Separate compilation and linker phases (#376)

Separate between compilation and linker phases to allow deploy time linking and back-porting era compiler changes to fix #91. Unlinked contract binaries (caused by missing libraries or missing factory dependencies in turn) are emitted as raw ELF object. Few drive by fixes: - #98 - A compiler panic on missing libraries definitions. - Fixes some incosistent type forwarding in JSON output (empty string vs. null object). - Remove the unused fallback for size optimization setting. - Remove the broken `--lvm-ir` mode. - CI workflow fixes. --------- Signed-off-by: Cyrill Leutwiler <bigcyrill@hotmail.com> Signed-off-by: xermicus <bigcyrill@hotmail.com> Signed-off-by: xermicus <cyrill@parity.io>
2026-06-11 10:21:01 +00:00 · 2025-09-27 20:52:22 +02:00
parent 13faedf08a
commit 94ec34c4d5
169 changed files with 6288 additions and 5206 deletions
@@ -0,0 +1,114 @@
+//! The revive ELF object linker library.
+
+use std::{ffi::CString, fs, path::PathBuf, sync::Mutex};
+
+use lld_sys::LLDELFLink;
+use tempfile::TempDir;
+
+use revive_builtins::COMPILER_RT;
+
+static GUARD: Mutex<()> = Mutex::new(());
+
+/// The revive ELF object linker.
+pub struct ElfLinker {
+    temporary_directory: TempDir,
+    output_path: PathBuf,
+    object_path: PathBuf,
+    symbols_path: PathBuf,
+    linker_script_path: PathBuf,
+}
+
+impl ElfLinker {
+    const LINKER_SCRIPT: &str = r#"
+SECTIONS {
+    .text : { KEEP(*(.text.polkavm_export)) *(.text .text.*) }
+}"#;
+
+    const BUILTINS_ARCHIVE_FILE: &str = "libclang_rt.builtins-riscv64.a";
+    const BUILTINS_LIB_NAME: &str = "clang_rt.builtins-riscv64";
+
+    /// The setup routine prepares a temporary working directory.
+    pub fn setup() -> anyhow::Result<Self> {
+        let temporary_directory = TempDir::new()?;
+        let object_path = temporary_directory.path().join("obj.o");
+        let output_path = temporary_directory.path().join("out.o");
+        let symbols_path = temporary_directory.path().join("sym.o");
+        let linker_script_path = temporary_directory.path().join("linker.ld");
+
+        fs::write(&linker_script_path, Self::LINKER_SCRIPT)
+            .map_err(|message| anyhow::anyhow!("{message} {linker_script_path:?}",))?;
+
+        let compiler_rt_path = temporary_directory.path().join(Self::BUILTINS_ARCHIVE_FILE);
+        fs::write(&compiler_rt_path, COMPILER_RT)
+            .map_err(|message| anyhow::anyhow!("{message} {compiler_rt_path:?}"))?;
+
+        Ok(Self {
+            temporary_directory,
+            output_path,
+            object_path,
+            symbols_path,
+            linker_script_path,
+        })
+    }
+
+    /// Link `input` with `symbols` and the `compiler_rt` via `LLD`.
+    pub fn link<T: AsRef<[u8]>>(self, input: T, symbols: T) -> anyhow::Result<Vec<u8>> {
+        fs::write(&self.object_path, input)
+            .map_err(|message| anyhow::anyhow!("{message} {:?}", self.object_path))?;
+
+        fs::write(&self.symbols_path, symbols)
+            .map_err(|message| anyhow::anyhow!("{message} {:?}", self.symbols_path))?;
+
+        if lld(self
+            .create_arguments()
+            .into_iter()
+            .map(|v| v.to_string())
+            .collect())
+        {
+            return Err(anyhow::anyhow!("ld.lld failed"));
+        }
+
+        Ok(fs::read(&self.output_path)?)
+    }
+
+    /// The argument creation helper function.
+    fn create_arguments(&self) -> Vec<String> {
+        [
+            "ld.lld",
+            "--error-limit=0",
+            "--relocatable",
+            "--emit-relocs",
+            "--no-relax",
+            "--unique",
+            "--gc-sections",
+            self.linker_script_path.to_str().expect("should be utf8"),
+            "-o",
+            self.output_path.to_str().expect("should be utf8"),
+            self.object_path.to_str().expect("should be utf8"),
+            self.symbols_path.to_str().expect("should be utf8"),
+            "--library-path",
+            self.temporary_directory
+                .path()
+                .to_str()
+                .expect("should be utf8"),
+            "--library",
+            Self::BUILTINS_LIB_NAME,
+        ]
+        .iter()
+        .map(ToString::to_string)
+        .collect()
+    }
+}
+
+/// The thread-safe LLD helper function.
+fn lld(arguments: Vec<String>) -> bool {
+    let c_strings = arguments
+        .into_iter()
+        .map(|arg| CString::new(arg).expect("ld.lld args should not contain null bytes"))
+        .collect::<Vec<_>>();
+
+    let args: Vec<*const libc::c_char> = c_strings.iter().map(|arg| arg.as_ptr()).collect();
+
+    let _lock = GUARD.lock().expect("ICE: linker mutex should not poison");
+    unsafe { LLDELFLink(args.as_ptr(), args.len()) == 0 }
+}
@@ -1,76 +1,4 @@
-use std::{env, ffi::CString, fs};
+//! The revive ELF object to PVM blob linker library.

-use lld_sys::LLDELFLink;
-use revive_builtins::COMPILER_RT;
-
-const LINKER_SCRIPT: &str = r#"
-SECTIONS {
-    .text : { KEEP(*(.text.polkavm_export)) *(.text .text.*) }
-}"#;
-
-const BUILTINS_ARCHIVE_FILE: &str = "libclang_rt.builtins-riscv64.a";
-const BUILTINS_LIB_NAME: &str = "clang_rt.builtins-riscv64";
-
-fn invoke_lld(cmd_args: &[&str]) -> bool {
-    let c_strings = cmd_args
-        .iter()
-        .map(|arg| CString::new(*arg).expect("ld.lld args should not contain null bytes"))
-        .collect::<Vec<_>>();
-
-    let args: Vec<*const libc::c_char> = c_strings.iter().map(|arg| arg.as_ptr()).collect();
-
-    unsafe { LLDELFLink(args.as_ptr(), args.len()) == 0 }
-}
-
-pub fn polkavm_linker<T: AsRef<[u8]>>(code: T, strip_binary: bool) -> anyhow::Result<Vec<u8>> {
-    let mut config = polkavm_linker::Config::default();
-    config.set_strip(strip_binary);
-    config.set_optimize(true);
-
-    polkavm_linker::program_from_elf(config, code.as_ref())
-        .map_err(|reason| anyhow::anyhow!("polkavm linker failed: {}", reason))
-}
-
-pub fn link<T: AsRef<[u8]>>(input: T) -> anyhow::Result<Vec<u8>> {
-    let dir = tempfile::tempdir().expect("failed to create temp directory for linking");
-    let output_path = dir.path().join("out.so");
-    let object_path = dir.path().join("out.o");
-    let linker_script_path = dir.path().join("linker.ld");
-    let compiler_rt_path = dir.path().join(BUILTINS_ARCHIVE_FILE);
-
-    fs::write(&object_path, input).map_err(|msg| anyhow::anyhow!("{msg} {object_path:?}"))?;
-
-    if env::var("PVM_LINKER_DUMP_OBJ").is_ok() {
-        fs::copy(&object_path, "/tmp/out.o")?;
-    }
-
-    fs::write(&linker_script_path, LINKER_SCRIPT)
-        .map_err(|msg| anyhow::anyhow!("{msg} {linker_script_path:?}"))?;
-
-    fs::write(&compiler_rt_path, COMPILER_RT)
-        .map_err(|msg| anyhow::anyhow!("{msg} {compiler_rt_path:?}"))?;
-
-    let ld_args = [
-        "ld.lld",
-        "--error-limit=0",
-        "--relocatable",
-        "--emit-relocs",
-        "--no-relax",
-        "--unique",
-        "--gc-sections",
-        "--library-path",
-        dir.path().to_str().expect("should be utf8"),
-        "--library",
-        BUILTINS_LIB_NAME,
-        linker_script_path.to_str().expect("should be utf8"),
-        object_path.to_str().expect("should be utf8"),
-        "-o",
-        output_path.to_str().expect("should be utf8"),
-    ];
-
-    if invoke_lld(&ld_args) {
-        return Err(anyhow::anyhow!("ld.lld failed"));
-    }
-
-    Ok(fs::read(&output_path)?)
-}
+pub mod elf;
+pub mod pvm;
@@ -0,0 +1,10 @@
+//! The revive PVM blob linker library.
+
+pub fn polkavm_linker<T: AsRef<[u8]>>(code: T, strip_binary: bool) -> anyhow::Result<Vec<u8>> {
+    let mut config = polkavm_linker::Config::default();
+    config.set_strip(strip_binary);
+    config.set_optimize(true);
+
+    polkavm_linker::program_from_elf(config, code.as_ref())
+        .map_err(|reason| anyhow::anyhow!("polkavm linker failed: {}", reason))
+}