cg: use thorin instead of llvm-dwp

`thorin` is a Rust implementation of a DWARF packaging utility that
supports reading DWARF objects from archive files (i.e. rlibs) and
therefore is better suited for integration into rustc.

Signed-off-by: David Wood <david.wood@huawei.com>

Cargo.lock

@@ -9,7 +9,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "3e61f2b7f93d2c7d2b08263acaa4a363b3e276806c68af6134c44f523bf1aacd"
 dependencies = [
  "compiler_builtins",
- "gimli",
+ "gimli 0.25.0",
  "rustc-std-workspace-alloc",
  "rustc-std-workspace-core",
 ]
@@ -87,9 +87,9 @@ dependencies = [
 
 [[package]]
 name = "anyhow"
-version = "1.0.34"
+version = "1.0.51"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bf8dcb5b4bbaa28653b647d8c77bd4ed40183b48882e130c1f1ffb73de069fd7"
+checksum = "8b26702f315f53b6071259e15dd9d64528213b44d61de1ec926eca7715d62203"
 
 [[package]]
 name = "array_tool"
@@ -1158,6 +1158,12 @@ version = "0.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e88a8acf291dafb59c2d96e8f59828f3838bb1a70398823ade51a84de6a6deed"
 
+[[package]]
+name = "fallible-iterator"
+version = "0.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4443176a9f2c162692bd3d352d745ef9413eec5782a80d8fd6f8a1ac692a07f7"
+
 [[package]]
 name = "filetime"
 version = "0.2.14"
@@ -1446,6 +1452,17 @@ dependencies = [
  "rustc-std-workspace-core",
 ]
 
+[[package]]
+name = "gimli"
+version = "0.26.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "78cc372d058dcf6d5ecd98510e7fbc9e5aec4d21de70f65fea8fecebcd881bd4"
+dependencies = [
+ "fallible-iterator",
+ "indexmap",
+ "stable_deref_trait",
+]
+
 [[package]]
 name = "git2"
 version = "0.13.23"
@@ -2339,6 +2356,18 @@ dependencies = [
  "rustc-std-workspace-core",
 ]
 
+[[package]]
+name = "object"
+version = "0.27.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "67ac1d3f9a1d3616fd9a60c8d74296f22406a238b6a72f5cc1e6f314df4ffbf9"
+dependencies = [
+ "crc32fast",
+ "flate2",
+ "indexmap",
+ "memchr",
+]
+
 [[package]]
 name = "odht"
 version = "0.3.1"
@@ -3725,10 +3754,11 @@ dependencies = [
  "itertools 0.9.0",
  "jobserver",
  "libc",
- "object",
+ "object 0.26.2",
  "pathdiff",
  "regex",
  "rustc_apfloat",
+ "rustc_arena",
  "rustc_ast",
  "rustc_attr",
  "rustc_data_structures",
@@ -3749,6 +3779,7 @@ dependencies = [
  "smallvec",
  "snap",
  "tempfile",
+ "thorin-dwp",
  "tracing",
 ]
 
@@ -4993,7 +5024,7 @@ dependencies = [
  "hermit-abi",
  "libc",
  "miniz_oxide",
- "object",
+ "object 0.26.2",
  "panic_abort",
  "panic_unwind",
  "profiler_builtins",
@@ -5057,9 +5088,9 @@ checksum = "8ea5119cdb4c55b55d432abb513a0429384878c15dde60cc77b1c99de1a95a6a"
 
 [[package]]
 name = "structopt"
-version = "0.3.16"
+version = "0.3.25"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "de5472fb24d7e80ae84a7801b7978f95a19ec32cb1876faea59ab711eb901976"
+checksum = "40b9788f4202aa75c240ecc9c15c65185e6a39ccdeb0fd5d008b98825464c87c"
 dependencies = [
  "clap",
  "lazy_static",
@@ -5068,9 +5099,9 @@ dependencies = [
 
 [[package]]
 name = "structopt-derive"
-version = "0.4.9"
+version = "0.4.18"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1e0eb37335aeeebe51be42e2dc07f031163fbabfa6ac67d7ea68b5c2f68d5f99"
+checksum = "dcb5ae327f9cc13b68763b5749770cb9e048a99bd9dfdfa58d0cf05d5f64afe0"
 dependencies = [
  "heck",
  "proc-macro-error",
@@ -5249,24 +5280,36 @@ dependencies = [
 
 [[package]]
 name = "thiserror"
-version = "1.0.20"
+version = "1.0.30"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7dfdd070ccd8ccb78f4ad66bf1982dc37f620ef696c6b5028fe2ed83dd3d0d08"
+checksum = "854babe52e4df1653706b98fcfc05843010039b406875930a70e4d9644e5c417"
 dependencies = [
  "thiserror-impl",
 ]
 
 [[package]]
 name = "thiserror-impl"
-version = "1.0.20"
+version = "1.0.30"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bd80fc12f73063ac132ac92aceea36734f04a1d93c1240c6944e23a3b8841793"
+checksum = "aa32fd3f627f367fe16f893e2597ae3c05020f8bba2666a4e6ea73d377e5714b"
 dependencies = [
  "proc-macro2",
  "quote",
  "syn",
 ]
 
+[[package]]
+name = "thorin-dwp"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "039d1fc0bfdb73910c2702893515580e38c192f47a987bc98ddd38a36f2d953a"
+dependencies = [
+ "gimli 0.26.1",
+ "indexmap",
+ "object 0.27.1",
+ "tracing",
+]
+
 [[package]]
 name = "thread_local"
 version = "1.0.1"
@@ -5394,9 +5437,9 @@ checksum = "360dfd1d6d30e05fda32ace2c8c70e9c0a9da713275777f5a4dbb8a1893930c6"
 
 [[package]]
 name = "tracing"
-version = "0.1.28"
+version = "0.1.29"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "84f96e095c0c82419687c20ddf5cb3eadb61f4e1405923c9dc8e53a1adacbda8"
+checksum = "375a639232caf30edfc78e8d89b2d4c375515393e7af7e16f01cd96917fb2105"
 dependencies = [
  "cfg-if 1.0.0",
  "pin-project-lite",
@@ -5406,9 +5449,9 @@ dependencies = [
 
 [[package]]
 name = "tracing-attributes"
-version = "0.1.17"
+version = "0.1.18"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c4f915eb6abf914599c200260efced9203504c4c37380af10cdf3b7d36970650"
+checksum = "f4f480b8f81512e825f337ad51e94c1eb5d3bbdf2b363dcd01e2b19a9ffe3f8e"
 dependencies = [
  "proc-macro2",
  "quote",

compiler/rustc_codegen_ssa/Cargo.toml

@@ -14,12 +14,14 @@ tracing = "0.1"
 libc = "0.2.50"
 jobserver = "0.1.22"
 tempfile = "3.2"
+thorin-dwp = "0.1.1"
 pathdiff = "0.2.0"
 snap = "1"
 smallvec = { version = "1.6.1", features = ["union", "may_dangle"] }
 regex = "1.4"
 
 rustc_serialize = { path = "../rustc_serialize" }
+rustc_arena = { path = "../rustc_arena" }
 rustc_ast = { path = "../rustc_ast" }
 rustc_span = { path = "../rustc_span" }
 rustc_middle = { path = "../rustc_middle" }

compiler/rustc_codegen_ssa/src/back/link.rs

@@ -1,4 +1,6 @@
+use rustc_arena::TypedArena;
 use rustc_data_structures::fx::{FxHashSet, FxIndexMap};
+use rustc_data_structures::memmap::Mmap;
 use rustc_data_structures::temp_dir::MaybeTempDir;
 use rustc_errors::{ErrorReported, Handler};
 use rustc_fs_util::fix_windows_verbatim_for_gcc;
@@ -32,7 +34,10 @@ use cc::windows_registry;
 use regex::Regex;
 use tempfile::Builder as TempFileBuilder;
 
-use std::ffi::{OsStr, OsString};
+use std::borrow::Borrow;
+use std::ffi::OsString;
+use std::fs::{File, OpenOptions};
+use std::io::{BufWriter, Write};
 use std::lazy::OnceCell;
 use std::path::{Path, PathBuf};
 use std::process::{ExitStatus, Output, Stdio};
@@ -261,10 +266,16 @@ fn link_rlib<'a, B: ArchiveBuilder<'a>>(
 
     let mut ab = <B as ArchiveBuilder>::new(sess, out_filename, None);
 
-    for obj in codegen_results.modules.iter().filter_map(|m| m.object.as_ref()) {
+    for m in &codegen_results.modules {
+        if let Some(obj) = m.object.as_ref() {
             ab.add_file(obj);
         }
 
+        if let Some(dwarf_obj) = m.dwarf_object.as_ref() {
+            ab.add_file(dwarf_obj);
+        }
+    }
+
     // Note that in this loop we are ignoring the value of `lib.cfg`. That is,
     // we may not be configured to actually include a static library if we're
     // adding it here. That's because later when we consume this rlib we'll
@@ -518,59 +529,108 @@ fn escape_stdout_stderr_string(s: &[u8]) -> String {
     })
 }
 
-const LLVM_DWP_EXECUTABLE: &'static str = "rust-llvm-dwp";
+/// Use `thorin` (rust implementation of a dwarf packaging utility) to link DWARF objects into a
-
+/// DWARF package.
-/// Invoke `llvm-dwp` (shipped alongside rustc) to link debuginfo in object files into a `dwp`
+fn link_dwarf_object<'a>(
-/// file.
+    sess: &'a Session,
-fn link_dwarf_object<'a, I>(sess: &'a Session, executable_out_filename: &Path, object_files: I)
+    cg_results: &CodegenResults,
-where
+    executable_out_filename: &Path,
-    I: IntoIterator<Item: AsRef<OsStr>>,
+) {
-{
-    info!("preparing dwp to {}.dwp", executable_out_filename.to_str().unwrap());
-
     let dwp_out_filename = executable_out_filename.with_extension("dwp");
-    let mut cmd = Command::new(LLVM_DWP_EXECUTABLE);
+    debug!(?dwp_out_filename, ?executable_out_filename);
-    cmd.arg("-o");
-    cmd.arg(&dwp_out_filename);
-    cmd.args(object_files);
 
-    let mut new_path = sess.get_tools_search_paths(false);
+    #[derive(Default)]
-    if let Some(path) = env::var_os("PATH") {
+    struct ThorinSession<Relocations> {
-        new_path.extend(env::split_paths(&path));
+        arena_data: TypedArena<Vec<u8>>,
+        arena_mmap: TypedArena<Mmap>,
+        arena_relocations: TypedArena<Relocations>,
     }
-    let new_path = env::join_paths(new_path).unwrap();
-    cmd.env("PATH", new_path);
 
-    info!("{:?}", &cmd);
+    impl<Relocations> ThorinSession<Relocations> {
-    match sess.time("run_dwp", || cmd.output()) {
+        fn alloc_mmap<'arena>(&'arena self, data: Mmap) -> &'arena Mmap {
-        Ok(prog) if !prog.status.success() => {
+            (*self.arena_mmap.alloc(data)).borrow()
-            sess.struct_err(&format!(
-                "linking dwarf objects with `{}` failed: {}",
-                LLVM_DWP_EXECUTABLE, prog.status
-            ))
-            .note(&format!("{:?}", &cmd))
-            .note(&escape_stdout_stderr_string(&prog.stdout))
-            .note(&escape_stdout_stderr_string(&prog.stderr))
-            .emit();
-            info!("linker stderr:\n{}", escape_stdout_stderr_string(&prog.stderr));
-            info!("linker stdout:\n{}", escape_stdout_stderr_string(&prog.stdout));
         }
-        Ok(_) => {}
+    }
+
+    impl<Relocations> thorin::Session<Relocations> for ThorinSession<Relocations> {
+        fn alloc_data<'arena>(&'arena self, data: Vec<u8>) -> &'arena [u8] {
+            (*self.arena_data.alloc(data)).borrow()
+        }
+
+        fn alloc_relocation<'arena>(&'arena self, data: Relocations) -> &'arena Relocations {
+            (*self.arena_relocations.alloc(data)).borrow()
+        }
+
+        fn read_input<'arena>(&'arena self, path: &Path) -> std::io::Result<&'arena [u8]> {
+            let file = File::open(&path)?;
+            let mmap = (unsafe { Mmap::map(file) })?;
+            Ok(self.alloc_mmap(mmap))
+        }
+    }
+
+    match sess.time("run_thorin", || -> Result<(), thorin::Error> {
+        let thorin_sess = ThorinSession::default();
+        let mut package = thorin::DwarfPackage::new(&thorin_sess);
+
+        // Input objs contain .o/.dwo files from the current crate.
+        match sess.opts.debugging_opts.split_dwarf_kind {
+            SplitDwarfKind::Single => {
+                for input_obj in cg_results.modules.iter().filter_map(|m| m.object.as_ref()) {
+                    package.add_input_object(input_obj)?;
+                }
+            }
+            SplitDwarfKind::Split => {
+                for input_obj in cg_results.modules.iter().filter_map(|m| m.dwarf_object.as_ref()) {
+                    package.add_input_object(input_obj)?;
+                }
+            }
+        }
+
+        // Input rlibs contain .o/.dwo files from dependencies.
+        let input_rlibs = cg_results
+            .crate_info
+            .used_crate_source
+            .values()
+            .filter_map(|csource| csource.rlib.as_ref())
+            .map(|(path, _)| path);
+        for input_rlib in input_rlibs {
+            debug!(?input_rlib);
+            package.add_input_object(input_rlib)?;
+        }
+
+        // Failing to read the referenced objects is expected for dependencies where the path in the
+        // executable will have been cleaned by Cargo, but the referenced objects will be contained
+        // within rlibs provided as inputs.
+        //
+        // If paths have been remapped, then .o/.dwo files from the current crate also won't be
+        // found, but are provided explicitly above.
+        //
+        // Adding an executable is primarily done to make `thorin` check that all the referenced
+        // dwarf objects are found in the end.
+        package.add_executable(
+            &executable_out_filename,
+            thorin::MissingReferencedObjectBehaviour::Skip,
+        )?;
+
+        let output = package.finish()?.write()?;
+        let mut output_stream = BufWriter::new(
+            OpenOptions::new()
+                .read(true)
+                .write(true)
+                .create(true)
+                .truncate(true)
+                .open(dwp_out_filename)?,
+        );
+        output_stream.write_all(&output)?;
+        output_stream.flush()?;
+
+        Ok(())
+    }) {
+        Ok(()) => {}
         Err(e) => {
-            let dwp_not_found = e.kind() == io::ErrorKind::NotFound;
+            sess.struct_err("linking dwarf objects with thorin failed")
-            let mut err = if dwp_not_found {
+                .note(&format!("{:?}", e))
-                sess.struct_err(&format!("linker `{}` not found", LLVM_DWP_EXECUTABLE))
+                .emit();
-            } else {
-                sess.struct_err(&format!("could not exec the linker `{}`", LLVM_DWP_EXECUTABLE))
-            };
-
-            err.note(&e.to_string());
-
-            if !dwp_not_found {
-                err.note(&format!("{:?}", &cmd));
-            }
-
-            err.emit();
         }
     }
 }
@@ -916,14 +976,11 @@ fn link_natively<'a, B: ArchiveBuilder<'a>>(
         SplitDebuginfo::Packed if sess.target.is_like_msvc => {}
 
         // ... and otherwise we're processing a `*.dwp` packed dwarf file.
+        //
         // We cannot rely on the .o paths in the exectuable because they may have been
-        // remapped by --remap-path-prefix and therefore invalid. So we need to provide
+        // remapped by --remap-path-prefix and therefore invalid, so we need to provide
-        // the .o paths explicitly
+        // the .o/.dwo paths explicitly.
-        SplitDebuginfo::Packed => link_dwarf_object(
+        SplitDebuginfo::Packed => link_dwarf_object(sess, codegen_results, out_filename),
-            sess,
-            &out_filename,
-            codegen_results.modules.iter().filter_map(|m| m.object.as_ref()),
-        ),
     }
 
     let strip = strip_value(sess);

src/test/run-make-fulldeps/split-debuginfo/Makefile

@@ -57,7 +57,45 @@ packed-single:
 	ls $(TMPDIR)/*.dwo && exit 1 || exit 0
 	rm -rf $(TMPDIR)/*.dwp
 
-unpacked: unpacked-split unpacked-single
+packed-remapped: packed-remapped-split packed-remapped-single
+
+packed-remapped-split:
+	$(RUSTC) -Z unstable-options -C split-debuginfo=packed -C debuginfo=2 \
+		-Z split-dwarf-kind=split --remap-path-prefix $(TMPDIR)=/a foo.rs -g
+	objdump -Wi $(TMPDIR)/foo | grep DW_AT_GNU_dwo_name | (! grep $(TMPDIR)) || exit 1
+
+packed-remapped-single:
+	$(RUSTC) -Z unstable-options -C split-debuginfo=packed -C debuginfo=2 \
+		-Z split-dwarf-kind=single --remap-path-prefix $(TMPDIR)=/a foo.rs -g
+	objdump -Wi $(TMPDIR)/foo | grep DW_AT_GNU_dwo_name | (! grep $(TMPDIR)) || exit 1
+
+packed-crosscrate: packed-crosscrate-split packed-crosscrate-single
+
+packed-crosscrate-split:
+	$(RUSTC) --crate-type lib -Z unstable-options -C split-debuginfo=packed \
+		-Zsplit-dwarf-kind=split -C debuginfo=2 -g bar.rs
+	ls $(TMPDIR)/*.rlib
+	ls $(TMPDIR)/*.dwo && exit 1 || exit 0
+	ls $(TMPDIR)/*.dwp && exit 1 || exit 0
+	$(RUSTC) --extern bar=$(TMPDIR)/libbar.rlib -Z unstable-options -C split-debuginfo=packed \
+		-Zsplit-dwarf-kind=split -C debuginfo=2 -g main.rs
+	rm $(TMPDIR)/*.dwo
+	rm $(TMPDIR)/main.dwp
+	rm $(TMPDIR)/$(call BIN,main)
+
+packed-crosscrate-single:
+	$(RUSTC) --crate-type lib -Z unstable-options -C split-debuginfo=packed \
+		-Zsplit-dwarf-kind=single -C debuginfo=2 -g bar.rs
+	ls $(TMPDIR)/*.rlib
+	ls $(TMPDIR)/*.dwo && exit 1 || exit 0
+	ls $(TMPDIR)/*.dwp && exit 1 || exit 0
+	$(RUSTC) --extern bar=$(TMPDIR)/libbar.rlib -Z unstable-options -C split-debuginfo=packed \
+		-Zsplit-dwarf-kind=single -C debuginfo=2 -g main.rs
+	ls $(TMPDIR)/*.dwo && exit 1 || exit 0
+	rm $(TMPDIR)/main.dwp
+	rm $(TMPDIR)/$(call BIN,main)
+
+unpacked: unpacked-split unpacked-single unpacked-remapped-split unpacked-remapped-single
 
 unpacked-split:
 	$(RUSTC) foo.rs -g -C split-debuginfo=unpacked -Z unstable-options -Zsplit-dwarf-kind=split
@@ -69,5 +107,15 @@ unpacked-single:
 	$(RUSTC) foo.rs -g -C split-debuginfo=unpacked -Z unstable-options -Zsplit-dwarf-kind=single
 	ls $(TMPDIR)/*.dwp && exit 1 || exit 0
 	ls $(TMPDIR)/*.dwo && exit 1 || exit 0
+
+unpacked-remapped-split:
+	$(RUSTC) -Z unstable-options -C split-debuginfo=unpacked -C debuginfo=2 \
+		-Z split-dwarf-kind=split --remap-path-prefix $(TMPDIR)=/a foo.rs -g
+	objdump -Wi $(TMPDIR)/foo | grep DW_AT_GNU_dwo_name | (! grep $(TMPDIR)) || exit 1
+
+unpacked-remapped-single:
+	$(RUSTC) -Z unstable-options -C split-debuginfo=unpacked -C debuginfo=2 \
+		-Z split-dwarf-kind=single --remap-path-prefix $(TMPDIR)=/a foo.rs -g
+	objdump -Wi $(TMPDIR)/foo | grep DW_AT_GNU_dwo_name | (! grep $(TMPDIR)) || exit 1
 endif
 endif

src/test/run-make-fulldeps/split-debuginfo/bar.rs

@@ -0,0 +1,13 @@
+pub struct Bar {
+    x: u32,
+}
+
+impl Bar {
+    pub fn print(&self) {
+        println!("{}", self.x);
+    }
+}
+
+pub fn make_bar(x: u32) -> Bar {
+    Bar { x }
+}

src/test/run-make-fulldeps/split-debuginfo/foo.rs

@@ -1 +1,15 @@
+pub struct Foo {
+    x: u32,
+}
+
+impl Foo {
+    pub fn print(&self) {
+        println!("{}", self.x);
+    }
+}
+
+pub fn make_foo(x: u32) -> Foo {
+    Foo { x }
+}
+
 fn main() {}

src/test/run-make-fulldeps/split-debuginfo/main.rs

@@ -0,0 +1,8 @@
+extern crate bar;
+
+use bar::{Bar, make_bar};
+
+fn main() {
+    let b = make_bar(3);
+    b.print();
+}

src/test/run-make-fulldeps/split-dwarf/Makefile

@@ -1,17 +0,0 @@
--include ../tools.mk
-
-# only-linux
-
-all: packed remapped
-
-remapped:
-	$(RUSTC) -Z unstable-options -C split-debuginfo=packed -C debuginfo=2 --remap-path-prefix $(TMPDIR)=/a foo.rs -g
-	objdump -Wi $(TMPDIR)/foo | grep DW_AT_GNU_dwo_name | (! grep $(TMPDIR)) || exit 1
-
-	$(RUSTC) -Z unstable-options -C split-debuginfo=unpacked -C debuginfo=2 --remap-path-prefix $(TMPDIR)=/a foo.rs -g
-	objdump -Wi $(TMPDIR)/foo | grep DW_AT_GNU_dwo_name | (! grep $(TMPDIR)) || exit 1
-
-packed:
-	$(RUSTC) -Z unstable-options -C split-debuginfo=packed -C debuginfo=2 foo.rs -g
-	rm $(TMPDIR)/foo.dwp
-	rm $(TMPDIR)/$(call BIN,foo)

src/test/run-make-fulldeps/split-dwarf/foo.rs

@@ -1 +0,0 @@
-fn main() {}

src/tools/tidy/src/deps.rs

@@ -109,6 +109,7 @@ const PERMITTED_DEPENDENCIES: &[&str] = &[
     "env_logger",
     "expect-test",
     "fake-simd",
+    "fallible-iterator", // dependency of `thorin`
     "filetime",
     "fixedbitset",
     "flate2",
@@ -201,6 +202,7 @@ const PERMITTED_DEPENDENCIES: &[&str] = &[
     "tempfile",
     "termcolor",
     "termize",
+    "thorin-dwp",
     "thread_local",
     "time",
     "tinyvec",