wx-cli/src/scanner/linux.rs

/// Linux WeChat 进程内存密钥扫描器
///
/// 通过 /proc/<pid>/maps 枚举内存区域，
/// 通过 /proc/<pid>/mem 读取内存内容，
/// 搜索 x'<64hex><32hex>' 格式的 SQLCipher 密钥
use anyhow::{Context, Result};
use std::collections::HashMap;
use std::io::{Read, Seek, SeekFrom};
use std::path::Path;

use super::{collect_db_salts, KeyEntry};

const HEX_PATTERN_LEN: usize = 96;
const CHUNK_SIZE: usize = 2 * 1024 * 1024;

/// 查找 WeChat 进程 PID
fn find_wechat_pid() -> Option<u32> {
    let proc_dir = std::fs::read_dir("/proc").ok()?;
    for entry in proc_dir.flatten() {
        let name = entry.file_name();
        let name_str = name.to_string_lossy();
        // 只处理数字目录（PID）
        if !name_str.chars().all(|c| c.is_ascii_digit()) {
            continue;
        }
        let comm_path = format!("/proc/{}/comm", name_str);
        if let Ok(comm) = std::fs::read_to_string(&comm_path) {
            let comm = comm.trim().to_lowercase();
            if comm == "wechat" || comm == "weixin" {
                if let Ok(pid) = name_str.parse::<u32>() {
                    return Some(pid);
                }
            }
        }
    }
    None
}

/// 解析 /proc/<pid>/maps 文件，返回可读的内存区域 (start, end)
fn parse_maps(pid: u32) -> Result<Vec<(u64, u64)>> {
    let maps_path = format!("/proc/{}/maps", pid);
    let content = std::fs::read_to_string(&maps_path)
        .with_context(|| format!("读取 {} 失败", maps_path))?;

    let mut regions = Vec::new();
    for line in content.lines() {
        // 格式: start-end perms offset dev inode pathname
        let parts: Vec<&str> = line.splitn(2, ' ').collect();
        if parts.len() < 2 {
            continue;
        }
        let perms = parts[1].trim_start();
        // 只选取 r 和 w 权限的区域
        if !perms.starts_with("rw") {
            continue;
        }
        let addr_parts: Vec<&str> = parts[0].splitn(2, '-').collect();
        if addr_parts.len() != 2 {
            continue;
        }
        if let (Ok(start), Ok(end)) = (
            u64::from_str_radix(addr_parts[0], 16),
            u64::from_str_radix(addr_parts[1], 16),
        ) {
            regions.push((start, end));
        }
    }
    Ok(regions)
}

pub fn scan_keys(db_dir: &Path) -> Result<Vec<KeyEntry>> {
    let pid = find_wechat_pid()
        .context("找不到 WeChat 进程，请确认 WeChat 正在运行")?;
    eprintln!("WeChat PID: {}", pid);

    let db_salts = collect_db_salts(db_dir);
    eprintln!("找到 {} 个加密数据库", db_salts.len());

    eprintln!("扫描进程内存...");
    let regions = parse_maps(pid)?;
    eprintln!("找到 {} 个可读写内存区域", regions.len());

    let mem_path = format!("/proc/{}/mem", pid);
    let mut mem_file = std::fs::File::open(&mem_path)
        .with_context(|| format!("打开 {} 失败，请以 root 权限运行", mem_path))?;

    let mut raw_keys: Vec<(String, String)> = Vec::new();
    for (start, end) in &regions {
        scan_region(&mut mem_file, *start, *end, &mut raw_keys);
    }
    eprintln!("找到 {} 个候选密钥", raw_keys.len());

    // 无法匹配 DB 文件的密钥用 _by_salt/<salt> 保存，daemon 会做 salt 回退查找，
    // 确保所有在内存中找到的密钥都不会因 DB 文件权限或路径问题而丢失。
    let db_salt_map: HashMap<&str, &str> = db_salts
        .iter()
        .map(|(salt, name)| (salt.as_str(), name.as_str()))
        .collect();

    let mut entries = Vec::new();
    for (key_hex, salt_hex) in &raw_keys {
        let db_name = db_salt_map
            .get(salt_hex.as_str())
            .map(|s| s.to_string())
            .unwrap_or_else(|| format!("_by_salt/{}", salt_hex));
        entries.push(KeyEntry {
            db_name,
            enc_key: key_hex.clone(),
            salt: salt_hex.clone(),
        });
    }

    let matched = entries.iter().filter(|e| !e.db_name.starts_with("_by_salt/")).count();
    eprintln!(
        "匹配到 {}/{} 个密钥（另有 {} 个按 salt 保存）",
        matched,
        raw_keys.len(),
        entries.len() - matched
    );
    Ok(entries)
}

fn scan_region(
    mem: &mut std::fs::File,
    start: u64,
    end: u64,
    results: &mut Vec<(String, String)>,
) {
    let total_len = (end - start) as usize;
    let overlap = HEX_PATTERN_LEN + 3;
    let mut offset = 0usize;

    loop {
        if offset >= total_len {
            break;
        }
        let chunk_size = std::cmp::min(CHUNK_SIZE, total_len - offset);
        let addr = start + offset as u64;

        if mem.seek(SeekFrom::Start(addr)).is_err() {
            break;
        }
        let mut buf = vec![0u8; chunk_size];
        match mem.read(&mut buf) {
            Ok(n) if n > 0 => {
                buf.truncate(n);
                search_pattern(&buf, results);
            }
            _ => {}
        }

        if chunk_size > overlap {
            offset += chunk_size - overlap;
        } else {
            offset += chunk_size;
        }
    }
}

#[inline]
fn is_hex_char(c: u8) -> bool {
    c.is_ascii_hexdigit()
}

fn search_pattern(buf: &[u8], results: &mut Vec<(String, String)>) {
    let total = HEX_PATTERN_LEN + 3;
    if buf.len() < total {
        return;
    }
    let mut i = 0;
    while i + total <= buf.len() {
        if buf[i] != b'x' || buf[i + 1] != b'\'' {
            i += 1;
            continue;
        }
        let hex_start = i + 2;
        let all_hex = buf[hex_start..hex_start + HEX_PATTERN_LEN]
            .iter()
            .all(|&c| is_hex_char(c));
        if !all_hex {
            i += 1;
            continue;
        }
        if buf[hex_start + HEX_PATTERN_LEN] != b'\'' {
            i += 1;
            continue;
        }
        let key_hex = String::from_utf8_lossy(&buf[hex_start..hex_start + 64])
            .to_lowercase();
        let salt_hex = String::from_utf8_lossy(&buf[hex_start + 64..hex_start + 96])
            .to_lowercase();
        let is_dup = results.iter().any(|(k, s)| k == &key_hex && s == &salt_hex);
        if !is_dup {
            results.push((key_hex, salt_hex));
        }
        i += total;
    }
}