Stream: git-wasmtime

Topic: wasmtime / issue #7997 libgcc_13: SIGABORT inside `Unwind...

Wasmtime GitHub notifications bot (Feb 26 2024 at 13:14):

Mr-Leshiy edited issue #7997:

We have faced a SIGABORT failure during executing simple benchmark which you can find here https://github.com/input-output-hk/hermes/blob/f4d20fc06f0b558be4805a106db32480bcef649d/hermes/bin/src/wasm/module.rs#L118.
According to the stack trace

Steps to Reproduce

You can fully reproduce it in already prepared environment and code but cloning our repo and specifically this branch libgcc-wasmtime-unwinding-bug. (repo: https://github.com/input-output-hk/hermes)

Before start you will need to install earthly tool.
To run it you will need to execute the following in the root of the repo

earthly -I ./hermes+alpine-3-19-fail

This will prepare an alpine:3.19 environment with rust:1.75 version and all our project configuration that we have an all source code and run a benchmarks.
Earthfile code https://github.com/input-output-hk/hermes/blob/libgcc-wasmtime-unwinding-bug/hermes/Earthfile#L18

alpine-3-19-fail:
    FROM rust:1.75-alpine3.19

    # Install necessary packages
    RUN apk add --no-cache \
            musl-dev \
            gdb

    COPY --dir .cargo .config crates bin .
    COPY Cargo.toml .
    COPY clippy.toml deny.toml rustfmt.toml .

    RUN mkdir /wasm
    COPY --dir ../wasm+wasi-src/wasi /wasm/wasi
    # Compiled WASM component for benchmarks
    COPY ../wasm/c+build/component.wasm /wasm/c/bench_component.wasm

    # Run benchmarks
    RUN cargo bench --features bench

After the failure it will run an interacting mode -i flag (similar to docker interactive mode).
So you will be able to collect a dump file inside this container.

ulimit -c unlimited && ./target/release/deps/module-<> --bench

Then inspect it and using gdb

gdb  ./target/release/deps/module-<> core -- --bench

Actual Results

SIGABORT

Versions and Environment

Wasmtime version or commit: 17.0.0

Operating system: alpine:3.19

Extra Info

This is totally works on alpine:3.18 version with the libgcc:12 version.
We have found that most probable it is an issue with libgcc because, there is a huge difference with the libgcc unwinding implementation between these two versions.

libgcc_12 - https://github.com/gcc-mirror/gcc/blob/8cbb2cade4c724760c868c9f493b169d6ec4168a/libgcc/unwind-dw2-fde.c#L201

libgcc_13 - https://github.com/gcc-mirror/gcc/blob/0e7bc3eaa36b81004b799124d2fe00137401a43b/libgcc/unwind-dw2-fde.c#L225

You can try it running already prepared earthly target https://github.com/input-output-hk/hermes/blob/libgcc-wasmtime-unwinding-bug/hermes/Earthfile#L38

earthly ./hermes+alpine-3-18-pass

Wasmtime GitHub notifications bot (Feb 26 2024 at 17:28):

alexcrichton commented on issue #7997:

Thanks for the report! Would you be able to upload the wasm file here directly? Also would you be able to share a copy/paste of the stack trace you're seeing? I can't reproduce locally with a "hello-world" style module loaded into alpine:3.19 so it seems that the issue may be related to the specific unwind table for the wasm file. I don't know much about earthly myself so I'm hopeful that the issue can be reduced relatively quickly to just a wasm file and some simple interactions with the wasmtime crate.

Wasmtime GitHub notifications bot (Feb 26 2024 at 17:55):

Mr-Leshiy edited issue #7997:

We have faced a SIGABORT failure during executing simple benchmark which you can find here https://github.com/input-output-hk/hermes/blob/f4d20fc06f0b558be4805a106db32480bcef649d/hermes/bin/src/wasm/module.rs#L118.
Stack trace:

#0  __restore_sigs (set=set@entry=0xffffc437ab50) at ./arch/aarch64/syscall_arch.h:48
#1  0x0000ffff9c2add5c in raise (sig=sig@entry=6) at src/signal/raise.c:11
#2  0x0000ffff9c27cbf4 in abort () at src/exit/abort.c:11
#3  0x0000ffff9c23e8fc in __deregister_frame_info_bases () from /usr/lib/libgcc_s.so.1
#4  0x0000ffff9c23e91c in __deregister_frame () from /usr/lib/libgcc_s.so.1
#5  0x0000aaaac2c0d0a8 in wasmtime_runtime::sys::unix::unwind::{impl#1}::drop (self=<optimized out>) at src/sys/unix/unwind.rs:87
#6  0x0000aaaac2ba2ba8 in core::ptr::drop_in_place<wasmtime_runtime::sys::unix::unwind::UnwindRegistration> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#7  core::ptr::drop_in_place<core::option::Option<wasmtime_runtime::sys::unix::unwind::UnwindRegistration>> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#8  0x0000aaaac2bb4d5c in core::mem::manually_drop::ManuallyDrop<core::option::Option<wasmtime_runtime::sys::unix::unwind::UnwindRegistration>>::drop<core::option::Option<wasmtime_runtime::sys::unix::unwind::UnwindRegistration>> (slot=0x0)
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/mem/manually_drop.rs:144
--Type <RET> for more, q to quit, c to continue without paging--
#9  wasmtime_jit::code_memory::{impl#0}::drop (self=0xffff9c216eb0) at src/code_memory.rs:42
#10 core::ptr::drop_in_place<wasmtime_jit::code_memory::CodeMemory> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#11 alloc::sync::Arc<wasmtime_jit::code_memory::CodeMemory, alloc::alloc::Global>::drop_slow<wasmtime_jit::code_memory::CodeMemory, alloc::alloc::Global> (self=<optimized out>) at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:1752
#12 0x0000aaaac27cfcb8 in alloc::sync::{impl#33}::drop<wasmtime_jit::code_memory::CodeMemory, alloc::alloc::Global> (
    self=0xffff9c213540) at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:2408
#13 core::ptr::drop_in_place<alloc::sync::Arc<wasmtime_jit::code_memory::CodeMemory, alloc::alloc::Global>> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#14 core::ptr::drop_in_place<wasmtime::code::CodeObject> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#15 alloc::sync::Arc<wasmtime::code::CodeObject, alloc::alloc::Global>::drop_slow<wasmtime::code::CodeObject, alloc::alloc::Global> (
    self=<optimized out>) at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:1752
--Type <RET> for more, q to quit, c to continue without paging--
#16 0x0000aaaac27d06a8 in alloc::sync::{impl#33}::drop<wasmtime::code::CodeObject, alloc::alloc::Global> (self=0x0)
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:2408
#17 core::ptr::drop_in_place<alloc::sync::Arc<wasmtime::code::CodeObject, alloc::alloc::Global>> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#18 core::ptr::drop_in_place<wasmtime::component::component::ComponentInner> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#19 alloc::sync::Arc<wasmtime::component::component::ComponentInner, alloc::alloc::Global>::drop_slow<wasmtime::component::component::ComponentInner, alloc::alloc::Global> (self=<optimized out>)
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:1752
#20 0x0000aaaac275d494 in alloc::sync::{impl#33}::drop<wasmtime::component::component::ComponentInner, alloc::alloc::Global> (
    self=0xffffc437ad40) at rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:2408
#21 core::ptr::drop_in_place<alloc::sync::Arc<wasmtime::component::component::ComponentInner, alloc::alloc::Global>> ()
    at rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
--Type <RET> for more, q to quit, c to continue without paging--
#22 core::ptr::drop_in_place<wasmtime::component::component::Component> ()
    at rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#23 core::ptr::drop_in_place<wasmtime::component::instance::InstancePre<hermes::state::HermesState>> ()
    at rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#24 0x0000aaaac275b390 in core::ptr::drop_in_place<hermes::wasm::module::Module> ()
    at rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#25 0x0000aaaac2761460 in hermes::wasm::module::bench::module_hermes_component_bench (b=0xffffc437ae30) at bin/src/wasm/module.rs:143

Steps to Reproduce

You can fully reproduce it in already prepared environment and code but cloning our repo and specifically this branch libgcc-wasmtime-unwinding-bug. (repo: https://github.com/input-output-hk/hermes)

Before start you will need to install earthly tool.
To run it you will need to execute the following in the root of the repo

earthly -I ./hermes+alpine-3-19-fail

This will prepare an alpine:3.19 environment with rust:1.75 version and all our project configuration that we have an all source code and run a benchmarks.
Earthfile code https://github.com/input-output-hk/hermes/blob/libgcc-wasmtime-unwinding-bug/hermes/Earthfile#L18

alpine-3-19-fail:
    FROM rust:1.75-alpine3.19

    # Install necessary packages
    RUN apk add --no-cache \
            musl-dev \
            gdb

    COPY --dir .cargo .config crates bin .
    COPY Cargo.toml .
    COPY clippy.toml deny.toml rustfmt.toml .

    RUN mkdir /wasm
    COPY --dir ../wasm+wasi-src/wasi /wasm/wasi
    # Compiled WASM component for benchmarks
    COPY ../wasm/c+build/component.wasm /wasm/c/bench_component.wasm

    # Run benchmarks
    RUN cargo bench --features bench

After the failure it will run an interacting mode -i flag (similar to docker interactive mode).
So you will be able to collect a dump file inside this container.

ulimit -c unlimited && ./target/release/deps/module-<> --bench

Then inspect it and using gdb

gdb  ./target/release/deps/module-<> core -- --bench

Actual Results

SIGABORT

Versions and Environment

Wasmtime version or commit: 17.0.0

Operating system: alpine:3.19

Extra Info

This is totally works on alpine:3.18 version with the libgcc:12 version.
We have found that most probable it is an issue with libgcc because, there is a huge difference with the libgcc unwinding implementation between these two versions.

libgcc_12 - https://github.com/gcc-mirror/gcc/blob/8cbb2cade4c724760c868c9f493b169d6ec4168a/libgcc/unwind-dw2-fde.c#L201

libgcc_13 - https://github.com/gcc-mirror/gcc/blob/0e7bc3eaa36b81004b799124d2fe00137401a43b/libgcc/unwind-dw2-fde.c#L225

You can try it running already prepared earthly target https://github.com/input-output-hk/hermes/blob/libgcc-wasmtime-unwinding-bug/hermes/Earthfile#L38

earthly ./hermes+alpine-3-18-pass

Wasmtime GitHub notifications bot (Feb 27 2024 at 07:57):

Mr-Leshiy commented on issue #7997:

@alexcrichton Yea, this bug appears with this specific WASM component.

Give me some time I will try to make a separate simple project reproducing this thing.

bench_component.wasm.zip

Wasmtime GitHub notifications bot (Feb 27 2024 at 20:23):

alexcrichton commented on issue #7997:

Ok thanks for the component! Looks like just compiling the component, dropping it, and doing that a few times is not sufficient to trigger any bugs. A more minimal repro would be much appreciated, even if it's just a high-level description of how the component is instantiated and worked with. A runnable repro would be perfect!

Otherwise though I'll try to dig in later this week and see if I can't see anything that's awry.

Wasmtime GitHub notifications bot (Feb 29 2024 at 02:16):

alexcrichton commented on issue #7997:

Ok I'm able to reproduce this with:

use wasmtime::*;

fn main() {
    for i in 0.. {
        println!("{i}");
        let mut config = Config::new();
        config.parallel_compilation(false);
        let e = Engine::new(&config).unwrap();
        let m = component::Component::from_file(&e, "/src/bench_component.wasm").unwrap();
    }
}

when specifically compiling with RUSTFLAGS='-Ctarget-feature=-crt-static' in a rust:1.75-alpine3.19 container.

Still trying to figure out what's going on here, but in the meantime setting this config to false should fix the issue.

Wasmtime GitHub notifications bot (Feb 29 2024 at 02:56):

alexcrichton commented on issue #7997:

Ok and I think that https://github.com/bytecodealliance/wasmtime/pull/8028 should completely fix this

Wasmtime GitHub notifications bot (Feb 29 2024 at 07:34):

Mr-Leshiy commented on issue #7997:

@alexcrichton BTW don't you think it could be a bug inside the libgcc as well ?

Wasmtime GitHub notifications bot (Feb 29 2024 at 16:20):

alexcrichton commented on issue #7997:

Perhaps! I wouldn't be able to conclude that with any certainty though. The libgcc code is pretty opaque to me. I think it's probably lucky that this worked before because libgcc's interface is clearly different than libunwind's and we were using the wrong one on musl. With the above PR we should correctly use the libgcc-desired interface on musl by seeing that libunwind isn't available.

Wasmtime GitHub notifications bot (Feb 29 2024 at 17:10):

Mr-Leshiy commented on issue #7997:

I see, thank a lot for your so quick response !
Will apply this patch to our code in days and let you know how it works.

Wasmtime GitHub notifications bot (Feb 29 2024 at 17:11):

Mr-Leshiy edited a comment on issue #7997:

I see, thanks a lot for your so quick response !
Will apply this patch to our code in days and let you know how it works.

Wasmtime GitHub notifications bot (Feb 29 2024 at 17:27):

alexcrichton closed issue #7997:

We have faced a SIGABORT failure during executing simple benchmark which you can find here https://github.com/input-output-hk/hermes/blob/f4d20fc06f0b558be4805a106db32480bcef649d/hermes/bin/src/wasm/module.rs#L118.
Stack trace:

#0  __restore_sigs (set=set@entry=0xffffc437ab50) at ./arch/aarch64/syscall_arch.h:48
#1  0x0000ffff9c2add5c in raise (sig=sig@entry=6) at src/signal/raise.c:11
#2  0x0000ffff9c27cbf4 in abort () at src/exit/abort.c:11
#3  0x0000ffff9c23e8fc in __deregister_frame_info_bases () from /usr/lib/libgcc_s.so.1
#4  0x0000ffff9c23e91c in __deregister_frame () from /usr/lib/libgcc_s.so.1
#5  0x0000aaaac2c0d0a8 in wasmtime_runtime::sys::unix::unwind::{impl#1}::drop (self=<optimized out>) at src/sys/unix/unwind.rs:87
#6  0x0000aaaac2ba2ba8 in core::ptr::drop_in_place<wasmtime_runtime::sys::unix::unwind::UnwindRegistration> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#7  core::ptr::drop_in_place<core::option::Option<wasmtime_runtime::sys::unix::unwind::UnwindRegistration>> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#8  0x0000aaaac2bb4d5c in core::mem::manually_drop::ManuallyDrop<core::option::Option<wasmtime_runtime::sys::unix::unwind::UnwindRegistration>>::drop<core::option::Option<wasmtime_runtime::sys::unix::unwind::UnwindRegistration>> (slot=0x0)
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/mem/manually_drop.rs:144
--Type <RET> for more, q to quit, c to continue without paging--
#9  wasmtime_jit::code_memory::{impl#0}::drop (self=0xffff9c216eb0) at src/code_memory.rs:42
#10 core::ptr::drop_in_place<wasmtime_jit::code_memory::CodeMemory> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#11 alloc::sync::Arc<wasmtime_jit::code_memory::CodeMemory, alloc::alloc::Global>::drop_slow<wasmtime_jit::code_memory::CodeMemory, alloc::alloc::Global> (self=<optimized out>) at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:1752
#12 0x0000aaaac27cfcb8 in alloc::sync::{impl#33}::drop<wasmtime_jit::code_memory::CodeMemory, alloc::alloc::Global> (
    self=0xffff9c213540) at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:2408
#13 core::ptr::drop_in_place<alloc::sync::Arc<wasmtime_jit::code_memory::CodeMemory, alloc::alloc::Global>> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#14 core::ptr::drop_in_place<wasmtime::code::CodeObject> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#15 alloc::sync::Arc<wasmtime::code::CodeObject, alloc::alloc::Global>::drop_slow<wasmtime::code::CodeObject, alloc::alloc::Global> (
    self=<optimized out>) at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:1752
--Type <RET> for more, q to quit, c to continue without paging--
#16 0x0000aaaac27d06a8 in alloc::sync::{impl#33}::drop<wasmtime::code::CodeObject, alloc::alloc::Global> (self=0x0)
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:2408
#17 core::ptr::drop_in_place<alloc::sync::Arc<wasmtime::code::CodeObject, alloc::alloc::Global>> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#18 core::ptr::drop_in_place<wasmtime::component::component::ComponentInner> ()
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#19 alloc::sync::Arc<wasmtime::component::component::ComponentInner, alloc::alloc::Global>::drop_slow<wasmtime::component::component::ComponentInner, alloc::alloc::Global> (self=<optimized out>)
    at /rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:1752
#20 0x0000aaaac275d494 in alloc::sync::{impl#33}::drop<wasmtime::component::component::ComponentInner, alloc::alloc::Global> (
    self=0xffffc437ad40) at rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/alloc/src/sync.rs:2408
#21 core::ptr::drop_in_place<alloc::sync::Arc<wasmtime::component::component::ComponentInner, alloc::alloc::Global>> ()
    at rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
--Type <RET> for more, q to quit, c to continue without paging--
#22 core::ptr::drop_in_place<wasmtime::component::component::Component> ()
    at rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#23 core::ptr::drop_in_place<wasmtime::component::instance::InstancePre<hermes::state::HermesState>> ()
    at rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#24 0x0000aaaac275b390 in core::ptr::drop_in_place<hermes::wasm::module::Module> ()
    at rustc/82e1608dfa6e0b5569232559e3d385fea5a93112/library/core/src/ptr/mod.rs:498
#25 0x0000aaaac2761460 in hermes::wasm::module::bench::module_hermes_component_bench (b=0xffffc437ae30) at bin/src/wasm/module.rs:143

Steps to Reproduce

You can fully reproduce it in already prepared environment and code but cloning our repo and specifically this branch libgcc-wasmtime-unwinding-bug. (repo: https://github.com/input-output-hk/hermes)

Before start you will need to install earthly tool.
To run it you will need to execute the following in the root of the repo

earthly -I ./hermes+alpine-3-19-fail

This will prepare an alpine:3.19 environment with rust:1.75 version and all our project configuration that we have an all source code and run a benchmarks.
Earthfile code https://github.com/input-output-hk/hermes/blob/libgcc-wasmtime-unwinding-bug/hermes/Earthfile#L18

alpine-3-19-fail:
    FROM rust:1.75-alpine3.19

    # Install necessary packages
    RUN apk add --no-cache \
            musl-dev \
            gdb

    COPY --dir .cargo .config crates bin .
    COPY Cargo.toml .
    COPY clippy.toml deny.toml rustfmt.toml .

    RUN mkdir /wasm
    COPY --dir ../wasm+wasi-src/wasi /wasm/wasi
    # Compiled WASM component for benchmarks
    COPY ../wasm/c+build/component.wasm /wasm/c/bench_component.wasm

    # Run benchmarks
    RUN cargo bench --features bench

After the failure it will run an interacting mode -i flag (similar to docker interactive mode).
So you will be able to collect a dump file inside this container.

ulimit -c unlimited && ./target/release/deps/module-<> --bench

Then inspect it and using gdb

gdb  ./target/release/deps/module-<> core -- --bench

Actual Results

SIGABORT

Versions and Environment

Wasmtime version or commit: 17.0.0

Operating system: alpine:3.19

Extra Info

This is totally works on alpine:3.18 version with the libgcc:12 version.
We have found that most probable it is an issue with libgcc because, there is a huge difference with the libgcc unwinding implementation between these two versions.

libgcc_12 - https://github.com/gcc-mirror/gcc/blob/8cbb2cade4c724760c868c9f493b169d6ec4168a/libgcc/unwind-dw2-fde.c#L201

libgcc_13 - https://github.com/gcc-mirror/gcc/blob/0e7bc3eaa36b81004b799124d2fe00137401a43b/libgcc/unwind-dw2-fde.c#L225

You can try it running already prepared earthly target https://github.com/input-output-hk/hermes/blob/libgcc-wasmtime-unwinding-bug/hermes/Earthfile#L38

earthly ./hermes+alpine-3-18-pass

Last updated: Dec 23 2024 at 12:05 UTC