rust-lang / rust-lang/backtrace-rs

Backtrace info missing when file descriptors exhausted

Open
#686 3 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Rust
Stars
628
Forks
291
PR merge metrics
No merged PRs in 30d

Description

When a program uses all available file descriptors and then panics, the backtrace contains no symbolic information, line numbers, and such like.

Example output

Consuming all available file descriptors with RUST_BACKTRACE=full
thread 'main' panicked at src/main.rs:19:13:
ran out of file descriptors
stack backtrace:
0:     0x61c61a98b56a - <unknown>
1:     0x61c61a9a7f33 - <unknown>
2:     0x61c61a988fd3 - <unknown>
3:     0x61c61a98b3b2 - <unknown>
4:     0x61c61a98c35c - <unknown>
5:     0x61c61a98c1a2 - <unknown>
6:     0x61c61a98c937 - <unknown>
7:     0x61c61a98c796 - <unknown>
8:     0x61c61a98ba49 - <unknown>
9:     0x61c61a98c45c - <unknown>
10:     0x61c61a959420 - <unknown>
11:     0x61c61a95b2d9 - <unknown>
12:     0x61c61a959ecb - <unknown>
13:     0x61c61a95ab2e - <unknown>
14:     0x61c61a95b961 - <unknown>
15:     0x61c61a98636e - <unknown>
16:     0x61c61a95b93a - <unknown>
17:     0x61c61a95b33e - <unknown>
18:     0x76526c753e08 - <unknown>
19:     0x76526c753ecc - <unknown>
20:     0x61c61a959c95 - <unknown>
21:                0x0 - <unknown>

Desired behaviour

Ideally, full backtrace information should be generated.

Alternatively, it would be helpful to include some clue in the backtrace as to why the information is missing.

Reproducing the issue

The following program reproduces the behaviour:

use libc;
use tempdir::TempDir;
use std::{ffi::CString, io::Write};
use std::{env, fs::File};

fn main() {
    env::set_var("RUST_BACKTRACE", "full");
    println!("Consuming all available file descriptors with RUST_BACKTRACE=full");
    let tmp_dir = TempDir::new("example").unwrap();
    let path : String = tmp_dir.path().join("test").to_owned().into_os_string().into_string().unwrap();
    let mut file = File::create_new(&path).unwrap();
    file.write_all(b"contents").unwrap();

    let mut fds = vec!();
    loop {
        let file_path = CString::new(path.clone()).unwrap();
        let fd = unsafe { libc::open(file_path.as_ptr(), libc::O_RDONLY) };
        if fd == -1 {
            panic!("ran out of file descriptors");
        }
        fds.push(fd);
    }
}

Reproduce the issue by cloning this repository and issuing cargo run from inside the cloned directory.

Environment

The issue occurred in the following environments.

Linux
  • Rust 1.83.0
  • Arch linux 6.12.4
  • x86_64
macOS
  • Rust 1.83.0
  • macOS Sonoma 14.7.2
  • Apple M1

Background

The issue was encountered while debugging a failing test. See this zuplip conversation for more details.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by running cargo run in the linked exhaustfd repository with the reproducer shown in the issue, then trace how backtrace-rs resolves symbols when file descriptors are exhausted. The issue is done when the full backtrace retains symbolic information and line numbers, or clearly explains why they are unavailable.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
operating-systems
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.