WIP libgccjit codegen backend for rust

This is a GCC codegen for rustc, which means it can be loaded by the existing rustc frontend, but benefits from GCC: more architectures are supported and GCC's optimizations are used.

Despite its name, libgccjit can be used for ahead-of-time compilation, as is used here.

Motivation

The primary goal of this project is to be able to compile Rust code on platforms unsupported by LLVM. A secondary goal is to check if using the gcc backend will provide any run-time speed improvement for the programs compiled using rustc.

Building

This requires a patched libgccjit in order to work. You need to use my fork of gcc which already includes these patches.

To build it (most of these instructions come from here, so don't hesitate to take a look there if you encounter an issue):

$ git clone https://github.com/antoyo/gcc
$ sudo apt install flex libmpfr-dev libgmp-dev libmpc3 libmpc-dev
$ mkdir gcc-build gcc-install
$ cd gcc-build
$ ../gcc/configure \
    --enable-host-shared \
    --enable-languages=jit \
    --enable-checking=release \ # it enables extra checks which allow to find bugs
    --disable-bootstrap \
    --disable-multilib \
    --prefix=$(pwd)/../gcc-install
$ make -j4 # You can replace `4` with another number depending on how many cores you have.

If you want to run libgccjit tests, you will need to also enable the C++ language in the configure:

--enable-languages=jit,c++

Then to run libgccjit tests:

$ cd gcc # from the `gcc-build` folder
$ make check-jit
# To run one specific test:
$ make check-jit RUNTESTFLAGS="-v -v -v jit.exp=jit.dg/test-asm.cc"

Put the path to your custom build of libgccjit in the file config.toml.

If you followed the instructions exactly as written (ie, you have created a gcc-build folder where gcc is built), the only thing you need to do is:

$ cp config.example.toml config.toml

But if you did something different, you also need to set the gcc-path value in config.toml with the result of this command:

$ dirname $(readlink -f `find . -name libgccjit.so`)

Then you can run commands like this:

$ ./y.sh prepare # download and patch sysroot src and install hyperfine for benchmarking
$ ./y.sh build --release

To run the tests:

$ ./y.sh test --release

Usage

$CG_GCCJIT_DIR is the directory you cloned this repo into in the following instructions:

export CG_GCCJIT_DIR=[the full path to rustc_codegen_gcc]

Cargo

$ CHANNEL="release" $CG_GCCJIT_DIR/y.sh cargo run

If you compiled cg_gccjit in debug mode (aka you didn't pass --release to ./y.sh test) you should use CHANNEL="debug" instead or omit CHANNEL="release" completely.

LTO

To use LTO, you need to set the variable FAT_LTO=1 and EMBED_LTO_BITCODE=1 in addition to setting lto = "fat" in the Cargo.toml. Don't set FAT_LTO when compiling the sysroot, though: only set EMBED_LTO_BITCODE=1.

Failing to set EMBED_LTO_BITCODE will give you the following error:

error: failed to copy bitcode to object file: No such file or directory (os error 2)

Rustc

You should prefer using the Cargo method.

$ LIBRARY_PATH="[gcc-path value]" LD_LIBRARY_PATH="[gcc-path value]" rustc +$(cat $CG_GCCJIT_DIR/rust-toolchain | grep 'channel' | cut -d '=' -f 2 | sed 's/"//g' | sed 's/ //g') -Cpanic=abort -Zcodegen-backend=$CG_GCCJIT_DIR/target/release/librustc_codegen_gcc.so --sysroot $CG_GCCJIT_DIR/build_sysroot/sysroot my_crate.rs

Env vars

CG_GCCJIT_INCR_CACHE_DISABLED

Don't cache object files in the incremental cache. Useful during development of cg_gccjit to make it possible to use incremental mode for all analyses performed by rustc without caching object files when their content should have been changed by a change to cg_gccjit.

CG_GCCJIT_DISPLAY_CG_TIME

Display the time it took to perform codegen for a crate

CG_RUSTFLAGS

Send additional flags to rustc. Can be used to build the sysroot without unwinding by setting `CG_RUSTFLAGS=-Cpanic=abort`.

CG_GCCJIT_DUMP_TO_FILE

Dump a C-like representation to /tmp/gccjit_dumps and enable debug info in order to debug this C-like representation.

Licensing

While this crate is licensed under a dual Apache/MIT license, it links to libgccjit which is under the GPLv3+ and thus, the resulting toolchain (rustc + GCC codegen) will need to be released under the GPL license.

However, programs compiled with rustc_codegen_gcc do not need to be released under a GPL license.

Debugging

Sometimes, libgccjit will crash and output an error like this:

during RTL pass: expand
libgccjit.so: error: in expmed_mode_index, at expmed.h:249
0x7f0da2e61a35 expmed_mode_index
	../../../gcc/gcc/expmed.h:249
0x7f0da2e61aa4 expmed_op_cost_ptr
	../../../gcc/gcc/expmed.h:271
0x7f0da2e620dc sdiv_cost_ptr
	../../../gcc/gcc/expmed.h:540
0x7f0da2e62129 sdiv_cost
	../../../gcc/gcc/expmed.h:558
0x7f0da2e73c12 expand_divmod(int, tree_code, machine_mode, rtx_def*, rtx_def*, rtx_def*, int)
	../../../gcc/gcc/expmed.c:4335
0x7f0da2ea1423 expand_expr_real_2(separate_ops*, rtx_def*, machine_mode, expand_modifier)
	../../../gcc/gcc/expr.c:9240
0x7f0da2cd1a1e expand_gimple_stmt_1
	../../../gcc/gcc/cfgexpand.c:3796
0x7f0da2cd1c30 expand_gimple_stmt
	../../../gcc/gcc/cfgexpand.c:3857
0x7f0da2cd90a9 expand_gimple_basic_block
	../../../gcc/gcc/cfgexpand.c:5898
0x7f0da2cdade8 execute
	../../../gcc/gcc/cfgexpand.c:6582

To see the code which causes this error, call the following function:

gcc_jit_context_dump_to_file(ctxt, "/tmp/output.c", 1 /* update_locations */)

This will create a C-like file and add the locations into the IR pointing to this C file. Then, rerun the program and it will output the location in the second line:

libgccjit.so: /tmp/something.c:61322:0: error: in expmed_mode_index, at expmed.h:249

Or add a breakpoint to add_error in gdb and print the line number using:

p loc->m_line
p loc->m_filename->m_buffer

To print a debug representation of a tree:

debug_tree(expr);

(defined in print-tree.h)

To print a debug representation of a gimple struct:

debug_gimple_stmt(gimple_struct)

To get the rustc command to run in gdb, add the --verbose flag to cargo build.

To have the correct file paths in gdb instead of /usr/src/debug/gcc/libstdc++-v3/libsupc++/eh_personality.cc:

Maybe by calling the following at the beginning of gdb:

set substitute-path /usr/src/debug/gcc /path/to/gcc-repo/gcc

TODO(antoyo): but that's not what I remember I was doing.

failed to build archive error

When you get this error:

error: failed to build archive: failed to open object file: No such file or directory (os error 2)

That can be caused by the fact that you try to compile with lto = "fat", but you didn't compile the sysroot with LTO. (Not sure if that's the reason since I cannot reproduce anymore. Maybe it happened when forgetting setting FAT_LTO.)

ld: cannot find crtbegin.o

When compiling an executable with libgccijt, if setting the *LIBRARY_PATH variables to the install directory, you will get the following errors:

ld: cannot find crtbegin.o: No such file or directory
ld: cannot find -lgcc: No such file or directory
ld: cannot find -lgcc: No such file or directory
libgccjit.so: error: error invoking gcc driver

To fix this, set the variables to gcc-build/build/gcc.

How to debug GCC LTO

Run do the command with -v -save-temps and then extract the lto1 line from the output and run that under the debugger.

How to send arguments to the GCC linker

CG_RUSTFLAGS="-Clink-args=-save-temps -v" ../y.sh cargo build

How to see the personality functions in the asm dump

CG_RUSTFLAGS="-Clink-arg=-save-temps -v -Clink-arg=-dA" ../y.sh cargo build

How to see the LLVM IR for a sysroot crate

cargo build -v --target x86_64-unknown-linux-gnu -Zbuild-std
# Take the command from the output and add --emit=llvm-ir

To prevent the linker from unmangling symbols

Run with:

COLLECT_NO_DEMANGLE=1

How to use a custom-build rustc

• Build the stage2 compiler (rustup toolchain link debug-current build/x86_64-unknown-linux-gnu/stage2).
• Clean and rebuild the codegen with debug-current in the file rust-toolchain.

How to use mem-trace

rustc needs to be built without jemalloc so that mem-trace can overload malloc since jemalloc is linked statically, so a LD_PRELOAD-ed library won't a chance to intercept the calls to malloc.

How to generate GIMPLE

If you need to check what gccjit is generating (GIMPLE), then take a look at how to generate it in gimple.md.

How to build a cross-compiling libgccjit

Building libgccjit

• Follow the instructions on this repo.

Configuring rustc_codegen_gcc

• Run ./y.sh prepare --cross so that the sysroot is patched for the cross-compiling case.
• Set the path to the cross-compiling libgccjit in gcc-path (in config.toml).
• Make sure you have the linker for your target (for instance m68k-unknown-linux-gnu-gcc) in your $PATH. Currently, the linker name is hardcoded as being $TARGET-gcc. Specify the target when building the sysroot: ./y.sh build --target-triple m68k-unknown-linux-gnu.
• Build your project by specifying the target: OVERWRITE_TARGET_TRIPLE=m68k-unknown-linux-gnu ../y.sh cargo build --target m68k-unknown-linux-gnu.

If the target is not yet supported by the Rust compiler, create a target specification file (note that the arch specified in this file must be supported by the rust compiler). Then, you can use it the following way:

• Add the target specification file using --target as an absolute path to build the sysroot: ./y.sh build --target-triple m68k-unknown-linux-gnu --target $(pwd)/m68k-unknown-linux-gnu.json
• Build your project by specifying the target specification file: OVERWRITE_TARGET_TRIPLE=m68k-unknown-linux-gnu ../y.sh cargo build --target path/to/m68k-unknown-linux-gnu.json.

If you get the following error:

/usr/bin/ld: unrecognised emulation mode: m68kelf

Make sure you set gcc-path (in config.toml) to the install directory.