end with a lowercase letter or digit".to_string(),
);
}
if !bytes[len - 1].is_ascii_lowercase() && !bytes[len - 1].is_ascii_digit() {
return Some(
"Bucket name must start and end with a lowercase letter or digit".to_string(),
);
}
for &b in bytes {
if !b.is_ascii_lowercase() && !b.is_ascii_digit() && b != b'.' && b != b'-' {
return Some(
"Bucket name can only contain lowercase letters, digits, dots, and hyphens"
.to_string(),
);
}
}
if bucket_name.contains("..") {
return Some("Bucket name must not contain consecutive periods".to_string());
}
if IP_REGEX.is_match(bucket_name) {
return Some("Bucket name must not be formatted as an IP address".to_string());
}
None
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_valid_bucket_names() {
assert!(validate_bucket_name("my-bucket").is_none());
assert!(validate_bucket_name("test123").is_none());
assert!(validate_bucket_name("my.bucket.name").is_none());
}
#[test]
fn test_invalid_bucket_names() {
assert!(validate_bucket_name("ab").is_some());
assert!(validate_bucket_name("My-Bucket").is_some());
assert!(validate_bucket_name("-bucket").is_some());
assert!(validate_bucket_name("bucket-").is_some());
assert!(validate_bucket_name("my..bucket").is_some());
assert!(validate_bucket_name("192.168.1.1").is_some());
}
#[test]
fn test_valid_object_keys() {
assert!(validate_object_key("file.txt", 1024, false, None).is_none());
assert!(validate_object_key("path/to/file.txt", 1024, false, None).is_none());
assert!(validate_object_key("a", 1024, false, None).is_none());
}
#[test]
fn test_invalid_object_keys() {
assert!(validate_object_key("", 1024, false, None).is_some());
assert!(validate_object_key("/leading-slash", 1024, false, None).is_some());
assert!(validate_object_key("path/../escape", 1024, false, None).is_some());
assert!(validate_object_key(".myfsio.sys/secret", 1024, false, None).is_some());
assert!(validate_object_key(".meta/data", 1024, false, None).is_some());
}
#[test]
fn test_object_key_max_length() {
let long_key = "a".repeat(1025);
assert!(validate_object_key(&long_key, 1024, false, None).is_some());
let ok_key = "a".repeat(1024);
assert!(validate_object_key(&ok_key, 1024, false, None).is_none());
}
#[test]
fn test_windows_validation() {
assert!(validate_object_key("CON", 1024, true, None).is_some());
assert!(validate_object_key("file<name", 1024, true, None).is_some());
assert!(validate_object_key("file.txt ", 1024, true, None).is_some());
}
}

View File

@@ -0,0 +1,10 @@
[package]
name = "myfsio-xml"
version = "0.1.0"
edition = "2021"
[dependencies]
myfsio-common = { path = "../myfsio-common" }
quick-xml = { workspace = true }
serde = { workspace = true }
chrono = { workspace = true }

View File

@@ -0,0 +1,14 @@
pub mod response;
pub mod request;
use quick_xml::Writer;
use std::io::Cursor;
pub fn write_xml_element(tag: &str, text: &str) -> String {
let mut writer = Writer::new(Cursor::new(Vec::new()));
writer
.create_element(tag)
.write_text_content(quick_xml::events::BytesText::new(text))
.unwrap();
String::from_utf8(writer.into_inner().into_inner()).unwrap()
}

View File

@@ -0,0 +1,159 @@
use quick_xml::events::Event;
use quick_xml::Reader;
#[derive(Debug, Default)]
pub struct DeleteObjectsRequest {
pub objects: Vec<ObjectIdentifier>,
pub quiet: bool,
}
#[derive(Debug)]
pub struct ObjectIdentifier {
pub key: String,
pub version_id: Option<String>,
}
#[derive(Debug, Default)]
pub struct CompleteMultipartUpload {
pub parts: Vec<CompletedPart>,
}
#[derive(Debug)]
pub struct CompletedPart {
pub part_number: u32,
pub etag: String,
}
pub fn parse_complete_multipart_upload(xml: &str) -> Result<CompleteMultipartUpload, String> {
let mut reader = Reader::from_str(xml);
let mut result = CompleteMultipartUpload::default();
let mut buf = Vec::new();
let mut current_tag = String::new();
let mut part_number: Option<u32> = None;
let mut etag: Option<String> = None;
let mut in_part = false;
loop {
match reader.read_event_into(&mut buf) {
Ok(Event::Start(ref e)) => {
let name = String::from_utf8_lossy(e.name().as_ref()).to_string();
current_tag = name.clone();
if name == "Part" {
in_part = true;
part_number = None;
etag = None;
}
}
Ok(Event::Text(ref e)) => {
if in_part {
let text = e.unescape().map_err(|e| e.to_string())?.to_string();
match current_tag.as_str() {
"PartNumber" => {
part_number = Some(text.trim().parse().map_err(|e: std::num::ParseIntError| e.to_string())?);
}
"ETag" => {
etag = Some(text.trim().trim_matches('"').to_string());
}
_ => {}
}
}
}
Ok(Event::End(ref e)) => {
let name = String::from_utf8_lossy(e.name().as_ref()).to_string();
if name == "Part" && in_part {
if let (Some(pn), Some(et)) = (part_number.take(), etag.take()) {
result.parts.push(CompletedPart {
part_number: pn,
etag: et,
});
}
in_part = false;
}
}
Ok(Event::Eof) => break,
Err(e) => return Err(format!("XML parse error: {}", e)),
_ => {}
}
buf.clear();
}
result.parts.sort_by_key(|p| p.part_number);
Ok(result)
}
pub fn parse_delete_objects(xml: &str) -> Result<DeleteObjectsRequest, String> {
let mut reader = Reader::from_str(xml);
let mut result = DeleteObjectsRequest::default();
let mut buf = Vec::new();
let mut current_tag = String::new();
let mut current_key: Option<String> = None;
let mut current_version_id: Option<String> = None;
let mut in_object = false;
loop {
match reader.read_event_into(&mut buf) {
Ok(Event::Start(ref e)) => {
let name = String::from_utf8_lossy(e.name().as_ref()).to_string();
current_tag = name.clone();
if name == "Object" {
in_object = true;
current_key = None;
current_version_id = None;
}
}
Ok(Event::Text(ref e)) => {
let text = e.unescape().map_err(|e| e.to_string())?.to_string();
match current_tag.as_str() {
"Key" if in_object => {
current_key = Some(text.trim().to_string());
}
"VersionId" if in_object => {
current_version_id = Some(text.trim().to_string());
}
"Quiet" => {
result.quiet = text.trim() == "true";
}
_ => {}
}
}
Ok(Event::End(ref e)) => {
let name = String::from_utf8_lossy(e.name().as_ref()).to_string();
if name == "Object" && in_object {
if let Some(key) = current_key.take() {
result.objects.push(ObjectIdentifier {
key,
version_id: current_version_id.take(),
});
}
in_object = false;
}
}
Ok(Event::Eof) => break,
Err(e) => return Err(format!("XML parse error: {}", e)),
_ => {}
}
buf.clear();
}
Ok(result)
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_parse_complete_multipart() {
let xml = r#"<CompleteMultipartUpload>
<Part><PartNumber>2</PartNumber><ETag>"etag2"</ETag></Part>
<Part><PartNumber>1</PartNumber><ETag>"etag1"</ETag></Part>
</CompleteMultipartUpload>"#;
let result = parse_complete_multipart_upload(xml).unwrap();
assert_eq!(result.parts.len(), 2);
assert_eq!(result.parts[0].part_number, 1);
assert_eq!(result.parts[0].etag, "etag1");
assert_eq!(result.parts[1].part_number, 2);
assert_eq!(result.parts[1].etag, "etag2");
}
}

View File

@@ -0,0 +1,269 @@
use myfsio_common::types::{BucketMeta, ObjectMeta};
use quick_xml::events::{BytesDecl, BytesEnd, BytesStart, BytesText, Event};
use quick_xml::Writer;
use std::io::Cursor;
pub fn list_buckets_xml(owner_id: &str, owner_name: &str, buckets: &[BucketMeta]) -> String {
let mut writer = Writer::new(Cursor::new(Vec::new()));
writer.write_event(Event::Decl(BytesDecl::new("1.0", Some("UTF-8"), None))).unwrap();
let start = BytesStart::new("ListAllMyBucketsResult")
.with_attributes([("xmlns", "http://s3.amazonaws.com/doc/2006-03-01/")]);
writer.write_event(Event::Start(start)).unwrap();
writer.write_event(Event::Start(BytesStart::new("Owner"))).unwrap();
write_text_element(&mut writer, "ID", owner_id);
write_text_element(&mut writer, "DisplayName", owner_name);
writer.write_event(Event::End(BytesEnd::new("Owner"))).unwrap();
writer.write_event(Event::Start(BytesStart::new("Buckets"))).unwrap();
for bucket in buckets {
writer.write_event(Event::Start(BytesStart::new("Bucket"))).unwrap();
write_text_element(&mut writer, "Name", &bucket.name);
write_text_element(&mut writer, "CreationDate", &bucket.creation_date.to_rfc3339());
writer.write_event(Event::End(BytesEnd::new("Bucket"))).unwrap();
}
writer.write_event(Event::End(BytesEnd::new("Buckets"))).unwrap();
writer.write_event(Event::End(BytesEnd::new("ListAllMyBucketsResult"))).unwrap();
String::from_utf8(writer.into_inner().into_inner()).unwrap()
}
pub fn list_objects_v2_xml(
bucket_name: &str,
prefix: &str,
delimiter: &str,
max_keys: usize,
objects: &[ObjectMeta],
common_prefixes: &[String],
is_truncated: bool,
continuation_token: Option<&str>,
next_continuation_token: Option<&str>,
key_count: usize,
) -> String {
let mut writer = Writer::new(Cursor::new(Vec::new()));
writer.write_event(Event::Decl(BytesDecl::new("1.0", Some("UTF-8"), None))).unwrap();
let start = BytesStart::new("ListBucketResult")
.with_attributes([("xmlns", "http://s3.amazonaws.com/doc/2006-03-01/")]);
writer.write_event(Event::Start(start)).unwrap();
write_text_element(&mut writer, "Name", bucket_name);
write_text_element(&mut writer, "Prefix", prefix);
if !delimiter.is_empty() {
write_text_element(&mut writer, "Delimiter", delimiter);
}
write_text_element(&mut writer, "MaxKeys", &max_keys.to_string());
write_text_element(&mut writer, "KeyCount", &key_count.to_string());
write_text_element(&mut writer, "IsTruncated", &is_truncated.to_string());
if let Some(token) = continuation_token {
write_text_element(&mut writer, "ContinuationToken", token);
}
if let Some(token) = next_continuation_token {
write_text_element(&mut writer, "NextContinuationToken", token);
}
for obj in objects {
writer.write_event(Event::Start(BytesStart::new("Contents"))).unwrap();
write_text_element(&mut writer, "Key", &obj.key);
write_text_element(&mut writer, "LastModified", &obj.last_modified.to_rfc3339());
if let Some(ref etag) = obj.etag {
write_text_element(&mut writer, "ETag", &format!("\"{}\"", etag));
}
write_text_element(&mut writer, "Size", &obj.size.to_string());
write_text_element(&mut writer, "StorageClass", obj.storage_class.as_deref().unwrap_or("STANDARD"));
writer.write_event(Event::End(BytesEnd::new("Contents"))).unwrap();
}
for prefix in common_prefixes {
writer.write_event(Event::Start(BytesStart::new("CommonPrefixes"))).unwrap();
write_text_element(&mut writer, "Prefix", prefix);
writer.write_event(Event::End(BytesEnd::new("CommonPrefixes"))).unwrap();
}
writer.write_event(Event::End(BytesEnd::new("ListBucketResult"))).unwrap();
String::from_utf8(writer.into_inner().into_inner()).unwrap()
}
fn write_text_element(writer: &mut Writer<Cursor<Vec<u8>>>, tag: &str, text: &str) {
writer.write_event(Event::Start(BytesStart::new(tag))).unwrap();
writer.write_event(Event::Text(BytesText::new(text))).unwrap();
writer.write_event(Event::End(BytesEnd::new(tag))).unwrap();
}
pub fn initiate_multipart_upload_xml(bucket: &str, key: &str, upload_id: &str) -> String {
let mut writer = Writer::new(Cursor::new(Vec::new()));
writer.write_event(Event::Decl(BytesDecl::new("1.0", Some("UTF-8"), None))).unwrap();
let start = BytesStart::new("InitiateMultipartUploadResult")
.with_attributes([("xmlns", "http://s3.amazonaws.com/doc/2006-03-01/")]);
writer.write_event(Event::Start(start)).unwrap();
write_text_element(&mut writer, "Bucket", bucket);
write_text_element(&mut writer, "Key", key);
write_text_element(&mut writer, "UploadId", upload_id);
writer.write_event(Event::End(BytesEnd::new("InitiateMultipartUploadResult"))).unwrap();
String::from_utf8(writer.into_inner().into_inner()).unwrap()
}
pub fn complete_multipart_upload_xml(
bucket: &str,
key: &str,
etag: &str,
location: &str,
) -> String {
let mut writer = Writer::new(Cursor::new(Vec::new()));
writer.write_event(Event::Decl(BytesDecl::new("1.0", Some("UTF-8"), None))).unwrap();
let start = BytesStart::new("CompleteMultipartUploadResult")
.with_attributes([("xmlns", "http://s3.amazonaws.com/doc/2006-03-01/")]);
writer.write_event(Event::Start(start)).unwrap();
write_text_element(&mut writer, "Location", location);
write_text_element(&mut writer, "Bucket", bucket);
write_text_element(&mut writer, "Key", key);
write_text_element(&mut writer, "ETag", &format!("\"{}\"", etag));
writer.write_event(Event::End(BytesEnd::new("CompleteMultipartUploadResult"))).unwrap();
String::from_utf8(writer.into_inner().into_inner()).unwrap()
}
pub fn copy_object_result_xml(etag: &str, last_modified: &str) -> String {
let mut writer = Writer::new(Cursor::new(Vec::new()));
writer.write_event(Event::Decl(BytesDecl::new("1.0", Some("UTF-8"), None))).unwrap();
let start = BytesStart::new("CopyObjectResult")
.with_attributes([("xmlns", "http://s3.amazonaws.com/doc/2006-03-01/")]);
writer.write_event(Event::Start(start)).unwrap();
write_text_element(&mut writer, "ETag", &format!("\"{}\"", etag));
write_text_element(&mut writer, "LastModified", last_modified);
writer.write_event(Event::End(BytesEnd::new("CopyObjectResult"))).unwrap();
String::from_utf8(writer.into_inner().into_inner()).unwrap()
}
pub fn delete_result_xml(
deleted: &[(String, Option<String>)],
errors: &[(String, String, String)],
quiet: bool,
) -> String {
let mut writer = Writer::new(Cursor::new(Vec::new()));
writer.write_event(Event::Decl(BytesDecl::new("1.0", Some("UTF-8"), None))).unwrap();
let start = BytesStart::new("DeleteResult")
.with_attributes([("xmlns", "http://s3.amazonaws.com/doc/2006-03-01/")]);
writer.write_event(Event::Start(start)).unwrap();
if !quiet {
for (key, version_id) in deleted {
writer.write_event(Event::Start(BytesStart::new("Deleted"))).unwrap();
write_text_element(&mut writer, "Key", key);
if let Some(vid) = version_id {
write_text_element(&mut writer, "VersionId", vid);
}
writer.write_event(Event::End(BytesEnd::new("Deleted"))).unwrap();
}
}
for (key, code, message) in errors {
writer.write_event(Event::Start(BytesStart::new("Error"))).unwrap();
write_text_element(&mut writer, "Key", key);
write_text_element(&mut writer, "Code", code);
write_text_element(&mut writer, "Message", message);
writer.write_event(Event::End(BytesEnd::new("Error"))).unwrap();
}
writer.write_event(Event::End(BytesEnd::new("DeleteResult"))).unwrap();
String::from_utf8(writer.into_inner().into_inner()).unwrap()
}
pub fn list_multipart_uploads_xml(
bucket: &str,
uploads: &[myfsio_common::types::MultipartUploadInfo],
) -> String {
let mut writer = Writer::new(Cursor::new(Vec::new()));
writer.write_event(Event::Decl(BytesDecl::new("1.0", Some("UTF-8"), None))).unwrap();
let start = BytesStart::new("ListMultipartUploadsResult")
.with_attributes([("xmlns", "http://s3.amazonaws.com/doc/2006-03-01/")]);
writer.write_event(Event::Start(start)).unwrap();
write_text_element(&mut writer, "Bucket", bucket);
for upload in uploads {
writer.write_event(Event::Start(BytesStart::new("Upload"))).unwrap();
write_text_element(&mut writer, "Key", &upload.key);
write_text_element(&mut writer, "UploadId", &upload.upload_id);
write_text_element(&mut writer, "Initiated", &upload.initiated.to_rfc3339());
writer.write_event(Event::End(BytesEnd::new("Upload"))).unwrap();
}
writer.write_event(Event::End(BytesEnd::new("ListMultipartUploadsResult"))).unwrap();
String::from_utf8(writer.into_inner().into_inner()).unwrap()
}
pub fn list_parts_xml(
bucket: &str,
key: &str,
upload_id: &str,
parts: &[myfsio_common::types::PartMeta],
) -> String {
let mut writer = Writer::new(Cursor::new(Vec::new()));
writer.write_event(Event::Decl(BytesDecl::new("1.0", Some("UTF-8"), None))).unwrap();
let start = BytesStart::new("ListPartsResult")
.with_attributes([("xmlns", "http://s3.amazonaws.com/doc/2006-03-01/")]);
writer.write_event(Event::Start(start)).unwrap();
write_text_element(&mut writer, "Bucket", bucket);
write_text_element(&mut writer, "Key", key);
write_text_element(&mut writer, "UploadId", upload_id);
for part in parts {
writer.write_event(Event::Start(BytesStart::new("Part"))).unwrap();
write_text_element(&mut writer, "PartNumber", &part.part_number.to_string());
write_text_element(&mut writer, "ETag", &format!("\"{}\"", part.etag));
write_text_element(&mut writer, "Size", &part.size.to_string());
if let Some(ref lm) = part.last_modified {
write_text_element(&mut writer, "LastModified", &lm.to_rfc3339());
}
writer.write_event(Event::End(BytesEnd::new("Part"))).unwrap();
}
writer.write_event(Event::End(BytesEnd::new("ListPartsResult"))).unwrap();
String::from_utf8(writer.into_inner().into_inner()).unwrap()
}
#[cfg(test)]
mod tests {
use super::*;
use chrono::Utc;
#[test]
fn test_list_buckets_xml() {
let buckets = vec![BucketMeta {
name: "test-bucket".to_string(),
creation_date: Utc::now(),
}];
let xml = list_buckets_xml("owner-id", "owner-name", &buckets);
assert!(xml.contains("<Name>test-bucket</Name>"));
assert!(xml.contains("<ID>owner-id</ID>"));
assert!(xml.contains("ListAllMyBucketsResult"));
}
#[test]
fn test_list_objects_v2_xml() {
let objects = vec![ObjectMeta::new("file.txt".to_string(), 1024, Utc::now())];
let xml = list_objects_v2_xml(
"my-bucket", "", "/", 1000, &objects, &[], false, None, None, 1,
);
assert!(xml.contains("<Key>file.txt</Key>"));
assert!(xml.contains("<Size>1024</Size>"));
assert!(xml.contains("<IsTruncated>false</IsTruncated>"));
}
}

56
run.py
View File

@@ -5,6 +5,7 @@ import argparse
import atexit import atexit
import os import os
import signal import signal
import subprocess
import sys import sys
import warnings import warnings
import multiprocessing import multiprocessing
@@ -74,6 +75,48 @@ def _serve_granian(target: str, port: int, config: Optional[AppConfig] = None) -
server.serve() server.serve()
def _find_rust_binary() -> Optional[Path]:
candidates = [
Path(__file__).parent / "myfsio-engine" / "target" / "release" / "myfsio-server.exe",
Path(__file__).parent / "myfsio-engine" / "target" / "release" / "myfsio-server",
Path(__file__).parent / "myfsio-engine" / "target" / "debug" / "myfsio-server.exe",
Path(__file__).parent / "myfsio-engine" / "target" / "debug" / "myfsio-server",
]
for p in candidates:
if p.exists():
return p
return None
def serve_rust_api(port: int, config: Optional[AppConfig] = None) -> None:
binary = _find_rust_binary()
if binary is None:
print("ERROR: Rust engine binary not found. Build it first:")
print(" cd myfsio-engine && cargo build --release")
sys.exit(1)
env = os.environ.copy()
env["PORT"] = str(port)
env["HOST"] = _server_host()
if config:
env["STORAGE_ROOT"] = str(config.storage_root)
env["AWS_REGION"] = config.aws_region
if config.secret_key:
env["SECRET_KEY"] = config.secret_key
env.setdefault("ENCRYPTION_ENABLED", str(config.encryption_enabled).lower())
env.setdefault("KMS_ENABLED", str(config.kms_enabled).lower())
env.setdefault("LIFECYCLE_ENABLED", str(config.lifecycle_enabled).lower())
env.setdefault("RUST_LOG", "info")
print(f"Starting Rust S3 engine: {binary}")
proc = subprocess.Popen([str(binary)], env=env)
try:
proc.wait()
except KeyboardInterrupt:
proc.terminate()
proc.wait(timeout=5)
def serve_api(port: int, prod: bool = False, config: Optional[AppConfig] = None) -> None: def serve_api(port: int, prod: bool = False, config: Optional[AppConfig] = None) -> None:
if prod: if prod:
_serve_granian("app:create_api_app", port, config) _serve_granian("app:create_api_app", port, config)
@@ -227,6 +270,7 @@ if __name__ == "__main__":
parser.add_argument("--ui-port", type=int, default=5100) parser.add_argument("--ui-port", type=int, default=5100)
parser.add_argument("--prod", action="store_true", help="Run in production mode using Granian") parser.add_argument("--prod", action="store_true", help="Run in production mode using Granian")
parser.add_argument("--dev", action="store_true", help="Force development mode (Flask dev server)") parser.add_argument("--dev", action="store_true", help="Force development mode (Flask dev server)")
parser.add_argument("--engine", choices=["python", "rust"], default="python", help="API engine: python (Flask) or rust (myfsio-engine)")
parser.add_argument("--check-config", action="store_true", help="Validate configuration and exit") parser.add_argument("--check-config", action="store_true", help="Validate configuration and exit")
parser.add_argument("--show-config", action="store_true", help="Show configuration summary and exit") parser.add_argument("--show-config", action="store_true", help="Show configuration summary and exit")
parser.add_argument("--reset-cred", action="store_true", help="Reset admin credentials and exit") parser.add_argument("--reset-cred", action="store_true", help="Reset admin credentials and exit")
@@ -280,9 +324,17 @@ if __name__ == "__main__":
else: else:
print("Running in development mode (Flask dev server)") print("Running in development mode (Flask dev server)")
use_rust = args.engine == "rust"
if args.mode in {"api", "both"}: if args.mode in {"api", "both"}:
print(f"Starting API server on port {args.api_port}...") if use_rust:
api_proc = Process(target=serve_api, args=(args.api_port, prod_mode, config)) print(f"Starting Rust API engine on port {args.api_port}...")
else:
print(f"Starting API server on port {args.api_port}...")
if use_rust:
api_proc = Process(target=serve_rust_api, args=(args.api_port, config))
else:
api_proc = Process(target=serve_api, args=(args.api_port, prod_mode, config))
api_proc.start() api_proc.start()
else: else:
api_proc = None api_proc = None