Latest cargo nightly instantly segfaults on FreeBSD in Cirrus CI

[asomers]

Problem

The latest nightly cargo instantly segfaults when run on FreeBSD in Cirrus CI. It doesn't even need to attempt a build. Simply cargo -h will segfault. I haven't managed to reproduce this in a local VM, so I don't know what's special about Cirrus CI. But the failures only began to happen within the last 24 hours.

Steps

In a FreeBSD 14.0 or 14.1 (other versions are untested) instance on Cirrus CI, install Rust and invoke Cargo like this:

$ fetch https://sh.rustup.rs -o rustup.sh
$ sh rustup.sh -y --profile=minimal --default-toolchain nightly-x86_64-unknown-freebsd
$ . $HOME/.cargo/env
$ cargo -h
Segmentation fault (core dumped)

Possible Solution(s)

No response

Notes

Here is the stack backtrace from gdb.

Starting program: /tmp/.cargo/bin/cargo -h
process 13657 is executing new program: /tmp/.rustup/toolchains/nightly-x86_64-unknown-freebsd/bin/cargo

Program received signal SIGSEGV, Segmentation fault.
Address not mapped to object.
libunwind::CFI_Parser<libunwind::LocalAddressSpace>::parseCIE (addressSpace=..., cie=18446744073502207747, 
    cieInfo=0x7fffffffa480) at /usr/src/contrib/llvm-project/libunwind/src/DwarfParser.hpp:332
332       pint_t cieLength = (pint_t)addressSpace.get32(p);
(gdb) bt
#0  libunwind::CFI_Parser<libunwind::LocalAddressSpace>::parseCIE (addressSpace=..., cie=18446744073502207747, 
    cieInfo=0x7fffffffa480) at /usr/src/contrib/llvm-project/libunwind/src/DwarfParser.hpp:332
#1  0x0000000802b9ae8c in libunwind::CFI_Parser<libunwind::LocalAddressSpace>::findFDE (addressSpace=..., 
    pc=pc@entry=19247820, ehSectionStart=42259904, sectionLength=<optimized out>, fdeHint=fdeHint@entry=0, 
    fdeInfo=fdeInfo@entry=0x7fffffffa4b8, cieInfo=0x7fffffffa480)
    at /usr/src/contrib/llvm-project/libunwind/src/DwarfParser.hpp:264
#2  0x0000000802b9ab22 in libunwind::UnwindCursor<libunwind::LocalAddressSpace, libunwind::Registers_x86_64>::getInfoFromDwarfSection (this=this@entry=0x7fffffffab00, pc=pc@entry=19247820, sects=..., 
    fdeSectionOffsetHint=fdeSectionOffsetHint@entry=0)
    at /usr/src/contrib/llvm-project/libunwind/src/UnwindCursor.hpp:1689
#3  0x0000000802b97fe0 in libunwind::UnwindCursor<libunwind::LocalAddressSpace, libunwind::Registers_x86_64>::setInfoBasedOnIPRegister (this=0x7fffffffab00, isReturnAddress=true)
    at /usr/src/contrib/llvm-project/libunwind/src/UnwindCursor.hpp:2634
#4  0x0000000802b97e32 in libunwind::UnwindCursor<libunwind::LocalAddressSpace, libunwind::Registers_x86_64>::step (
    this=0x7fffffffab00, stage2=<optimized out>) at /usr/src/contrib/llvm-project/libunwind/src/UnwindCursor.hpp:2946
#5  0x0000000802b95379 in _Unwind_Backtrace (
    callback=0x25ac480 <std::backtrace_rs::backtrace::libunwind::trace::trace_fn>, ref=0x7fffffffaca8)
    at /usr/src/contrib/llvm-project/libunwind/src/UnwindLevel1-gcc-ext.c:156
#6  0x0000000002592725 in std::backtrace_rs::backtrace::libunwind::trace ()
    at library/std/src/../../backtrace/src/backtrace/libunwind.rs:116
#7  std::backtrace_rs::backtrace::trace_unsynchronized<std::backtrace::{impl#4}::create::{closure_env#0}> ()
    at library/std/src/../../backtrace/src/backtrace/mod.rs:66
#8  std::backtrace::Backtrace::create () at library/std/src/backtrace.rs:331
#9  0x0000000002592590 in std::backtrace::Backtrace::capture () at library/std/src/backtrace.rs:296
#10 0x000000000123b2f8 in <anyhow::Error as core::convert::From<clap_builder::error::Error>>::from ()
#11 0x0000000001a30856 in <cargo::util::errors::CliError as core::convert::From<clap_builder::error::Error>>::from ()
#12 0x0000000001327f67 in cargo::cli::main ()
#13 0x00000000012ff725 in cargo::main ()
#14 0x00000000012e4463 in std::sys::backtrace::__rust_begin_short_backtrace::<fn(), ()> ()
#15 0x000000000132fb29 in std::rt::lang_start::<()>::{closure#0} ()
#16 0x0000000002591230 in core::ops::function::impls::{impl#2}::call_once<(), (dyn core::ops::function::Fn<(), Output=i32>

And here are some links to recent builds demonstrating the problem:

• https://cirrus-ci.com/task/4617134287880192
• https://cirrus-ci.com/task/5279838094753792

Version

# cargo --version --verbose
cargo 1.86.0-nightly (045bf21b3 2025-01-17)
release: 1.86.0-nightly
commit-hash: 045bf21b36a2e1f3ed85e38278d1c3cc4305e134
commit-date: 2025-01-17
host: x86_64-unknown-freebsd
libgit2: 1.9.0 (sys:0.20.0 vendored)
libcurl: 8.9.0-DEV (sys:0.4.74+curl-8.9.0 vendored ssl:OpenSSL/1.1.1w)
ssl: OpenSSL 1.1.1w  11 Sep 2023
os: FreeBSD 14.1-RELEASE [64-bit]

[weihanglo]

But the failures only began to happen within the last 24 hours.

Nightly cargo hasn't changed since 2025-01-17, so if it worked one day ago it should generally work now.
To bisect which nightly start failing, is it possible to configure CI to download a specific nightly version? Like rustup install nightly-2024-01-22.

[ehuss]

FreeBSD builds were switched from 12 to 13 in rust-lang/rust#132232 about a day ago.

cc @asomers FYI

[asomers]

Using rustup, I've bisected the problem down. It happens with nightly-2025-01-23 but not nightly-2025-01-22. And it only happens when RUST_BACKTRACE=1 or RUST_BACKTRACE=full. So it's probably a regression of rust-lang/rust#132185 .