Files
guara/crates/guara-core/src/raid_detect.rs
T
Lucas RayanandClaude Opus 5.5 beb83fef92
Release / Cria a release (push) Canceled after 0s
Release / CLI (linux-aarch64) (push) Canceled after 0s
Release / CLI (linux-x86_64) (push) Canceled after 0s
Release / CLI (macos-aarch64) (push) Canceled after 0s
Release / CLI (macos-x86_64) (push) Canceled after 0s
Release / CLI (windows-x86_64) (push) Canceled after 0s
Release / App desktop (linux-x86_64) (push) Canceled after 0s
Release / App desktop (macos-aarch64) (push) Canceled after 0s
Release / App desktop (macos-x86_64) (push) Canceled after 0s
Release / App desktop (windows-x86_64) (push) Canceled after 0s
Release / Publica a release (push) Canceled after 0s
Formatação (cargo fmt) e ajustes do clippy para o CI da 1.0.0
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-19 11:41:58 -03:00

283 lines
8.4 KiB
Rust

//! Descobre sozinho nível, ordem dos discos, tamanho de faixa e layout de um RAID sem
//! metadados, testando as combinações e medindo a coerência do sistema de arquivos que sai.
use crate::partition;
use crate::raid::{Layout, Level, Member, RaidParams, RaidSource};
use crate::source::{read_vec, MemSource, Source};
use crate::types::FsKind;
use std::sync::Arc;
#[derive(Clone, Debug)]
pub struct Candidate {
pub level: Level,
pub chunk: u64,
pub layout: Layout,
/// Índices dos membros fornecidos, na ordem do array.
pub order: Vec<usize>,
pub score: f64,
}
impl Candidate {
pub fn params(&self, members: &[Arc<dyn Source>]) -> RaidParams {
RaidParams::simple(
self.level,
self.chunk,
self.layout,
self.order
.iter()
.map(|&i| {
Some(Member {
src: members[i].clone(),
data_offset: 0,
})
})
.collect(),
)
}
}
const PREFIX: u64 = 24 << 20;
const CHUNKS: [u64; 9] = [
4096, 8192, 16384, 32768, 65536, 131072, 262144, 524288, 1048576,
];
fn permutations(n: usize) -> Vec<Vec<usize>> {
fn rec(cur: &mut Vec<usize>, used: &mut Vec<bool>, n: usize, out: &mut Vec<Vec<usize>>) {
if cur.len() == n {
out.push(cur.clone());
return;
}
for i in 0..n {
if !used[i] {
used[i] = true;
cur.push(i);
rec(cur, used, n, out);
cur.pop();
used[i] = false;
}
}
}
let mut out = vec![];
rec(&mut vec![], &mut vec![false; n], n, &mut out);
out
}
fn ntfs_score(src: &dyn Source, start: u64) -> f64 {
let Ok(b) = read_vec(src, start, 512) else {
return 0.0;
};
if b.len() < 512 || &b[3..11] != b"NTFS " {
return 0.0;
}
let bps = u16::from_le_bytes([b[11], b[12]]) as u64;
let spc = b[13] as u64;
let mft = u64::from_le_bytes(b[0x30..0x38].try_into().unwrap());
let rs = b[0x40] as i8;
let rec = if rs > 0 {
rs as u64 * bps * spc
} else {
1u64 << (-rs)
};
if bps == 0 || spc == 0 || rec == 0 || rec > 65536 {
return 0.0;
}
let mft_off = start + mft * bps * spc;
let (mut ok, mut total) = (0, 0);
for k in [
0u64, 1, 2, 3, 5, 11, 16, 24, 40, 100, 250, 600, 1500, 3000, 6000, 12000, 24000, 50000,
] {
let off = mft_off + k * rec;
if off + rec > src.size() {
break;
}
let Ok(r) = read_vec(src, off, 64) else {
continue;
};
if r.len() < 64 {
continue;
}
// Registros além do fim do MFT estão zerados: não contam contra.
if r.iter().all(|&x| x == 0) {
continue;
}
total += 1;
if &r[..4] == b"FILE" && u32::from_le_bytes(r[0x2C..0x30].try_into().unwrap()) as u64 == k {
ok += 1;
}
}
if total == 0 {
1.0
} else {
1.0 + 10.0 * ok as f64 / total as f64
}
}
/// Coerência do volume que começa em `start`.
fn volume_score(src: &dyn Source, start: u64, fs: FsKind) -> f64 {
match fs {
FsKind::Ntfs => ntfs_score(src, start),
FsKind::Ext => crate::fs::ext::consistency_score(src, start).unwrap_or(0.5),
FsKind::Unknown => 0.0,
_ => 1.0,
}
}
pub fn fs_score(src: &dyn Source) -> f64 {
let layout = partition::analyze(src, false);
let mut best = 0.0f64;
for p in &layout.partitions {
best = best.max(
volume_score(src, p.start, p.fs) + if p.fs != FsKind::Unknown { 0.5 } else { 0.0 },
);
}
best
}
fn first_sector_plausible(src: &dyn Source) -> bool {
let Ok(b) = read_vec(src, 0, 4096) else {
return false;
};
b.len() >= 1100 && ((b[510] == 0x55 && b[511] == 0xAA) || partition::detect_fs_in(&b).is_some())
}
/// Testa as combinações e devolve as melhores (maior pontuação primeiro).
pub fn detect(
members: &[Arc<dyn Source>],
cancelled: &(dyn Fn() -> bool + Sync),
) -> Vec<Candidate> {
let n = members.len();
if n < 2 {
return vec![];
}
// Trabalha sobre um prefixo em memória de cada disco.
let cached: Vec<Arc<dyn Source>> = members
.iter()
.map(|m| {
let len = m.size().min(PREFIX) as usize;
Arc::new(MemSource(read_vec(m.as_ref(), 0, len).unwrap_or_default())) as Arc<dyn Source>
})
.collect();
let min_len = cached.iter().map(|c| c.size()).min().unwrap_or(0);
if min_len < 1 << 20 {
return vec![];
}
let mut out: Vec<Candidate> = vec![];
// RAID 1: cópias idênticas.
let probe = |s: &Arc<dyn Source>| read_vec(s.as_ref(), 0, 1 << 20).unwrap_or_default();
let first = probe(&cached[0]);
if first.iter().any(|&x| x != 0) && cached.iter().skip(1).all(|c| probe(c) == first) {
let c = Candidate {
level: Level::R1,
chunk: 65536,
layout: Layout::LeftSymmetric,
order: (0..n).collect(),
score: fs_score(cached[0].as_ref()) + 1.0,
};
out.push(c);
}
// RAID 5/4 satisfaz "XOR de todos os discos = 0" em toda faixa.
let sample = 4usize << 20;
let bufs: Vec<Vec<u8>> = cached
.iter()
.map(|c| read_vec(c.as_ref(), 0, sample).unwrap_or_default())
.collect();
let mut xor_zero = 0usize;
let mut nonzero_rows = 0usize;
for blk in 0..sample / 4096 {
let mut any = false;
let mut z = true;
for i in 0..4096 {
let mut x = 0u8;
for b in &bufs {
x ^= b[blk * 4096 + i];
any |= b[blk * 4096 + i] != 0;
}
if x != 0 {
z = false;
break;
}
}
if any {
nonzero_rows += 1;
if z {
xor_zero += 1;
}
}
}
let parity5 = nonzero_rows > 8 && xor_zero * 100 >= nonzero_rows * 95;
let mut levels: Vec<Level> = vec![];
if parity5 && n >= 3 {
levels.push(Level::R5);
} else {
levels.push(Level::R0);
if n >= 4 {
levels.push(Level::R6);
}
}
let ok_first: Vec<bool> = cached
.iter()
.map(|c| first_sector_plausible(c.as_ref()))
.collect();
let any_first = ok_first.iter().any(|&b| b);
let perms = permutations(n);
for level in levels {
let layouts: Vec<Layout> = if matches!(level, Level::R0) {
vec![Layout::LeftSymmetric]
} else {
Layout::all_raid5().to_vec()
};
for &chunk in &CHUNKS {
for layout in &layouts {
for order in &perms {
if cancelled() {
return finish(out);
}
let cand = Candidate {
level,
chunk,
layout: *layout,
order: order.clone(),
score: 0.0,
};
let params = cand.params(&cached);
let Ok(r) = RaidSource::new(params) else {
continue;
};
// O primeiro setor do volume precisa estar num disco que o tenha.
if any_first {
let mut b = [0u8; 512];
let _ = r.read_at(0, &mut b);
let m0 = order.iter().position(|&i| {
let mut c = [0u8; 512];
let _ = cached[i].read_at(0, &mut c);
c == b
});
match m0 {
Some(p) if ok_first[order[p]] => {}
_ => continue,
}
}
let score = fs_score(&r);
if score > 0.0 {
out.push(Candidate { score, ..cand });
}
}
}
}
}
finish(out)
}
fn finish(mut out: Vec<Candidate>) -> Vec<Candidate> {
out.sort_by(|a, b| {
b.score
.partial_cmp(&a.score)
.unwrap_or(std::cmp::Ordering::Equal)
});
out.truncate(10);
out
}