Move from an hardcoded extern _start to a dynamic proc_macro

Cargo.toml

@@ -1,3 +1,6 @@
+[workspace]
+members = ["macros"]
+
 [package]
 name = "origin"
 version = "0.12.1"
@@ -21,6 +24,9 @@ memoffset = { version = "0.9.0", optional = true }
 log = { version = "0.4.14", default-features = false, optional = true }
 rustix-futex-sync = "0.1.0"
 
+# Provides "origin::main" macro
+origin-macros = { version = "0.1.0", path = "./macros", optional = true }
+
 # Optional logging backends. You can use any external logger, but using these
 # features allows origin to initialize the logger before main, so that you can
 # see the log messages emitted before main is called.
@@ -55,6 +61,8 @@ rustc-dep-of-std = [
     "libc/rustc-dep-of-std",
 ]
 
+macros = ["dep:origin-macros"]
+
 # Use origin's implementation of program startup and shutdown.
 origin-program = []
 

example-crates/origin-start-lto/Cargo.toml

@@ -7,7 +7,7 @@ publish = false
 [dependencies]
 # Origin can be depended on just like any other crate. For no_std, disable
 # the default features, and the desired features.
-origin = { path = "../..", default-features = false, features = ["origin-program", "origin-thread", "origin-start"] }
+origin = { path = "../..", default-features = false, features = ["origin-program", "origin-thread", "origin-start", "macros"] }
 
 # Crates to help writing no_std code.
 atomic-dbg = { version = "0.1.8", default-features = false }

example-crates/origin-start-lto/src/main.rs

@@ -5,6 +5,7 @@
 #![allow(internal_features)]
 #![feature(lang_items)]
 #![feature(core_intrinsics)]
+#![feature(naked_functions)]
 
 extern crate alloc;
 extern crate compiler_builtins;
@@ -26,8 +27,8 @@ extern "C" fn eh_personality() {}
 #[global_allocator]
 static GLOBAL_ALLOCATOR: rustix_dlmalloc::GlobalDlmalloc = rustix_dlmalloc::GlobalDlmalloc;
 
-#[no_mangle]
-extern "C" fn main(_argc: i32, _argv: *const *const u8) -> i32 {
+#[origin::main]
+fn main() -> i32 {
     eprintln!("Hello from main thread");
 
     at_exit(Box::new(|| eprintln!("Hello from an at_exit handler")));
@@ -53,5 +54,6 @@ extern "C" fn main(_argc: i32, _argv: *const *const u8) -> i32 {
     }
 
     eprintln!("Goodbye from main");
-    exit(0);
+
+    0
 }

example-crates/origin-start-no-alloc/Cargo.toml

@@ -7,7 +7,7 @@ publish = false
 [dependencies]
 # Origin can be depended on just like any other crate. For no_std, disable
 # the default features, and the desired features.
-origin = { path = "../..", default-features = false, features = ["origin-program", "origin-start"] }
+origin = { path = "../..", default-features = false, features = ["origin-program", "origin-start", "macros"] }
 
 # Crates to help writing no_std code.
 atomic-dbg = { version = "0.1.8", default-features = false }

example-crates/origin-start-no-alloc/src/main.rs

@@ -5,11 +5,11 @@
 #![allow(internal_features)]
 #![feature(lang_items)]
 #![feature(core_intrinsics)]
+#![feature(naked_functions)]
 
 extern crate compiler_builtins;
 
 use atomic_dbg::{dbg, eprintln};
-use origin::program::*;
 
 #[panic_handler]
 fn panic(panic: &core::panic::PanicInfo<'_>) -> ! {
@@ -20,12 +20,12 @@ fn panic(panic: &core::panic::PanicInfo<'_>) -> ! {
 #[lang = "eh_personality"]
 extern "C" fn eh_personality() {}
 
-#[no_mangle]
-extern "C" fn main(_argc: i32, _argv: *const *const u8) -> i32 {
+#[origin::main]
+fn start() -> i32 {
     eprintln!("Hello!");
 
     // Unlike origin-start, this example can't create threads because origin's
     // thread support requires an allocator.
 
-    exit(0);
+    0
 }

example-crates/origin-start-tiny/Cargo.toml

@@ -7,7 +7,7 @@ publish = false
 [dependencies]
 # Origin can be depended on just like any other crate. For no_std, disable
 # the default features, and the desired features.
-origin = { path = "../..", default-features = false, features = ["origin-program", "origin-start"] }
+origin = { path = "../..", default-features = false, features = ["origin-program", "origin-start", "macros"] }
 
 # Crates to help writing no_std code.
 compiler_builtins = { version = "0.1.101", features = ["mem"] }

example-crates/origin-start-tiny/src/main.rs

@@ -5,9 +5,10 @@
 #![allow(internal_features)]
 #![feature(lang_items)]
 #![feature(core_intrinsics)]
+#![feature(naked_functions)]
 
-extern crate origin;
 extern crate compiler_builtins;
+extern crate origin;
 
 #[panic_handler]
 fn panic(_panic: &core::panic::PanicInfo<'_>) -> ! {
@@ -17,7 +18,7 @@ fn panic(_panic: &core::panic::PanicInfo<'_>) -> ! {
 #[lang = "eh_personality"]
 extern "C" fn eh_personality() {}
 
-#[no_mangle]
-extern "C" fn main(_argc: i32, _argv: *const *const u8) -> i32 {
+#[origin::start]
+fn main() -> i32 {
     42
 }

example-crates/origin-start/Cargo.toml

@@ -7,7 +7,7 @@ publish = false
 [dependencies]
 # Origin can be depended on just like any other crate. For no_std, disable
 # the default features, and the desired features.
-origin = { path = "../..", default-features = false, features = ["origin-program", "origin-thread", "origin-start"] }
+origin = { path = "../..", default-features = false, features = ["origin-program", "origin-thread", "origin-start", "macros"] }
 
 # Crates to help writing no_std code.
 atomic-dbg = { version = "0.1.8", default-features = false }

example-crates/origin-start/src/main.rs

@@ -5,6 +5,7 @@
 #![allow(internal_features)]
 #![feature(lang_items)]
 #![feature(core_intrinsics)]
+#![feature(naked_functions)]
 
 extern crate alloc;
 extern crate compiler_builtins;
@@ -26,8 +27,8 @@ extern "C" fn eh_personality() {}
 #[global_allocator]
 static GLOBAL_ALLOCATOR: rustix_dlmalloc::GlobalDlmalloc = rustix_dlmalloc::GlobalDlmalloc;
 
-#[no_mangle]
-extern "C" fn main(_argc: i32, _argv: *const *const u8) -> i32 {
+#[origin::main]
+fn main() -> i32 {
     eprintln!("Hello from main thread");
 
     at_exit(Box::new(|| eprintln!("Hello from an at_exit handler")));

macros/Cargo.toml

@@ -0,0 +1,23 @@
+[package]
+name = "origin-macros"
+version = "0.1.0"
+edition = "2021"
+authors = [
+    "Dan Gohman <dev@sunfishcode.online>",
+    "Federico Maria Morrone <contact@morrone.dev>",
+]
+description = "Macros for origin crate"
+documentation = "https://docs.rs/origin-macros"
+license = "Apache-2.0 WITH LLVM-exception OR Apache-2.0 OR MIT"
+repository = "https://github.com/sunfishcode/origin"
+keywords = ["linux"]
+categories = ["no-std"]
+include = ["src", "Cargo.toml", "../COPYRIGHT", "../LICENSE*", "README.md"]
+
+[lib]
+proc-macro = true
+
+[dependencies]
+proc-macro2 = "1.0.66"
+quote = "1.0.33"
+syn = { version = "2.0.31", features = ["full"] }

macros/README.md

@@ -0,0 +1 @@
+# Origin Macros

macros/src/lib.rs

@@ -0,0 +1,88 @@
+use proc_macro::TokenStream;
+use proc_macro2::TokenStream as TokenStream2;
+use quote::quote;
+use syn::{parse_macro_input, ItemFn};
+
+#[proc_macro_attribute]
+pub fn main(_attr: TokenStream, input: TokenStream) -> TokenStream {
+    let main_fn: ItemFn = parse_macro_input!(input);
+    let main_fn_ident = &main_fn.sig.ident;
+    let asm_impl = asm_impl();
+
+    quote! {
+        #main_fn
+
+        #[naked]
+        #[no_mangle]
+        unsafe extern "C" fn _start() -> ! {
+            unsafe fn entry(mem: *mut usize) -> ! {
+                let (argc, argv, envp) = origin::program::compute_args(mem);
+                origin::program::init_runtime(mem, envp);
+                origin::program::exit(#main_fn_ident())
+            }
+
+            #asm_impl
+        }
+    }
+    .into()
+}
+
+/// Provides the asm implementation to start the program
+fn asm_impl() -> TokenStream2 {
+    quote! {
+        // Jump to `entry`, passing it the initial stack pointer value as an
+        // argument, a null return address, a null frame pointer, and an aligned
+        // stack pointer. On many architectures, the incoming frame pointer is
+        // already null.
+
+        #[cfg(target_arch = "x86_64")]
+        core::arch::asm!(
+            "mov rdi, rsp", // Pass the incoming `rsp` as the arg to `entry`.
+            "push rbp",     // Set the return address to zero.
+            "jmp {entry}",  // Jump to `entry`.
+            entry = sym entry,
+            options(noreturn),
+        );
+
+        #[cfg(target_arch = "aarch64")]
+        core::arch::asm!(
+            "mov x0, sp",   // Pass the incoming `sp` as the arg to `entry`.
+            "mov x30, xzr", // Set the return address to zero.
+            "b {entry}",    // Jump to `entry`.
+            entry = sym entry,
+            options(noreturn),
+        );
+
+        #[cfg(target_arch = "arm")]
+        core::arch::asm!(
+            "mov r0, sp\n", // Pass the incoming `sp` as the arg to `entry`.
+            "mov lr, #0",   // Set the return address to zero.
+            "b {entry}",    // Jump to `entry`.
+            entry = sym entry,
+            options(noreturn),
+        );
+
+        #[cfg(target_arch = "riscv64")]
+        core::arch::asm!(
+            "mv a0, sp",    // Pass the incoming `sp` as the arg to `entry`.
+            "mv ra, zero",  // Set the return address to zero.
+            "mv fp, zero",  // Set the frame address to zero.
+            "tail {entry}", // Jump to `entry`.
+            entry = sym entry,
+            options(noreturn),
+        );
+
+        #[cfg(target_arch = "x86")]
+        core::arch::asm!(
+            "mov eax, esp", // Save the incoming `esp` value.
+            "push ebp",     // Pad for alignment.
+            "push ebp",     // Pad for alignment.
+            "push ebp",     // Pad for alignment.
+            "push eax",     // Pass saved the incoming `esp` as the arg to `entry`.
+            "push ebp",     // Set the return address to zero.
+            "jmp {entry}",  // Jump to `entry`.
+            entry = sym entry,
+            options(noreturn),
+        );
+    }
+}

src/lib.rs

@@ -10,6 +10,9 @@
 #![deny(lossy_provenance_casts)]
 #![no_std]
 
+#![allow(dead_code)] // FIXME: this is just so tests pass while I implement everything
+
+
 #[cfg(all(feature = "alloc", not(feature = "rustc-dep-of-std")))]
 extern crate alloc;
 
@@ -40,76 +43,8 @@ pub mod thread;
 #[cfg_attr(target_arch = "arm", path = "arch/arm.rs")]
 mod arch;
 
-/// The program entry point.
-///
-/// # Safety
-///
-/// This function should never be called explicitly. It is the first thing
-/// executed in the program, and it assumes that memory is laid out according
-/// to the operating system convention for starting a new program.
-#[cfg(feature = "origin-program")]
-#[cfg(feature = "origin-start")]
-#[naked]
-#[no_mangle]
-unsafe extern "C" fn _start() -> ! {
-    use core::arch::asm;
-    use program::entry;
-
-    // Jump to `entry`, passing it the initial stack pointer value as an
-    // argument, a null return address, a null frame pointer, and an aligned
-    // stack pointer. On many architectures, the incoming frame pointer is
-    // already null.
-
-    #[cfg(target_arch = "x86_64")]
-    asm!(
-        "mov rdi, rsp", // Pass the incoming `rsp` as the arg to `entry`.
-        "push rbp",     // Set the return address to zero.
-        "jmp {entry}",  // Jump to `entry`.
-        entry = sym entry,
-        options(noreturn),
-    );
-
-    #[cfg(target_arch = "aarch64")]
-    asm!(
-        "mov x0, sp",   // Pass the incoming `sp` as the arg to `entry`.
-        "mov x30, xzr", // Set the return address to zero.
-        "b {entry}",    // Jump to `entry`.
-        entry = sym entry,
-        options(noreturn),
-    );
-
-    #[cfg(target_arch = "arm")]
-    asm!(
-        "mov r0, sp\n", // Pass the incoming `sp` as the arg to `entry`.
-        "mov lr, #0",   // Set the return address to zero.
-        "b {entry}",    // Jump to `entry`.
-        entry = sym entry,
-        options(noreturn),
-    );
-
-    #[cfg(target_arch = "riscv64")]
-    asm!(
-        "mv a0, sp",    // Pass the incoming `sp` as the arg to `entry`.
-        "mv ra, zero",  // Set the return address to zero.
-        "mv fp, zero",  // Set the frame address to zero.
-        "tail {entry}", // Jump to `entry`.
-        entry = sym entry,
-        options(noreturn),
-    );
-
-    #[cfg(target_arch = "x86")]
-    asm!(
-        "mov eax, esp", // Save the incoming `esp` value.
-        "push ebp",     // Pad for alignment.
-        "push ebp",     // Pad for alignment.
-        "push ebp",     // Pad for alignment.
-        "push eax",     // Pass saved the incoming `esp` as the arg to `entry`.
-        "push ebp",     // Set the return address to zero.
-        "jmp {entry}",  // Jump to `entry`.
-        entry = sym entry,
-        options(noreturn),
-    );
-}
+#[cfg(feature = "macros")]
+pub use origin_macros::main;
 
 /// A program entry point similar to `_start`, but which is meant to be called
 /// by something else in the program rather than the OS.