Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

SIGSEGV when using non-trivial thread_local! with -C target-cpu=haswell on macOS #46355

Closed
swolchok opened this issue Nov 29, 2017 · 5 comments · Fixed by #51828
Closed

SIGSEGV when using non-trivial thread_local! with -C target-cpu=haswell on macOS #46355

swolchok opened this issue Nov 29, 2017 · 5 comments · Fixed by #51828
Labels
A-thread-locals Area: Thread local storage (TLS) C-bug Category: This is a bug. I-crash Issue: The compiler crashes (SIGSEGV, SIGABRT, etc). Use I-ICE instead when the compiler panics. O-macos Operating system: macOS O-x86_64 Target: x86-64 processors (like x86_64-*) P-medium Medium priority regression-from-stable-to-stable Performance or correctness regression from one stable version to another.

Comments

@swolchok
Copy link
Contributor

swolchok commented Nov 29, 2017 •
edited
Loading

Here's a code sample, narrowed down a bit from my program:

extern crate fnv;
extern crate rayon;

use fnv::FnvHashMap;
use rayon::prelude::*;
use std::cell::RefCell;
use std::sync::Arc;

#[derive(Clone)]
struct Scratchpad {
    vals: FnvHashMap<u32, u32>,
}

fn main() {
    let per_thread_state = Scratchpad { vals: Default::default() };
    let per_thread_state = Arc::new(per_thread_state);

    thread_local! {
        static STATE : RefCell<Option<Scratchpad>> = RefCell::new(None);
    }

    {
        let pts = per_thread_state.clone();
        let configuration = rayon::Configuration::new().start_handler(move |_| {
            STATE.with(|current_value| {
                assert!(current_value.borrow().is_none());
                *current_value.borrow_mut() = Some((*pts).clone());
            });
        });
        rayon::initialize(configuration);
    }
    let sum: u32 = vec![0; 100].par_iter().map(|&item| item).sum();
    println!("Sum is {}", sum);
}

When built with RUSTFLAGS="-C target-cpu=native" cargo build --release and run, it gets SIGSEGV on my Mac. (I poked around a little bit with lldb and it looks like tlv_get_addr is returning something we can't dereference. I couldn't investigate thoroughly because it wasn't showing source line information for whatever reason.) Without the RUSTFLAGS, it doesn't segfault.

Here's an lldb session with some basic information about the crash:

rust-lldb target/release/threadlocalrepro
(lldb) command source -s 0 '/tmp/rust-lldb-commands.vC8eMe'
Executing commands in '/tmp/rust-lldb-commands.vC8eMe'.
(lldb) command script import "/usr/local/Cellar/rust/1.22.1/lib/rustlib/etc/lldb_rust_formatters.py"
(lldb) type summary add --no-value --python-function lldb_rust_formatters.print_val -x ".*" --category Rust
(lldb) type category enable Rust
(lldb) target create "target/release/threadlocalrepro"
Current executable set to 'target/release/threadlocalrepro' (x86_64).
(lldb) run
Process 53759 launched: '/Users/swolchok/threadlocalrepro/target/release/threadlocalrepro' (x86_64)
Process 53759 stopped
* thread #2, stop reason = EXC_BAD_ACCESS (code=EXC_I386_GPFLT)
    frame #0: 0x00000001000042ae threadlocalrepro`threadlocalrepro::main::_$u7b$$u7b$closure$u7d$$u7d$::h3e975efef0e15fc4 + 302
threadlocalrepro`threadlocalrepro::main::_$u7b$$u7b$closure$u7d$$u7d$::h3e975efef0e15fc4:
->  0x1000042ae <+302>: vmovdqa (%rax), %ymm0
    0x1000042b2 <+306>: movl   $0x1, %ecx
    0x1000042b7 <+311>: vmovq  %rcx, %xmm1
    0x1000042bc <+316>: vmovdqa %ymm1, (%rax)
Target 0: (threadlocalrepro) stopped.
(lldb) register read
General Purpose Registers:
       rax = 0x0000000100202890
       rbx = 0x0000000100202890
       rcx = 0x0000000000000003
       rdx = 0x0000000100202890
       rdi = 0x000000010005db08  threadlocalrepro`threadlocalrepro::main::STATE::__getit::__KEY::h3fa22d7abc868d4a
       rsi = 0x0000000100200000
       rbp = 0x0000700004e80c50
       rsp = 0x0000700004e80bf0
        r8 = 0x0000000000000004
        r9 = 0x0000000000000000
       r10 = 0x0000000000000001
       r11 = 0x0000000000000005
       r12 = 0x0000000100624000
       r13 = 0x0000000100624090
       r14 = 0x0000000100617008
       r15 = 0x0000700004e80d08
       rip = 0x00000001000042ae  threadlocalrepro`threadlocalrepro::main::_$u7b$$u7b$closure$u7d$$u7d$::h3e975efef0e15fc4 + 302
    rflags = 0x0000000000010206
        cs = 0x000000000000002b
        fs = 0x0000000000000000
        gs = 0x0000000000000000

(lldb) bt
* thread #2, stop reason = EXC_BAD_ACCESS (code=EXC_I386_GPFLT)
  * frame #0: 0x00000001000042ae threadlocalrepro`threadlocalrepro::main::_$u7b$$u7b$closure$u7d$$u7d$::h3e975efef0e15fc4 + 302
    frame #1: 0x00000001000221fd threadlocalrepro`panic_unwind::__rust_maybe_catch_panic at lib.rs:99 [opt]
    frame #2: 0x000000010000564b threadlocalrepro`std::sys_common::backtrace::__rust_begin_short_backtrace::h25c1221ba9601a19 + 475
    frame #3: 0x0000000100005e6d threadlocalrepro`std::panicking::try::do_call::habed391336ff94ed + 29
    frame #4: 0x00000001000221fd threadlocalrepro`panic_unwind::__rust_maybe_catch_panic at lib.rs:99 [opt]
    frame #5: 0x000000010000750b threadlocalrepro`_$LT$F$u20$as$u20$alloc..boxed..FnBox$LT$A$GT$$GT$::call_box::hc3095b26e437282e + 171
    frame #6: 0x000000010001a97c threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] alloc::boxed::{{impl}}::call_once<(),()> at boxed.rs:736 [opt]
    frame #7: 0x000000010001a979 threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] std::sys_common::thread::start_thread at thread.rs:24 [opt]
    frame #8: 0x000000010001a8fe threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start at thread.rs:90 [opt]
    frame #9: 0x00007fffe2c9993b libsystem_pthread.dylib`_pthread_body + 180
    frame #10: 0x00007fffe2c99887 libsystem_pthread.dylib`_pthread_start + 286
    frame #11: 0x00007fffe2c9908d libsystem_pthread.dylib`thread_start + 13
(lldb) bt all
  thread #1, queue = 'com.apple.main-thread'
    frame #0: 0x00007fffe2baebf2 libsystem_kernel.dylib`__psynch_cvwait + 10
    frame #1: 0x00007fffe2c9a7fa libsystem_pthread.dylib`_pthread_cond_wait + 712
    frame #2: 0x0000000100007993 threadlocalrepro`rayon_core::latch::LockLatch::wait::h1a0d2eb553df7e2a + 163
    frame #3: 0x000000010000a1cc threadlocalrepro`rayon_core::initialize::h33e8eefa046130dd + 348
    frame #4: 0x0000000100003e2b threadlocalrepro`threadlocalrepro::main::h5914660853824a52 + 379
    frame #5: 0x00000001000221fd threadlocalrepro`panic_unwind::__rust_maybe_catch_panic at lib.rs:99 [opt]
    frame #6: 0x000000010001bc09 threadlocalrepro`std::rt::lang_start [inlined] std::panicking::try<(),closure> at panicking.rs:459 [opt]
    frame #7: 0x000000010001bbd6 threadlocalrepro`std::rt::lang_start [inlined] std::panic::catch_unwind<closure,()> at panic.rs:361 [opt]
    frame #8: 0x000000010001bbd6 threadlocalrepro`std::rt::lang_start at rt.rs:59 [opt]
    frame #9: 0x00007fffe2a80235 libdyld.dylib`start + 1
    frame #10: 0x00007fffe2a80235 libdyld.dylib`start + 1
* thread #2, stop reason = EXC_BAD_ACCESS (code=EXC_I386_GPFLT)
  * frame #0: 0x00000001000042ae threadlocalrepro`threadlocalrepro::main::_$u7b$$u7b$closure$u7d$$u7d$::h3e975efef0e15fc4 + 302
    frame #1: 0x00000001000221fd threadlocalrepro`panic_unwind::__rust_maybe_catch_panic at lib.rs:99 [opt]
    frame #2: 0x000000010000564b threadlocalrepro`std::sys_common::backtrace::__rust_begin_short_backtrace::h25c1221ba9601a19 + 475
    frame #3: 0x0000000100005e6d threadlocalrepro`std::panicking::try::do_call::habed391336ff94ed + 29
    frame #4: 0x00000001000221fd threadlocalrepro`panic_unwind::__rust_maybe_catch_panic at lib.rs:99 [opt]
    frame #5: 0x000000010000750b threadlocalrepro`_$LT$F$u20$as$u20$alloc..boxed..FnBox$LT$A$GT$$GT$::call_box::hc3095b26e437282e + 171
    frame #6: 0x000000010001a97c threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] alloc::boxed::{{impl}}::call_once<(),()> at boxed.rs:736 [opt]
    frame #7: 0x000000010001a979 threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] std::sys_common::thread::start_thread at thread.rs:24 [opt]
    frame #8: 0x000000010001a8fe threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start at thread.rs:90 [opt]
    frame #9: 0x00007fffe2c9993b libsystem_pthread.dylib`_pthread_body + 180
    frame #10: 0x00007fffe2c99887 libsystem_pthread.dylib`_pthread_start + 286
    frame #11: 0x00007fffe2c9908d libsystem_pthread.dylib`thread_start + 13
  thread #3
    frame #0: 0x00007fffe2baebf2 libsystem_kernel.dylib`__psynch_cvwait + 10
    frame #1: 0x00007fffe2c9a7fa libsystem_pthread.dylib`_pthread_cond_wait + 712
    frame #2: 0x0000000100009df5 threadlocalrepro`rayon_core::sleep::Sleep::sleep::he476e9c9ebf0bba4 + 277
    frame #3: 0x0000000100009986 threadlocalrepro`rayon_core::registry::WorkerThread::wait_until::h438f955d0a4ae18c + 630
    frame #4: 0x0000000100005701 threadlocalrepro`std::sys_common::backtrace::__rust_begin_short_backtrace::h25c1221ba9601a19 + 657
    frame #5: 0x0000000100005e6d threadlocalrepro`std::panicking::try::do_call::habed391336ff94ed + 29
    frame #6: 0x00000001000221fd threadlocalrepro`panic_unwind::__rust_maybe_catch_panic at lib.rs:99 [opt]
    frame #7: 0x000000010000750b threadlocalrepro`_$LT$F$u20$as$u20$alloc..boxed..FnBox$LT$A$GT$$GT$::call_box::hc3095b26e437282e + 171
    frame #8: 0x000000010001a97c threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] alloc::boxed::{{impl}}::call_once<(),()> at boxed.rs:736 [opt]
    frame #9: 0x000000010001a979 threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] std::sys_common::thread::start_thread at thread.rs:24 [opt]
    frame #10: 0x000000010001a8fe threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start at thread.rs:90 [opt]
    frame #11: 0x00007fffe2c9993b libsystem_pthread.dylib`_pthread_body + 180
    frame #12: 0x00007fffe2c99887 libsystem_pthread.dylib`_pthread_start + 286
    frame #13: 0x00007fffe2c9908d libsystem_pthread.dylib`thread_start + 13
  thread #4
    frame #0: 0x00007fffe2bb0246 libsystem_kernel.dylib`read + 10
    frame #1: 0x0000000100011637 threadlocalrepro`std::fs::{{impl}}::read [inlined] std::sys::imp::fd::{{impl}}::read at fd.rs:58 [opt]
    frame #2: 0x000000010001161a threadlocalrepro`std::fs::{{impl}}::read [inlined] std::sys::imp::fs::{{impl}}::read at fs.rs:496 [opt]
    frame #3: 0x000000010001161a threadlocalrepro`std::fs::{{impl}}::read at fs.rs:449 [opt]
    frame #4: 0x000000010000af71 threadlocalrepro`_$LT$rand..os..imp..OsRng$u20$as$u20$rand..Rng$GT$::fill_bytes::h8532113fd6db4404 + 65
    frame #5: 0x000000010000b1b1 threadlocalrepro`rand::StdRng::new::hcd2107c72b382959 + 337
    frame #6: 0x000000010000b9c2 threadlocalrepro`rand::thread_rng::THREAD_RNG_KEY::__init::he04c67a54866e26e + 34
    frame #7: 0x000000010000b91d threadlocalrepro`rand::thread_rng::h5fb42a8cb8008a3e + 93
    frame #8: 0x000000010000b26a threadlocalrepro`rand::weak_rng::h1e1cad0723c27821 + 26
    frame #9: 0x00000001000054ae threadlocalrepro`std::sys_common::backtrace::__rust_begin_short_backtrace::h25c1221ba9601a19 + 62
    frame #10: 0x0000000100005e6d threadlocalrepro`std::panicking::try::do_call::habed391336ff94ed + 29
    frame #11: 0x00000001000221fd threadlocalrepro`panic_unwind::__rust_maybe_catch_panic at lib.rs:99 [opt]
    frame #12: 0x000000010000750b threadlocalrepro`_$LT$F$u20$as$u20$alloc..boxed..FnBox$LT$A$GT$$GT$::call_box::hc3095b26e437282e + 171
    frame #13: 0x000000010001a97c threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] alloc::boxed::{{impl}}::call_once<(),()> at boxed.rs:736 [opt]
    frame #14: 0x000000010001a979 threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] std::sys_common::thread::start_thread at thread.rs:24 [opt]
    frame #15: 0x000000010001a8fe threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start at thread.rs:90 [opt]
    frame #16: 0x00007fffe2c9993b libsystem_pthread.dylib`_pthread_body + 180
    frame #17: 0x00007fffe2c99887 libsystem_pthread.dylib`_pthread_start + 286
    frame #18: 0x00007fffe2c9908d libsystem_pthread.dylib`thread_start + 13
  thread #5
    frame #0: 0x00007fffe2bb0246 libsystem_kernel.dylib`read + 10
    frame #1: 0x0000000100011637 threadlocalrepro`std::fs::{{impl}}::read [inlined] std::sys::imp::fd::{{impl}}::read at fd.rs:58 [opt]
    frame #2: 0x000000010001161a threadlocalrepro`std::fs::{{impl}}::read [inlined] std::sys::imp::fs::{{impl}}::read at fs.rs:496 [opt]
    frame #3: 0x000000010001161a threadlocalrepro`std::fs::{{impl}}::read at fs.rs:449 [opt]
    frame #4: 0x000000010000af71 threadlocalrepro`_$LT$rand..os..imp..OsRng$u20$as$u20$rand..Rng$GT$::fill_bytes::h8532113fd6db4404 + 65
    frame #5: 0x000000010000b1b1 threadlocalrepro`rand::StdRng::new::hcd2107c72b382959 + 337
    frame #6: 0x000000010000b9c2 threadlocalrepro`rand::thread_rng::THREAD_RNG_KEY::__init::he04c67a54866e26e + 34
    frame #7: 0x000000010000b91d threadlocalrepro`rand::thread_rng::h5fb42a8cb8008a3e + 93
    frame #8: 0x000000010000b26a threadlocalrepro`rand::weak_rng::h1e1cad0723c27821 + 26
    frame #9: 0x00000001000054ae threadlocalrepro`std::sys_common::backtrace::__rust_begin_short_backtrace::h25c1221ba9601a19 + 62
    frame #10: 0x0000000100005e6d threadlocalrepro`std::panicking::try::do_call::habed391336ff94ed + 29
    frame #11: 0x00000001000221fd threadlocalrepro`panic_unwind::__rust_maybe_catch_panic at lib.rs:99 [opt]
    frame #12: 0x000000010000750b threadlocalrepro`_$LT$F$u20$as$u20$alloc..boxed..FnBox$LT$A$GT$$GT$::call_box::hc3095b26e437282e + 171
    frame #13: 0x000000010001a97c threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] alloc::boxed::{{impl}}::call_once<(),()> at boxed.rs:736 [opt]
    frame #14: 0x000000010001a979 threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] std::sys_common::thread::start_thread at thread.rs:24 [opt]
    frame #15: 0x000000010001a8fe threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start at thread.rs:90 [opt]
    frame #16: 0x00007fffe2c9993b libsystem_pthread.dylib`_pthread_body + 180
    frame #17: 0x00007fffe2c99887 libsystem_pthread.dylib`_pthread_start + 286
    frame #18: 0x00007fffe2c9908d libsystem_pthread.dylib`thread_start + 13
  thread #6
    frame #0: 0x00007fffe2baebf2 libsystem_kernel.dylib`__psynch_cvwait + 10
    frame #1: 0x00007fffe2c9a7fa libsystem_pthread.dylib`_pthread_cond_wait + 712
    frame #2: 0x0000000100009df5 threadlocalrepro`rayon_core::sleep::Sleep::sleep::he476e9c9ebf0bba4 + 277
    frame #3: 0x0000000100009986 threadlocalrepro`rayon_core::registry::WorkerThread::wait_until::h438f955d0a4ae18c + 630
    frame #4: 0x0000000100005701 threadlocalrepro`std::sys_common::backtrace::__rust_begin_short_backtrace::h25c1221ba9601a19 + 657
    frame #5: 0x0000000100005e6d threadlocalrepro`std::panicking::try::do_call::habed391336ff94ed + 29
    frame #6: 0x00000001000221fd threadlocalrepro`panic_unwind::__rust_maybe_catch_panic at lib.rs:99 [opt]
    frame #7: 0x000000010000750b threadlocalrepro`_$LT$F$u20$as$u20$alloc..boxed..FnBox$LT$A$GT$$GT$::call_box::hc3095b26e437282e + 171
    frame #8: 0x000000010001a97c threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] alloc::boxed::{{impl}}::call_once<(),()> at boxed.rs:736 [opt]
    frame #9: 0x000000010001a979 threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] std::sys_common::thread::start_thread at thread.rs:24 [opt]
    frame #10: 0x000000010001a8fe threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start at thread.rs:90 [opt]
    frame #11: 0x00007fffe2c9993b libsystem_pthread.dylib`_pthread_body + 180
    frame #12: 0x00007fffe2c99887 libsystem_pthread.dylib`_pthread_start + 286
    frame #13: 0x00007fffe2c9908d libsystem_pthread.dylib`thread_start + 13
  thread #7
    frame #0: 0x00007fffe2baebf2 libsystem_kernel.dylib`__psynch_cvwait + 10
    frame #1: 0x00007fffe2c9a7fa libsystem_pthread.dylib`_pthread_cond_wait + 712
    frame #2: 0x0000000100009df5 threadlocalrepro`rayon_core::sleep::Sleep::sleep::he476e9c9ebf0bba4 + 277
    frame #3: 0x0000000100009986 threadlocalrepro`rayon_core::registry::WorkerThread::wait_until::h438f955d0a4ae18c + 630
    frame #4: 0x0000000100005701 threadlocalrepro`std::sys_common::backtrace::__rust_begin_short_backtrace::h25c1221ba9601a19 + 657
    frame #5: 0x0000000100005e6d threadlocalrepro`std::panicking::try::do_call::habed391336ff94ed + 29
    frame #6: 0x00000001000221fd threadlocalrepro`panic_unwind::__rust_maybe_catch_panic at lib.rs:99 [opt]
    frame #7: 0x000000010000750b threadlocalrepro`_$LT$F$u20$as$u20$alloc..boxed..FnBox$LT$A$GT$$GT$::call_box::hc3095b26e437282e + 171
    frame #8: 0x000000010001a97c threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] alloc::boxed::{{impl}}::call_once<(),()> at boxed.rs:736 [opt]
    frame #9: 0x000000010001a979 threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start [inlined] std::sys_common::thread::start_thread at thread.rs:24 [opt]
    frame #10: 0x000000010001a8fe threadlocalrepro`std::sys::imp::thread::{{impl}}::new::thread_start at thread.rs:90 [opt]
    frame #11: 0x00007fffe2c9993b libsystem_pthread.dylib`_pthread_body + 180
    frame #12: 0x00007fffe2c99887 libsystem_pthread.dylib`_pthread_start + 286
    frame #13: 0x00007fffe2c9908d libsystem_pthread.dylib`thread_start + 13
(lldb)

rustc --version --verbose
rustc 1.22.1
binary: rustc
commit-hash: unknown
commit-date: unknown
host: x86_64-apple-darwin
release: 1.22.1
LLVM version: 4.0

I'm on macOS 10.12.6. target-cpu=native on my machine seems to mean haswell, according to rustc --print target-cpus.
@jethrogb
Copy link
Contributor

Seems to be macOS-specific, can't reproduce this on Linux.

@tcr
Copy link

tcr commented Nov 29, 2017 •
edited
Loading

I can repro on my Mac. Early 2015 Macbook Pro, running macOS Sierra.

$ sysctl -n machdep.cpu.brand_string
Intel(R) Core(TM) i7-5557U CPU @ 3.10GHz

It did not crash on my 1.18.0 install; first starts crashing on 1.20.0 installed via rustup:

rustc --version --verbose
rustc 1.20.0 (f3d6973f4 2017-08-27)
binary: rustc
commit-hash: f3d6973f41a7d1fb83029c9c0ceaf0f5d4fd7208
commit-date: 2017-08-27
host: x86_64-apple-darwin
release: 1.20.0
LLVM version: 4.0

@kennytm kennytm added A-thread-locals Area: Thread local storage (TLS) C-bug Category: This is a bug. I-crash Issue: The compiler crashes (SIGSEGV, SIGABRT, etc). Use I-ICE instead when the compiler panics. O-macos Operating system: macOS O-x86_64 Target: x86-64 processors (like x86_64-*) regression-from-stable-to-stable Performance or correctness regression from one stable version to another. labels Nov 29, 2017
@alexcrichton
Copy link
Member

This may be #44056 which was diagnosed as an LLVM bug.

@nikomatsakis
Copy link
Contributor

triage: P-medium

Seems likely to be an LLVM bug. If more data or dups turn up, we can re-evaluate.

@rust-highfive rust-highfive added the P-medium Medium priority label Jan 11, 2018
@josephDunne
Copy link
Contributor

josephDunne commented Feb 14, 2018 •
edited
Loading

I am experiencing something similar to this when compiling native on OSX.

thread #2, stop reason = EXC_BAD_ACCESS (code=EXC_I386_GPFLT)
    frame #0: 0x0000000100196852 nap`std::sys_common::thread_info::set::h958c88701b41d520 + 114
nap`std::sys_common::thread_info::set::h958c88701b41d520:
->  0x100196852 <+114>: vmovaps (%rax), %ymm0
    0x100196856 <+118>: vmovaps 0x60302(%rip), %ymm1
    0x10019685e <+126>: vmovaps %ymm1, (%rax)
    0x100196862 <+130>: movq   0x20(%rax), %rax

OSX

sysctl -n machdep.cpu.brand_string
Intel(R) Xeon(R) CPU E5-1650 v2 @ 3.50GHz

rustc --version --verbose
rustc 1.25.0-nightly (4d2d3fc5d 2018-02-13)
binary: rustc
commit-hash: 4d2d3fc5dadf894a8ad709a5860a549f2c0b1032
commit-date: 2018-02-13
host: x86_64-apple-darwin
release: 1.25.0-nightly
LLVM version: 6.0

[Edit] problem does not exists in nightly v 1.24. So perhaps a regression in LLVM 6.0?

rustc 1.24.0-nightly (9389e23a8 2017-12-31)
binary: rustc
commit-hash: 9389e23a8a754097e233c7bf3ea1bb404ccf1075
commit-date: 2017-12-31
host: x86_64-apple-darwin
release: 1.24.0-nightly
LLVM version: 4.0

@kennytm kennytm mentioned this issue Jun 28, 2018
Merged
bors added a commit that referenced this issue Jun 28, 2018
@bors
Auto merge of #51828 - kennytm:no-simd-swap-for-mac, r=<try>
29f834a 
[DO NOT MERGE] Do not allow LLVM to increase a TLS's alignment on macOS.

This addresses the various TLS segfault on macOS 10.10.

Fix #51794.
Fix #51758.
Fix #50867.
Fix #48866.
Fix #46355.
Fix #44056.
Mark-Simulacrum added a commit to Mark-Simulacrum/rust that referenced this issue Jun 30, 2018
@Mark-Simulacrum
Rollup merge of rust-lang#51828 - kennytm:no-simd-swap-for-mac, r=ale…
8c30769 
…xcrichton

Do not allow LLVM to increase a TLS's alignment on macOS.

This addresses the various TLS segfault on macOS 10.10.

Fix rust-lang#51794.
Fix rust-lang#51758.
Fix rust-lang#50867.
Fix rust-lang#48866.
Fix rust-lang#46355.
Fix rust-lang#44056.
bors added a commit that referenced this issue Jun 30, 2018
@bors
Auto merge of #51828 - kennytm:no-simd-swap-for-mac, r=alexcrichton
96b4733 
Do not allow LLVM to increase a TLS's alignment on macOS.

This addresses the various TLS segfault on macOS 10.10.

Fix #51794.
Fix #51758.
Fix #50867.
Fix #48866.
Fix #46355.
Fix #44056.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
A-thread-locals Area: Thread local storage (TLS) C-bug Category: This is a bug. I-crash Issue: The compiler crashes (SIGSEGV, SIGABRT, etc). Use I-ICE instead when the compiler panics. O-macos Operating system: macOS O-x86_64 Target: x86-64 processors (like x86_64-*) P-medium Medium priority regression-from-stable-to-stable Performance or correctness regression from one stable version to another.
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Do not allow LLVM to increase a TLS's alignment on macOS. kennytm/rust
8 participants
@swolchok @alexcrichton @tcr @kennytm @nikomatsakis @jethrogb @josephDunne @rust-highfive