cg_llvm: implement split dwarf support

This commit implements Split DWARF support, wiring up the flag (added in
earlier commits) to the modified FFI wrapper (also from earlier
commits).

Signed-off-by: David Wood <david@davidtw.co>

compiler/rustc_codegen_llvm/src/back/lto.rs

@@ -6,7 +6,9 @@ use crate::llvm::{self, build_string, False, True};
 use crate::{LlvmCodegenBackend, ModuleLlvm};
 use rustc_codegen_ssa::back::lto::{LtoModuleCodegen, SerializedModule, ThinModule, ThinShared};
 use rustc_codegen_ssa::back::symbol_export;
-use rustc_codegen_ssa::back::write::{CodegenContext, FatLTOInput, ModuleConfig};
+use rustc_codegen_ssa::back::write::{
+    CodegenContext, FatLTOInput, ModuleConfig, TargetMachineFactoryConfig,
+};
 use rustc_codegen_ssa::traits::*;
 use rustc_codegen_ssa::{looks_like_rust_object_file, ModuleCodegen, ModuleKind};
 use rustc_data_structures::fx::FxHashMap;
@@ -728,7 +730,14 @@ pub unsafe fn optimize_thin_module(
     cgcx: &CodegenContext<LlvmCodegenBackend>,
 ) -> Result<ModuleCodegen<ModuleLlvm>, FatalError> {
     let diag_handler = cgcx.create_diag_handler();
-    let tm = (cgcx.tm_factory)().map_err(|e| write::llvm_err(&diag_handler, &e))?;
+
+    let module_name = &thin_module.shared.module_names[thin_module.idx];
+    let split_dwarf_file = cgcx
+        .output_filenames
+        .split_dwarf_file(cgcx.split_dwarf_kind, Some(module_name.to_str().unwrap()));
+    let tm_factory_config = TargetMachineFactoryConfig { split_dwarf_file };
+    let tm =
+        (cgcx.tm_factory)(tm_factory_config).map_err(|e| write::llvm_err(&diag_handler, &e))?;
 
     // Right now the implementation we've got only works over serialized
     // modules, so we create a fresh new LLVM context and parse the module
@@ -736,12 +745,8 @@ pub unsafe fn optimize_thin_module(
     // crates but for locally codegened modules we may be able to reuse
     // that LLVM Context and Module.
     let llcx = llvm::LLVMRustContextCreate(cgcx.fewer_names);
-    let llmod_raw = parse_module(
+    let llmod_raw =
-        llcx,
+        parse_module(llcx, &module_name, thin_module.data(), &diag_handler)? as *const _;
-        &thin_module.shared.module_names[thin_module.idx],
-        thin_module.data(),
-        &diag_handler,
-    )? as *const _;
     let module = ModuleCodegen {
         module_llvm: ModuleLlvm { llmod_raw, llcx, tm },
         name: thin_module.name().to_string(),

compiler/rustc_codegen_llvm/src/back/write.rs

@@ -12,7 +12,8 @@ use crate::type_::Type;
 use crate::LlvmCodegenBackend;
 use crate::ModuleLlvm;
 use rustc_codegen_ssa::back::write::{
-    BitcodeSection, CodegenContext, EmitObj, ModuleConfig, TargetMachineFactoryFn,
+    BitcodeSection, CodegenContext, EmitObj, ModuleConfig, TargetMachineFactoryConfig,
+    TargetMachineFactoryFn,
 };
 use rustc_codegen_ssa::traits::*;
 use rustc_codegen_ssa::{CompiledModule, ModuleCodegen};
@@ -22,7 +23,9 @@ use rustc_fs_util::{link_or_copy, path_to_c_string};
 use rustc_hir::def_id::LOCAL_CRATE;
 use rustc_middle::bug;
 use rustc_middle::ty::TyCtxt;
-use rustc_session::config::{self, Lto, OutputType, Passes, SanitizerSet, SwitchWithOptPath};
+use rustc_session::config::{
+    self, Lto, OutputType, Passes, SanitizerSet, SplitDwarfKind, SwitchWithOptPath,
+};
 use rustc_session::Session;
 use rustc_span::symbol::sym;
 use rustc_span::InnerSpan;
@@ -51,18 +54,31 @@ pub fn write_output_file(
     pm: &llvm::PassManager<'ll>,
     m: &'ll llvm::Module,
     output: &Path,
+    dwo_output: Option<&Path>,
     file_type: llvm::FileType,
 ) -> Result<(), FatalError> {
     unsafe {
         let output_c = path_to_c_string(output);
-        let result = llvm::LLVMRustWriteOutputFile(
+        let result = if let Some(dwo_output) = dwo_output {
-            target,
+            let dwo_output_c = path_to_c_string(dwo_output);
-            pm,
+            llvm::LLVMRustWriteOutputFile(
-            m,
+                target,
-            output_c.as_ptr(),
+                pm,
-            std::ptr::null(),
+                m,
-            file_type,
+                output_c.as_ptr(),
-        );
+                dwo_output_c.as_ptr(),
+                file_type,
+            )
+        } else {
+            llvm::LLVMRustWriteOutputFile(
+                target,
+                pm,
+                m,
+                output_c.as_ptr(),
+                std::ptr::null(),
+                file_type,
+            )
+        };
         result.into_result().map_err(|()| {
             let msg = format!("could not write output to {}", output.display());
             llvm_err(handler, &msg)
@@ -71,12 +87,17 @@ pub fn write_output_file(
 }
 
 pub fn create_informational_target_machine(sess: &Session) -> &'static mut llvm::TargetMachine {
-    target_machine_factory(sess, config::OptLevel::No)()
+    let config = TargetMachineFactoryConfig { split_dwarf_file: None };
+    target_machine_factory(sess, config::OptLevel::No)(config)
         .unwrap_or_else(|err| llvm_err(sess.diagnostic(), &err).raise())
 }
 
-pub fn create_target_machine(tcx: TyCtxt<'_>) -> &'static mut llvm::TargetMachine {
+pub fn create_target_machine(tcx: TyCtxt<'_>, mod_name: &str) -> &'static mut llvm::TargetMachine {
-    target_machine_factory(&tcx.sess, tcx.backend_optimization_level(LOCAL_CRATE))()
+    let split_dwarf_file = tcx
+        .output_filenames(LOCAL_CRATE)
+        .split_dwarf_file(tcx.sess.opts.debugging_opts.split_dwarf, Some(mod_name));
+    let config = TargetMachineFactoryConfig { split_dwarf_file };
+    target_machine_factory(&tcx.sess, tcx.backend_optimization_level(LOCAL_CRATE))(config)
         .unwrap_or_else(|err| llvm_err(tcx.sess.diagnostic(), &err).raise())
 }
 
@@ -172,8 +193,10 @@ pub fn target_machine_factory(
     let use_init_array =
         !sess.opts.debugging_opts.use_ctors_section.unwrap_or(sess.target.use_ctors_section);
 
-    Arc::new(move || {
+    Arc::new(move |config: TargetMachineFactoryConfig| {
-        let split_dwarf_file = std::ptr::null();
+        let split_dwarf_file = config.split_dwarf_file.unwrap_or_default();
+        let split_dwarf_file = CString::new(split_dwarf_file.to_str().unwrap()).unwrap();
+
         let tm = unsafe {
             llvm::LLVMRustCreateTargetMachine(
                 triple.as_ptr(),
@@ -192,7 +215,7 @@ pub fn target_machine_factory(
                 emit_stack_size_section,
                 relax_elf_relocations,
                 use_init_array,
-                split_dwarf_file,
+                split_dwarf_file.as_ptr(),
             )
         };
 
@@ -796,7 +819,15 @@ pub(crate) unsafe fn codegen(
                 llmod
             };
             with_codegen(tm, llmod, config.no_builtins, |cpm| {
-                write_output_file(diag_handler, tm, cpm, llmod, &path, llvm::FileType::AssemblyFile)
+                write_output_file(
+                    diag_handler,
+                    tm,
+                    cpm,
+                    llmod,
+                    &path,
+                    None,
+                    llvm::FileType::AssemblyFile,
+                )
             })?;
         }
 
@@ -805,6 +836,15 @@ pub(crate) unsafe fn codegen(
                 let _timer = cgcx
                     .prof
                     .generic_activity_with_arg("LLVM_module_codegen_emit_obj", &module.name[..]);
+
+                let dwo_out = cgcx.output_filenames.temp_path_dwo(module_name);
+                let dwo_out = match cgcx.split_dwarf_kind {
+                    // Don't change how DWARF is emitted in single mode (or when disabled).
+                    SplitDwarfKind::None | SplitDwarfKind::Single => None,
+                    // Emit (a subset of the) DWARF into a separate file in split mode.
+                    SplitDwarfKind::Split => Some(dwo_out.as_path()),
+                };
+
                 with_codegen(tm, llmod, config.no_builtins, |cpm| {
                     write_output_file(
                         diag_handler,
@@ -812,6 +852,7 @@ pub(crate) unsafe fn codegen(
                         cpm,
                         llmod,
                         &obj_out,
+                        dwo_out,
                         llvm::FileType::ObjectFile,
                     )
                 })?;
@@ -839,6 +880,7 @@ pub(crate) unsafe fn codegen(
 
     Ok(module.into_compiled_module(
         config.emit_obj != EmitObj::None,
+        cgcx.split_dwarf_kind == SplitDwarfKind::Split,
         config.emit_bc,
         &cgcx.output_filenames,
     ))

compiler/rustc_codegen_llvm/src/debuginfo/metadata.rs

@@ -995,7 +995,11 @@ pub fn compile_unit_metadata(
     let name_in_debuginfo = name_in_debuginfo.to_string_lossy();
     let work_dir = tcx.sess.working_dir.0.to_string_lossy();
     let flags = "\0";
-    let split_name = "";
+    let split_name = tcx
+        .output_filenames(LOCAL_CRATE)
+        .split_dwarf_file(tcx.sess.opts.debugging_opts.split_dwarf, Some(codegen_unit_name))
+        .unwrap_or_default();
+    let split_name = split_name.to_str().unwrap();
 
     // FIXME(#60020):
     //
@@ -1040,7 +1044,7 @@ pub fn compile_unit_metadata(
             split_name.len(),
             kind,
             0,
-            true,
+            tcx.sess.opts.debugging_opts.split_dwarf_inlining,
         );
 
         if tcx.sess.opts.debugging_opts.profile {

compiler/rustc_codegen_llvm/src/lib.rs

@@ -20,7 +20,7 @@ pub use llvm_util::target_features;
 use rustc_ast::expand::allocator::AllocatorKind;
 use rustc_codegen_ssa::back::lto::{LtoModuleCodegen, SerializedModule, ThinModule};
 use rustc_codegen_ssa::back::write::{
-    CodegenContext, FatLTOInput, ModuleConfig, TargetMachineFactoryFn,
+    CodegenContext, FatLTOInput, ModuleConfig, TargetMachineFactoryConfig, TargetMachineFactoryFn,
 };
 use rustc_codegen_ssa::traits::*;
 use rustc_codegen_ssa::ModuleCodegen;
@@ -332,7 +332,7 @@ impl ModuleLlvm {
         unsafe {
             let llcx = llvm::LLVMRustContextCreate(tcx.sess.fewer_names());
             let llmod_raw = context::create_module(tcx, llcx, mod_name) as *const _;
-            ModuleLlvm { llmod_raw, llcx, tm: create_target_machine(tcx) }
+            ModuleLlvm { llmod_raw, llcx, tm: create_target_machine(tcx, mod_name) }
         }
     }
 
@@ -353,7 +353,13 @@ impl ModuleLlvm {
         unsafe {
             let llcx = llvm::LLVMRustContextCreate(cgcx.fewer_names);
             let llmod_raw = back::lto::parse_module(llcx, name, buffer, handler)?;
-            let tm = match (cgcx.tm_factory)() {
+
+            let split_dwarf_file = cgcx
+                .output_filenames
+                .split_dwarf_file(cgcx.split_dwarf_kind, Some(name.to_str().unwrap()));
+            let tm_factory_config = TargetMachineFactoryConfig { split_dwarf_file };
+
+            let tm = match (cgcx.tm_factory)(tm_factory_config) {
                 Ok(m) => m,
                 Err(e) => {
                     handler.struct_err(&e).emit();

compiler/rustc_codegen_ssa/src/back/link.rs

@@ -21,7 +21,9 @@ use super::archive::ArchiveBuilder;
 use super::command::Command;
 use super::linker::{self, Linker};
 use super::rpath::{self, RPathConfig};
-use crate::{looks_like_rust_object_file, CodegenResults, CrateInfo, METADATA_FILENAME};
+use crate::{
+    looks_like_rust_object_file, CodegenResults, CompiledModule, CrateInfo, METADATA_FILENAME,
+};
 
 use cc::windows_registry;
 use tempfile::Builder as TempFileBuilder;
@@ -96,6 +98,9 @@ pub fn link_binary<'a, B: ArchiveBuilder<'a>>(
                         path.as_ref(),
                         target_cpu,
                     );
+                    if sess.opts.debugging_opts.split_dwarf == config::SplitDwarfKind::Split {
+                        link_dwarf_object(sess, &out_filename);
+                    }
                 }
             }
             if sess.opts.json_artifact_notifications {
@@ -107,22 +112,30 @@ pub fn link_binary<'a, B: ArchiveBuilder<'a>>(
     // Remove the temporary object file and metadata if we aren't saving temps
     sess.time("link_binary_remove_temps", || {
         if !sess.opts.cg.save_temps {
+            let remove_temps_from_module = |module: &CompiledModule| {
+                if let Some(ref obj) = module.object {
+                    remove(sess, obj);
+                }
+
+                if let Some(ref obj) = module.dwarf_object {
+                    remove(sess, obj);
+                }
+            };
+
             if sess.opts.output_types.should_codegen()
                 && !preserve_objects_for_their_debuginfo(sess)
             {
-                for obj in codegen_results.modules.iter().filter_map(|m| m.object.as_ref()) {
+                for module in &codegen_results.modules {
-                    remove(sess, obj);
+                    remove_temps_from_module(module);
                 }
             }
+
             if let Some(ref metadata_module) = codegen_results.metadata_module {
-                if let Some(ref obj) = metadata_module.object {
+                remove_temps_from_module(metadata_module);
-                    remove(sess, obj);
-                }
             }
+
             if let Some(ref allocator_module) = codegen_results.allocator_module {
-                if let Some(ref obj) = allocator_module.object {
+                remove_temps_from_module(allocator_module);
-                    remove(sess, obj);
-                }
             }
         }
     });
@@ -446,6 +459,69 @@ fn link_staticlib<'a, B: ArchiveBuilder<'a>>(
     }
 }
 
+fn escape_stdout_stderr_string(s: &[u8]) -> String {
+    str::from_utf8(s).map(|s| s.to_owned()).unwrap_or_else(|_| {
+        let mut x = "Non-UTF-8 output: ".to_string();
+        x.extend(s.iter().flat_map(|&b| ascii::escape_default(b)).map(char::from));
+        x
+    })
+}
+
+const LLVM_DWP_EXECUTABLE: &'static str = "rust-llvm-dwp";
+
+/// Invoke `llvm-dwp` (shipped alongside rustc) to link `dwo` files from Split DWARF into a `dwp`
+/// file.
+fn link_dwarf_object<'a>(sess: &'a Session, executable_out_filename: &Path) {
+    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);
+    cmd.arg("-e");
+    cmd.arg(executable_out_filename);
+    cmd.arg("-o");
+    cmd.arg(&dwp_out_filename);
+
+    let mut new_path = sess.host_filesearch(PathKind::All).get_tools_search_paths(false);
+    if let Some(path) = env::var_os("PATH") {
+        new_path.extend(env::split_paths(&path));
+    }
+    let new_path = env::join_paths(new_path).unwrap();
+    cmd.env("PATH", new_path);
+
+    info!("{:?}", &cmd);
+    match sess.time("run_dwp", || cmd.output()) {
+        Ok(prog) if !prog.status.success() => {
+            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(_) => {}
+        Err(e) => {
+            let dwp_not_found = e.kind() == io::ErrorKind::NotFound;
+            let mut err = if dwp_not_found {
+                sess.struct_err(&format!("linker `{}` not found", LLVM_DWP_EXECUTABLE))
+            } 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();
+        }
+    }
+}
+
 /// Create a dynamic library or executable.
 ///
 /// This will invoke the system linker/cc to create the resulting file. This links to all upstream
@@ -661,7 +737,7 @@ fn link_natively<'a, B: ArchiveBuilder<'a>>(
                     prog.status
                 ))
                 .note(&format!("{:?}", &cmd))
-                .note(&escape_string(&output))
+                .note(&escape_stdout_stderr_string(&output))
                 .emit();
 
                 // If MSVC's `link.exe` was expected but the return code
@@ -714,8 +790,8 @@ fn link_natively<'a, B: ArchiveBuilder<'a>>(
 
                 sess.abort_if_errors();
             }
-            info!("linker stderr:\n{}", escape_string(&prog.stderr));
+            info!("linker stderr:\n{}", escape_stdout_stderr_string(&prog.stderr));
-            info!("linker stdout:\n{}", escape_string(&prog.stdout));
+            info!("linker stdout:\n{}", escape_stdout_stderr_string(&prog.stdout));
         }
         Err(e) => {
             let linker_not_found = e.kind() == io::ErrorKind::NotFound;

compiler/rustc_codegen_ssa/src/back/write.rs

@@ -274,8 +274,19 @@ impl ModuleConfig {
     }
 }
 
-pub type TargetMachineFactoryFn<B> =
+/// Configuration passed to the function returned by the `target_machine_factory`.
-    Arc<dyn Fn() -> Result<<B as WriteBackendMethods>::TargetMachine, String> + Send + Sync>;
+pub struct TargetMachineFactoryConfig {
+    /// Split DWARF is enabled in LLVM by checking that `TM.MCOptions.SplitDwarfFile` isn't empty,
+    /// so the path to the dwarf object has to be provided when we create the target machine.
+    /// This can be ignored by backends which do not need it for their Split DWARF support.
+    pub split_dwarf_file: Option<PathBuf>,
+}
+
+pub type TargetMachineFactoryFn<B> = Arc<
+    dyn Fn(TargetMachineFactoryConfig) -> Result<<B as WriteBackendMethods>::TargetMachine, String>
+        + Send
+        + Sync,
+>;
 
 pub type ExportedSymbols = FxHashMap<CrateNum, Arc<Vec<(String, SymbolExportLevel)>>>;
 
@@ -303,6 +314,7 @@ pub struct CodegenContext<B: WriteBackendMethods> {
     pub target_pointer_width: u32,
     pub target_arch: String,
     pub debuginfo: config::DebugInfo,
+    pub split_dwarf_kind: config::SplitDwarfKind,
 
     // Number of cgus excluding the allocator/metadata modules
     pub total_cgus: usize,
@@ -619,6 +631,12 @@ fn produce_final_output_artifacts(
                 }
             }
 
+            if let Some(ref path) = module.dwarf_object {
+                if !keep_numbered_objects {
+                    remove(sess, path);
+                }
+            }
+
             if let Some(ref path) = module.bytecode {
                 if !keep_numbered_bitcode {
                     remove(sess, path);
@@ -841,6 +859,7 @@ fn execute_copy_from_cache_work_item<B: ExtraBackendMethods>(
         name: module.name,
         kind: ModuleKind::Regular,
         object,
+        dwarf_object: None,
         bytecode: None,
     }))
 }
@@ -1019,6 +1038,7 @@ fn start_executing_work<B: ExtraBackendMethods>(
         target_pointer_width: tcx.sess.target.pointer_width,
         target_arch: tcx.sess.target.arch.clone(),
         debuginfo: tcx.sess.opts.debuginfo,
+        split_dwarf_kind: tcx.sess.opts.debugging_opts.split_dwarf,
     };
 
     // This is the "main loop" of parallel work happening for parallel codegen.

compiler/rustc_codegen_ssa/src/lib.rs

@@ -64,13 +64,15 @@ impl<M> ModuleCodegen<M> {
     pub fn into_compiled_module(
         self,
         emit_obj: bool,
+        emit_dwarf_obj: bool,
         emit_bc: bool,
         outputs: &OutputFilenames,
     ) -> CompiledModule {
         let object = emit_obj.then(|| outputs.temp_path(OutputType::Object, Some(&self.name)));
+        let dwarf_object = emit_dwarf_obj.then(|| outputs.temp_path_dwo(Some(&self.name)));
         let bytecode = emit_bc.then(|| outputs.temp_path(OutputType::Bitcode, Some(&self.name)));
 
-        CompiledModule { name: self.name.clone(), kind: self.kind, object, bytecode }
+        CompiledModule { name: self.name.clone(), kind: self.kind, object, dwarf_object, bytecode }
     }
 }
 
@@ -79,6 +81,7 @@ pub struct CompiledModule {
     pub name: String,
     pub kind: ModuleKind,
     pub object: Option<PathBuf>,
+    pub dwarf_object: Option<PathBuf>,
     pub bytecode: Option<PathBuf>,
 }
 

compiler/rustc_session/src/config.rs

@@ -550,6 +550,7 @@ impl_stable_hash_via_hash!(OutputFilenames);
 
 pub const RLINK_EXT: &str = "rlink";
 pub const RUST_CGU_EXT: &str = "rcgu";
+pub const DWARF_OBJECT_EXT: &str = "dwo";
 
 impl OutputFilenames {
     pub fn new(
@@ -583,7 +584,12 @@ impl OutputFilenames {
         self.temp_path_ext(extension, codegen_unit_name)
     }
 
-    /// Like temp_path, but also supports things where there is no corresponding
+    /// Like `temp_path`, but specifically for dwarf objects.
+    pub fn temp_path_dwo(&self, codegen_unit_name: Option<&str>) -> PathBuf {
+        self.temp_path_ext(DWARF_OBJECT_EXT, codegen_unit_name)
+    }
+
+    /// Like `temp_path`, but also supports things where there is no corresponding
     /// OutputType, like noopt-bitcode or lto-bitcode.
     pub fn temp_path_ext(&self, ext: &str, codegen_unit_name: Option<&str>) -> PathBuf {
         let mut extension = String::new();
@@ -610,6 +616,26 @@ impl OutputFilenames {
         path.set_extension(extension);
         path
     }
+
+    /// Returns the path for the Split DWARF file - this can differ depending on which Split DWARF
+    /// mode is being used, which is the logic that this function is intended to encapsulate.
+    pub fn split_dwarf_file(
+        &self,
+        split_dwarf_kind: SplitDwarfKind,
+        cgu_name: Option<&str>,
+    ) -> Option<PathBuf> {
+        let obj_out = self.temp_path(OutputType::Object, cgu_name);
+        let dwo_out = self.temp_path_dwo(cgu_name);
+        match split_dwarf_kind {
+            SplitDwarfKind::None => None,
+            // Single mode doesn't change how DWARF is emitted, but does add Split DWARF attributes
+            // (pointing at the path which is being determined here). Use the path to the current
+            // object file.
+            SplitDwarfKind::Single => Some(obj_out),
+            // Split mode emits the DWARF into a different file, use that path.
+            SplitDwarfKind::Split => Some(dwo_out),
+        }
+    }
 }
 
 pub fn host_triple() -> &'static str {