Test rustc bootstrapping on AArch64 #1184

bjorn3 · 2021-07-02T13:04:44Z

No description provided.

afonso360 · 2021-07-02T14:05:39Z

Hey, I mentioned I had an issue natively running test_rustc_tests.sh but the same issue presents when running test_bootstrap.sh.

The issue shows up when compiling the byteorder-1.3.4 crate.

   Compiling proc-macro-error v1.0.4
error: failed to run custom build command for `byteorder v1.3.4`

Caused by:
  process didn't exit successfully: `/home/ubuntu/git/rustc_codegen_cranelift/rust/build/bootstrap/debug/build/byteorder-80e24b4ddeadaf84/build-script-build` (signal: 6, SIGABRT: process abort signal)
  --- stderr
  thread 'main' panicked at 'called `Option::unwrap()` on a `None` value', /home/ubuntu/.cargo/registry/src/jiasu.xzqcsaa.nyc.mn-1ecc6299db9ec823/byteorder-1.3.4/build.rs:54:14
  note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
warning: build failed, waiting for other jobs to finish...
error: build failed
failed to run: /home/ubuntu/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/bin/cargo build --manifest-path /home/ubuntu/git/rustc_codegen_cranelift/rust/src/bootstrap/Cargo.toml
Build completed unsuccessfully in 0:00:33

Here's the offending line.

This is a bit weird because running --version on cg_clif works:

ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ ./build/bin/cg_clif --version
rustc 1.55.0-nightly (6d820866a 2021-06-29)

Even if i set $RUSTC to cg_clif I get the same error:

ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ RUSTC="/home/ubuntu/git/rustc_codegen_cranelift/build/bin/cg_clif" ./scripts/test_bootstrap.sh
...
same error as above

Am I missing anything?

bjorn3 · 2021-07-02T16:27:08Z

Isn't this the offending line: https://github.com/BurntSushi/byteorder/blob/eaaad370715c04d78334d6a5b49c637549c84c71/build.rs#L54 (line 54, not 45) Doesn't look like a None unwrap should be possible at that point though. Can you get a backtrace?

bjorn3 · 2021-07-02T18:47:26Z

I just finished compiling cg_clif for aarch64 using qemu. The panic happens in build_script_build::Version::parse. It doesn't like the cranelift version in the cg_clif -vV output I guess.

afonso360 · 2021-07-02T18:47:59Z

Oh, right, I think I got mixed up, because the issue I was having in test_rustc_tests.sh was about the --version flag not giving us anything.

It is indeed in line 54. Here's the backtrace:

RUST_BACKTRACE=1 ./scripts/test_bootstrap.sh
...
   Compiling crossbeam-utils v0.8.3
error: failed to run custom build command for `byteorder v1.3.4`

Caused by:
  process didn't exit successfully: `/home/ubuntu/git/rustc_codegen_cranelift/rust/build/bootstrap/debug/build/byteorder-80e24b4ddeadaf84/build-script-build` (signal: 6, SIGABRT: process abort signal)
  --- stderr
  thread 'main' panicked at 'called `Option::unwrap()` on a `None` value', /home/ubuntu/.cargo/registry/src/jiasu.xzqcsaa.nyc.mn-1ecc6299db9ec823/byteorder-1.3.4/build.rs:54:14
  stack backtrace:
     0: rust_begin_unwind
               at /home/ubuntu/git/rustc_codegen_cranelift/build_sysroot/sysroot_src/library/std/src/panicking.rs:515:5
     1: core::panicking::panic_fmt
               at /home/ubuntu/git/rustc_codegen_cranelift/build_sysroot/sysroot_src/library/core/src/panicking.rs:92:14
     2: core::panicking::panic
               at /home/ubuntu/git/rustc_codegen_cranelift/build_sysroot/sysroot_src/library/core/src/panicking.rs:50:5
     3: core::str::slice_error_fail
               at /home/ubuntu/git/rustc_codegen_cranelift/build_sysroot/sysroot_src/library/core/src/option.rs:388:21
     4: core::str::traits::<impl core::slice::index::SliceIndex<str> for core::ops::range::RangeFrom<usize>>::index
               at /home/ubuntu/git/rustc_codegen_cranelift/build_sysroot/sysroot_src/library/core/src/str/traits.rs:365:21
     5: core::str::traits::<impl core::ops::index::Index<I> for str>::index
               at /home/ubuntu/git/rustc_codegen_cranelift/build_sysroot/sysroot_src/library/core/src/str/traits.rs:64:9
     6: build_script_build::Version::parse
               at ./build.rs:54:14
     7: build_script_build::Version::read
               at ./build.rs:47:9
     8: build_script_build::main
               at ./build.rs:10:25
     9: core::ops::function::FnOnce::call_once
               at /home/ubuntu/git/rustc_codegen_cranelift/build_sysroot/sysroot_src/library/core/src/ops/function.rs:227:5
  note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
warning: build failed, waiting for other jobs to finish...
error: build failed
failed to run: /home/ubuntu/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/bin/cargo build --manifest-path /home/ubuntu/git/rustc_codegen_cranelift/rust/src/bootstrap/Cargo.toml
Build completed unsuccessfully in 0:00:07

bjorn3 · 2021-07-02T18:51:08Z

It doesn't like the cranelift version in the cg_clif -vV output I guess.

Never mind. That only happens when you explicitly pass -Zcodegen-backend together with the -vV.

bjorn3 · 2021-07-02T19:00:34Z

The error seems to happen at https://github.com/rust-lang/rust/blob/851c82e88ade86bfe3b4ee785d5e5ab1d954b61c/library/core/src/str/mod.rs#L112 suggesting a non-utf8 string.

bjorn3 · 2021-07-02T19:12:08Z

With a debug sysroot:

   Compiling byteorder v1.3.4 (/home/bjorn/Projects/cg_clif2/byteorder)
error: failed to run custom build command for `byteorder v1.3.4 (/home/bjorn/Projects/cg_clif2/byteorder)`

Caused by:
  process didn't exit successfully: `/home/bjorn/Projects/cg_clif2/byteorder/target/debug/build/byteorder-25926a0b98921ea4/build-script-build` (signal: 15, SIGTERM: termination signal)
  --- stderr
  thread 'main' panicked at 'byte index 6 is not a char boundary; it is inside 'r' (bytes 0..1) of `rustc 1.55.0-nightly (6d820866a 2021-06-29)
  `', build.rs:54:14
  stack backtrace:
  thread 'main' panicked at 'byte index 7 is not a char boundary; it is inside '.' (bytes 0..1) of `.debug_addr`', sysroot_src/library/std/src/../../backtrace/src/symbolize/gimli/elf.rs:121:26

and then it hangs. Couldn't even see this until I killed qemu.

afonso360 · 2021-07-02T19:18:59Z

Hmm, I modified the build.rs file to include a dbg!(s). It seems to be correct:

[build.rs:51] s = "rustc 1.55.0-nightly (6d820866a 2021-06-29)\n"

afonso360 · 2021-07-02T19:25:59Z

Here's a minimal repro case:

fn main() {
    let s = "rustc 1.55.0-nightly (6d820866a 2021-06-29)\n";
    &s[6..];
}

This is enough to crash:

ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ ./build/bin/cg_clif ./test.rs
ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ ./test
thread 'main' panicked at 'called `Option::unwrap()` on a `None` value', ./test.rs:3:6
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
Aborted

Edit:
We are misscompiling str::is_char_boundary

fn main() {
    let s = "rustc 1.55.0-nightly (6d820866a 2021-06-29)\n";
    println!("{}", s.is_char_boundary(6));
}

ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ rustc ./test.rs
ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ ./test
true
ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ ./build/bin/cg_clif ./test.rs
ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ ./test
false

Edit 2:

Narrowing it down, it seems that we are miss compiling a comparison after a dereference. I'm going to try to get a clif test file out of this.

fn main() {
        dbg!({
                let a: i32 = 49;
                let a_ref = &a;
                (*a_ref as i8) >= -0x40
        });
        dbg!((49 as i8) >= -0x40);
}

ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ ./build/bin/cg_clif ./test.rs
ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ ./test
[./test.rs:2] { let a: i32 = 49; let a_ref = &a; (*a_ref as i8) >= -0x40 } = false
[./test.rs:7] (49 as i8) >= -0x40 = true
ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ rustc ./test.rs
ubuntu@ubuntu:~/git/rustc_codegen_cranelift$ ./test
[./test.rs:2] { let a: i32 = 49; let a_ref = &a; (*a_ref as i8) >= -0x40 } = true
[./test.rs:7] (49 as i8) >= -0x40 = true

Edit 3:
Here's the clif test file:

test run
target aarch64
target x86_64 machinst

function %test() -> b1 {
    ss0 = explicit_slot 16

block0:
    v0 = iconst.i32 49
    stack_store v0, ss0
    v1 = stack_addr.i64 ss0
    v2 = load.i32 notrap v1
    v3 = ireduce.i8 v2
    v4 = iconst.i8 192
    v5 = icmp sge v3, v4
    return v5
}
; run: %test() == true

This passes on x64 and fails on aarch64. I'll keep working on this tommorow.

bjorn3 · 2021-07-03T09:04:48Z

Thanks for the debugging!

When encoding constants as immediates into an RSE Imm12 instruction we need to take special care to check if the value that we are trying to input does not overflow its type when viewed as a signed value. (i.e. iconst.i8 200) We cannot both put an immediate and sign extend it, so we need to lower it into a separate reg, and emit the sign extend into the instruction. For more details see the [cg_clif bug report](https://github.com/bjorn3/rustc_codegen_cranelift/issues/1184#issuecomment-873214796).

This has a fix for a miscompilation on AArch64 cc #1184

bjorn3 · 2021-07-04T16:35:53Z

Now getting

   Compiling merge_derive v0.1.0
qemu: uncaught target signal 11 (Segmentation fault) - core dumped
error: could not compile `merge_derive`

Caused by:
  process didn't exit successfully: `/home/bjorn/Projects/cg_clif2/rust/../build/bin/cg_clif --crate-name merge_derive --edition=2018 /home/bjorn/.cargo/registry/src/jiasu.xzqcsaa.nyc.mn-1ecc6299db9ec823/merge_derive-0.1.0/src/lib.rs --error-format=json --json=diagnostic-rendered-ansi --crate-type proc-macro --emit=dep-info,link -C prefer-dynamic -C embed-bitcode=no -C debuginfo=2 -C metadata=9ead013e1ae016e2 -C extra-filename=-9ead013e1ae016e2 --out-dir /home/bjorn/Projects/cg_clif2/rust/build/bootstrap/debug/deps -C linker=aarch64-linux-gnu-gcc -L dependency=/home/bjorn/Projects/cg_clif2/rust/build/bootstrap/debug/deps --extern proc_macro_error=/home/bjorn/Projects/cg_clif2/rust/build/bootstrap/debug/deps/libproc_macro_error-d02fc3c73f17a371.rlib --extern proc_macro2=/home/bjorn/Projects/cg_clif2/rust/build/bootstrap/debug/deps/libproc_macro2-de20e217372061ee.rlib --extern quote=/home/bjorn/Projects/cg_clif2/rust/build/bootstrap/debug/deps/libquote-51534bb9a71bfbd9.rlib --extern syn=/home/bjorn/Projects/cg_clif2/rust/build/bootstrap/debug/deps/libsyn-f967af3d96e2e23c.rlib --extern proc_macro --cap-lints allow -Cdebuginfo=2 -Wrust_2018_idioms -Wunused_lifetimes -Wsemicolon_in_expressions_from_macros` (signal: 11, SIGSEGV: invalid memory reference)

bjorn3 · 2021-07-04T16:39:48Z

This seems to be an abi incompatibility with cg_llvm:

Backtrace

#0  0x000000400827a204 in ?? () from /usr/aarch64-linux-gnu/lib/libc.so.6
#1  0x0000004019b256e4 in core::slice::<impl [T]>::copy_from_slice () at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/core/src/intrinsics.rs:1861
#2  0x0000004019b1d6ec in <u32 as proc_macro::bridge::rpc::DecodeMut<S>>::decode () at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/rpc.rs:39
#3  0x0000004019b17b20 in <core::num::nonzero::NonZeroU32 as proc_macro::bridge::rpc::DecodeMut<S>>::decode ()
    at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/rpc.rs:168
#4  0x0000004019b159fc in <proc_macro::bridge::client::TokenStream as proc_macro::bridge::rpc::DecodeMut<S>>::decode ()
    at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/client.rs:116
#5  0x0000004019b16bf8 in <(A,B) as proc_macro::bridge::rpc::DecodeMut<S>>::decode () at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/rpc.rs:183
#6  0x0000004019b15d64 in proc_macro::bridge::client::run_client::{{closure}}::{{closure}} ()
    at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/client.rs:372
#7  0x0000004019b1c100 in proc_macro::bridge::scoped_cell::ScopedCell<T>::set::{{closure}} ()
    at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/scoped_cell.rs:80
#8  0x0000004019b1c41c in proc_macro::bridge::scoped_cell::ScopedCell<T>::replace ()
    at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/scoped_cell.rs:75
#9  0x0000004019b1c088 in proc_macro::bridge::scoped_cell::ScopedCell<T>::set () at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/scoped_cell.rs:80
#10 0x0000004019b16914 in proc_macro::bridge::client::<impl proc_macro::bridge::Bridge>::enter::{{closure}} ()
    at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/client.rs:325
#11 0x0000004019b1bec4 in std::thread::local::LocalKey<T>::try_with () at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/std/src/thread/local.rs:399
#12 0x0000004019b1b3e0 in std::thread::local::LocalKey<T>::with () at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/std/src/thread/local.rs:375
#13 0x0000004019b19580 in proc_macro::bridge::client::<impl proc_macro::bridge::Bridge>::enter ()
    at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/client.rs:325
#14 0x0000004019b15c24 in proc_macro::bridge::client::run_client::{{closure}} () at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/client.rs:370
#15 0x0000004019b2386c in <std::panic::AssertUnwindSafe<F> as core::ops::function::FnOnce<()>>::call_once ()
    at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/std/src/panic.rs:341
#16 0x0000004019b21364 in std::panicking::try::do_call () at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/std/src/panicking.rs:401
#17 0x0000004019b21170 in std::panicking::try () at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/std/src/panicking.rs:365
#18 0x0000004019b15b18 in proc_macro::bridge::client::run_client () at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/std/src/panic.rs:428
#19 0x0000004019b16048 in proc_macro::bridge::client::Client<fn(proc_macro::TokenStream,proc_macro::TokenStream) -> proc_macro::TokenStream>::expand2::run ()
    at /home/bjorn/Projects/cg_clif2/build_sysroot/sysroot_src/library/proc_macro/src/bridge/client.rs:425
#20 0x00000040054f48f0 in proc_macro::bridge::server::run_server () from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#21 0x00000040054d6c74 in <rustc_expand::proc_macro::AttrProcMacro as rustc_expand::base::AttrProcMacro>::expand ()
   from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#22 0x00000040054c4a5c in rustc_expand::expand::MacroExpander::fully_expand_fragment ()
   from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#23 0x00000040054c2630 in rustc_expand::expand::MacroExpander::expand_crate ()
   from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#24 0x0000004001b79d90 in rustc_session::utils::<impl rustc_session::session::Session>::time ()
   from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#25 0x0000004001b9d9b8 in rustc_interface::passes::configure_and_expand_inner ()
   from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#26 0x0000004001b77e44 in rustc_interface::passes::boxed_resolver::BoxedResolver::new ()
   from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#27 0x0000004001b83f60 in rustc_interface::queries::Queries::expansion ()
   from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#28 0x0000004001abba3c in rustc_interface::queries::<impl rustc_interface::interface::Compiler>::enter ()
   from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#29 0x0000004001a892ac in rustc_span::with_source_map () from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#30 0x0000004001abcc04 in rustc_interface::interface::create_compiler_and_run () from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#31 0x0000004001a8e3ec in scoped_tls::ScopedKey<T>::set () from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#32 0x0000004001abd740 in std::sys_common::backtrace::__rust_begin_short_backtrace () from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#33 0x0000004001a70fcc in core::ops::function::FnOnce::call_once{{vtable-shim}} () from /home/bjorn/.rustup/toolchains/nightly-2021-06-30-aarch64-unknown-linux-gnu/lib/librustc_driver-5839d4fe983c4c9c.so
#34 0x0000004008100674 in <alloc::boxed::Box<F,A> as core::ops::function::FnOnce<Args>>::call_once () at /rustc/6d820866a27b1949e237be79b9c8c0145fe728b7/library/alloc/src/boxed.rs:1575
#35 <alloc::boxed::Box<F,A> as core::ops::function::FnOnce<Args>>::call_once () at /rustc/6d820866a27b1949e237be79b9c8c0145fe728b7/library/alloc/src/boxed.rs:1575
#36 std::sys::unix::thread::Thread::new::thread_start () at library/std/src/sys/unix/thread.rs:72
#37 0x00000040083867e4 in start_thread () from /usr/aarch64-linux-gnu/lib/libpthread.so.0
#38 0x00000040082c698c in ?? () from /usr/aarch64-linux-gnu/lib/libc.so.6

bjorn3 · 2021-07-04T17:06:52Z

At the moment of the crash the register state is:

x0             0x4008d61820        275026155552
x1             0x910003fda9bf7bfd  -7998388550200493059
x2             0x4                 4

For a memcpy that represents the dst, src and count arguments. They may have already been clobbered by this point though. I can't seem to get attaching a debugger to qemu emulated programs working natively though, so I can't really tell if it was clobbered or not.

…jorn3 Sync rustc_codegen_cranelift The main hightlight this sync is basic support for AArch64. Most things should work on Linux, but there does seem to be an ABI incompatibility causing proc-macros to crash, see https://github.com/bjorn3/rustc_codegen_cranelift/issues/1184. Thanks to `@afonso360` for implementing all Cranelift features that were necessary to compile for AArch64 using cg_clif. Also thanks to `@shamatar` for implementing the `llvm.x86.addcarry.64` and `llvm.x86.subborrow.64` llvm intrinsics used by num-bigint (https://github.com/bjorn3/rustc_codegen_cranelift/pull/1178) and `@eggyal` for implementing multi-threading support for the lazy jit mode. (https://github.com/bjorn3/rustc_codegen_cranelift/pull/1166) r? `@ghost` `@rustbot` label +A-codegen +A-cranelift +T-compiler

…jorn3 Sync rustc_codegen_cranelift The main hightlight this sync is basic support for AArch64. Most things should work on Linux, but there does seem to be an ABI incompatibility causing proc-macros to crash, see https://github.com/bjorn3/rustc_codegen_cranelift/issues/1184. Thanks to ``@afonso360`` for implementing all Cranelift features that were necessary to compile for AArch64 using cg_clif. Also thanks to ``@shamatar`` for implementing the `llvm.x86.addcarry.64` and `llvm.x86.subborrow.64` llvm intrinsics used by num-bigint (https://github.com/bjorn3/rustc_codegen_cranelift/pull/1178) and ``@eggyal`` for implementing multi-threading support for the lazy jit mode. (https://github.com/bjorn3/rustc_codegen_cranelift/pull/1166) r? ``@ghost`` ``@rustbot`` label +A-codegen +A-cranelift +T-compiler

…jorn3 Sync rustc_codegen_cranelift The main hightlight this sync is basic support for AArch64. Most things should work on Linux, but there does seem to be an ABI incompatibility causing proc-macros to crash, see https://github.com/bjorn3/rustc_codegen_cranelift/issues/1184. Thanks to ```@afonso360``` for implementing all Cranelift features that were necessary to compile for AArch64 using cg_clif. Also thanks to ```@shamatar``` for implementing the `llvm.x86.addcarry.64` and `llvm.x86.subborrow.64` llvm intrinsics used by num-bigint (https://github.com/bjorn3/rustc_codegen_cranelift/pull/1178) and ```@eggyal``` for implementing multi-threading support for the lazy jit mode. (https://github.com/bjorn3/rustc_codegen_cranelift/pull/1166) r? ```@ghost``` ```@rustbot``` label +A-codegen +A-cranelift +T-compiler

bjorn3 · 2022-09-01T16:41:54Z

With de7342b and 244455d, all rustc tests now pass on AArch64 with the Cranelift main branch. Bootstrapping rustc also works.

bjorn3 added C-enhancement Category: An issue proposing an enhancement or a PR with one. O-arm Target: ARM processors (arm, thumb and AArch64 targets) A-testing Area: Testing of cg_clif itself labels Jul 2, 2021

afonso360 mentioned this issue Jul 3, 2021

aarch64: Fix incorrect encoding of large const values in icmp. bytecodealliance/wasmtime#3056

Merged

bjorn3 added a commit that referenced this issue Jul 4, 2021

Update Cranelift

ac730b4

This has a fix for a miscompilation on AArch64 cc #1184

bjorn3 mentioned this issue Jul 7, 2021

Sync rustc_codegen_cranelift rust-lang/rust#86927

Merged

afonso360 mentioned this issue Aug 6, 2022

Add abi-checker to y.rs and run it on CI #1255

Merged

bjorn3 mentioned this issue Sep 1, 2022

Tracking issue for fixes in Cranelift 0.88.0 #1276

Closed

bjorn3 mentioned this issue Sep 27, 2022

Update to Cranelift 0.88.1 #1283

Merged

bjorn3 closed this as completed in #1283 Sep 27, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Test rustc bootstrapping on AArch64 #1184

Test rustc bootstrapping on AArch64 #1184

bjorn3 commented Jul 2, 2021

afonso360 commented Jul 2, 2021

bjorn3 commented Jul 2, 2021

bjorn3 commented Jul 2, 2021

afonso360 commented Jul 2, 2021

bjorn3 commented Jul 2, 2021

bjorn3 commented Jul 2, 2021

bjorn3 commented Jul 2, 2021

afonso360 commented Jul 2, 2021

afonso360 commented Jul 2, 2021 •

edited

Loading

bjorn3 commented Jul 3, 2021

bjorn3 commented Jul 4, 2021

bjorn3 commented Jul 4, 2021

bjorn3 commented Jul 4, 2021

bjorn3 commented Sep 1, 2022

Test rustc bootstrapping on AArch64 #1184

Test rustc bootstrapping on AArch64 #1184

Comments

bjorn3 commented Jul 2, 2021

afonso360 commented Jul 2, 2021

bjorn3 commented Jul 2, 2021

bjorn3 commented Jul 2, 2021

afonso360 commented Jul 2, 2021

bjorn3 commented Jul 2, 2021

bjorn3 commented Jul 2, 2021

bjorn3 commented Jul 2, 2021

afonso360 commented Jul 2, 2021

afonso360 commented Jul 2, 2021 • edited Loading

bjorn3 commented Jul 3, 2021

bjorn3 commented Jul 4, 2021

bjorn3 commented Jul 4, 2021

bjorn3 commented Jul 4, 2021

bjorn3 commented Sep 1, 2022

afonso360 commented Jul 2, 2021 •

edited

Loading