如何追踪Rust程序内存使用？引入memory-stats库遇编译错误

问题背景 #

对Rust了解不多，想查看Rust程序的当前内存使用情况，引入memory-stats库后，编译时出现以下错误：

Compiling memory-stats v1.1.0
error[E0432]: unresolved imports `libc::mach_task_basic_info_data_t`, `libc::mach_task_self`, `libc::task_info`, `libc::task_info_t`, `libc::MACH_TASK_BASIC_INFO`, `libc::MACH_TASK_BASIC_INFO_COUNT`
 --> /Users/meij/.cargo/registry/src/index.crates.io-6f17d22bba15001f/memory-stats-1.1.0/src/darwin.rs:4:29
  |
4 |     mach_msg_type_number_t, mach_task_basic_info_data_t, mach_task_self, task_info, task_info_t, K...
  |                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^  ^^^^^^^^^^^^^^  ^^^^^^^^^  ^^^^^^^^^^^ no `task_info_t` in the root
  |                             |                            |               |
  |                             |                            |               no `task_info` in the root
  |                             |                            no `mach_task_self` in the root
  |                             no `mach_task_basic_info_data_t` in the root
5 |     MACH_TASK_BASIC_INFO, MACH_TASK_BASIC_INFO_COUNT,
  |     ^^^^^^^^^^^^^^^^^^^^  ^^^^^^^^^^^^^^^^^^^^^^^^^^ no `MACH_TASK_BASIC_INFO_COUNT` in the root
  |     |
  |     no `MACH_TASK_BASIC_INFO` in the root
  |
help: a similar name exists in the module
  |
4 |     mach_msg_type_number_t, mach_timebase_info_data_t, mach_task_self, task_info, task_info_t, KERN_SUCCESS,
  |                             ~~~~~~~~~~~~~~~~~~~~~~~~~
help: a similar name exists in the module
  |
4 |     mach_msg_type_number_t, mach_task_basic_info_data_t, mach_host_self, task_info, task_info_t, KERN_SUCCESS,
  |                                                          ~~~~~~~~~~~~~~

For more information about this error, try `rustc --explain E0432`.
error: could not compile `memory-stats` (lib) due to previous error

本地环境版本：

• rustc 1.72.1 (d5c2e9c34 2023-09-13)
• cargo 1.72.1 (103a7ff2e 2023-08-15)

原因分析 #

这个错误是因为memory-stats v1.1.0的macOS平台实现依赖的libc版本较旧，而新版libc中macOS的mach相关API命名或结构发生了变化，导致库无法找到对应的导入项。

解决方法 #

方案1：指定兼容的libc版本 #

在Cargo.toml中明确指定与memory-stats v1.1.0兼容的libc版本：

[dependencies]
memory-stats = "1.1.0"
libc = "0.2.146"

然后执行以下命令重新编译：

cargo clean && cargo build

方案2：改用更活跃的跨平台库 #

如果方案1无效，推荐使用sysinfo库，它跨平台支持，API更友好，能获取内存、CPU等多种系统信息：

1. 在Cargo.toml中添加依赖：

[dependencies]
sysinfo = "0.29"

2. 示例代码：

use sysinfo::{System, SystemExt};

fn main() {
    let mut sys = System::new_all();
    sys.refresh_memory();
    println!("当前程序内存使用: {} KB", sys.used_memory());
    println!("系统总内存: {} KB", sys.total_memory());
}

方案3：手动实现macOS内存监控 #

如果不想依赖第三方库，可以直接调用mach API实现内存监控，示例代码如下：

1. 先在Cargo.toml中添加libc依赖：

[dependencies]
libc = "0.2"

2. 实现代码：

#[cfg(target_os = "macos")]
fn get_memory_usage() -> Option<u64> {
    use libc::{c_int, c_void, mach_msg_type_number_t};

    const MACH_TASK_BASIC_INFO: c_int = 4;
    const MACH_TASK_BASIC_INFO_COUNT: mach_msg_type_number_t = 6;
    type MachTaskBasicInfo = [u64; MACH_TASK_BASIC_INFO_COUNT as usize];

    unsafe {
        let mut info = MachTaskBasicInfo::default();
        let mut count = MACH_TASK_BASIC_INFO_COUNT;
        let task = libc::mach_task_self();

        let ret = libc::task_info(
            task,
            MACH_TASK_BASIC_INFO,
            &mut info as *mut _ as *mut c_void,
            &mut count,
        );

        if ret == libc::KERN_SUCCESS {
            Some(info[0]) // 返回常驻内存大小（字节）
        } else {
            None
        }
    }
}

#[cfg(target_os = "macos")]
fn main() {
    match get_memory_usage() {
        Some(mem_bytes) => println!("当前内存使用: {} MB", mem_bytes / 1024 / 1024),
        None => println!("无法获取内存使用信息"),
    }
}

内容的提问来源于stack exchange，提问作者May