Link object files that use #[used]

compiler/rustc_codegen_ssa/src/back/link.rs

@@ -9,6 +9,7 @@ use std::{env, fmt, fs, io, mem, str};
 
 use cc::windows_registry;
 use itertools::Itertools;
+use object::read::archive::ArchiveFile;
 use regex::Regex;
 use rustc_arena::TypedArena;
 use rustc_ast::CRATE_NODE_ID;
@@ -78,6 +79,7 @@ pub fn link_binary(
     let _timer = sess.timer("link_binary");
     let output_metadata = sess.opts.output_types.contains_key(&OutputType::Metadata);
     let mut tempfiles_for_stdout_output: Vec<PathBuf> = Vec::new();
+    let mut tempfiles_for_linked_objects: Vec<PathBuf> = Vec::new();
     for &crate_type in &codegen_results.crate_info.crate_types {
         // Ignore executable crates if we have -Z no-codegen, as they will error.
         if (sess.opts.unstable_opts.no_codegen || !sess.opts.output_types.should_codegen())
@@ -142,6 +144,8 @@ pub fn link_binary(
                         &out_filename,
                         &codegen_results,
                         path.as_ref(),
+                        &mut tempfiles_for_linked_objects,
+                        outputs,
                     );
                 }
             }
@@ -214,6 +218,10 @@ pub fn link_binary(
             ensure_removed(sess.dcx(), &temp);
         }
 
+        for temp in tempfiles_for_linked_objects {
+            ensure_removed(sess.dcx(), &temp);
+        }
+
         // If no requested outputs require linking, then the object temporaries should
         // be kept.
         if !sess.opts.output_types.should_link() {
@@ -771,6 +779,8 @@ fn link_natively(
     out_filename: &Path,
     codegen_results: &CodegenResults,
     tmpdir: &Path,
+    tempfiles_for_linked_objects: &mut Vec<PathBuf>,
+    outputs: &OutputFilenames,
 ) {
     info!("preparing {:?} to {:?}", crate_type, out_filename);
     let (linker_path, flavor) = linker_and_flavor(sess);
@@ -795,6 +805,8 @@ fn link_natively(
         temp_filename,
         codegen_results,
         self_contained_components,
+        tempfiles_for_linked_objects,
+        outputs,
     );
 
     linker::disable_localization(&mut cmd);
@@ -2254,6 +2266,8 @@ fn linker_with_args(
     out_filename: &Path,
     codegen_results: &CodegenResults,
     self_contained_components: LinkSelfContainedComponents,
+    tempfiles_for_linked_objects: &mut Vec<PathBuf>,
+    outputs: &OutputFilenames,
 ) -> Command {
     let self_contained_crt_objects = self_contained_components.is_crt_objects_enabled();
     let cmd = &mut *super::linker::get_linker(
@@ -2329,6 +2343,13 @@ fn linker_with_args(
     add_local_crate_regular_objects(cmd, codegen_results);
     add_local_crate_metadata_objects(cmd, crate_type, codegen_results);
     add_local_crate_allocator_objects(cmd, codegen_results);
+    add_local_crate_linked_objects(
+        cmd,
+        codegen_results,
+        crate_type,
+        tempfiles_for_linked_objects,
+        outputs,
+    );
 
     // Avoid linking to dynamic libraries unless they satisfy some undefined symbols
     // at the point at which they are specified on the command line.
@@ -2925,6 +2946,35 @@ fn rehome_lib_path(sess: &Session, path: &Path) -> PathBuf {
     }
 }
 
+fn add_local_crate_linked_objects(
+    cmd: &mut dyn Linker,
+    codegen_results: &CodegenResults,
+    crate_type: CrateType,
+    tempfiles_for_linked_objects: &mut Vec<PathBuf>,
+    outputs: &OutputFilenames,
+) {
+    for (cnum, objects) in &codegen_results.crate_info.linked_objects[&crate_type] {
+        let src = &codegen_results.crate_info.used_crate_source[cnum];
+        let cratepath = &src.rlib.as_ref().unwrap().0;
+        let archive_map = unsafe { Mmap::map(File::open(cratepath).unwrap()).unwrap() };
+        let archive = ArchiveFile::parse(&*archive_map)
+            .map_err(|err| io::Error::new(io::ErrorKind::InvalidData, err))
+            .unwrap();
+        for member in archive.members() {
+            let member = member.unwrap();
+            let name = std::str::from_utf8(member.name()).unwrap();
+            if objects.contains(name) {
+                let data = member.data(&*archive_map).unwrap();
+                let obj =
+                    outputs.temp_path(OutputType::Object, Some(&format!("{name}.linked_object")));
+                fs::write(&obj, data).unwrap();
+                cmd.add_object(&obj);
+                tempfiles_for_linked_objects.push(obj);
+            }
+        }
+    }
+}
+
 // Adds the static "rlib" versions of all crates to the command line.
 // There's a bit of magic which happens here specifically related to LTO,
 // namely that we remove upstream object files.

compiler/rustc_codegen_ssa/src/back/linker.rs

@@ -5,14 +5,16 @@ use std::path::{Path, PathBuf};
 use std::{env, io, iter, mem, str};
 
 use cc::windows_registry;
+use rustc_data_structures::fx::{FxHashSet, FxIndexMap};
 use rustc_hir::def_id::{CrateNum, LOCAL_CRATE};
 use rustc_metadata::{
     find_native_static_library, try_find_native_dynamic_library, try_find_native_static_library,
 };
 use rustc_middle::bug;
 use rustc_middle::middle::dependency_format::Linkage;
-use rustc_middle::middle::exported_symbols;
-use rustc_middle::middle::exported_symbols::{ExportedSymbol, SymbolExportInfo, SymbolExportKind};
+use rustc_middle::middle::exported_symbols::{
+    self, ExportedSymbol, SymbolExportInfo, SymbolExportKind,
+};
 use rustc_middle::ty::TyCtxt;
 use rustc_session::Session;
 use rustc_session::config::{self, CrateType, DebugInfo, LinkerPluginLto, Lto, OptLevel, Strip};
@@ -21,6 +23,7 @@ use rustc_target::spec::{Cc, LinkOutputKind, LinkerFlavor, Lld};
 use tracing::{debug, warn};
 
 use super::command::Command;
+use super::link::are_upstream_rust_objects_already_included;
 use super::symbol_export;
 use crate::errors;
 
@@ -1753,17 +1756,15 @@ impl<'a> Linker for AixLinker<'a> {
 fn for_each_exported_symbols_include_dep<'tcx>(
     tcx: TyCtxt<'tcx>,
     crate_type: CrateType,
-    mut callback: impl FnMut(ExportedSymbol<'tcx>, SymbolExportInfo, CrateNum),
+    mut callback: impl FnMut(&'tcx [(ExportedSymbol<'tcx>, SymbolExportInfo)], CrateNum),
 ) {
     let formats = tcx.dependency_formats(());
     let deps = &formats[&crate_type];
 
     for (cnum, dep_format) in deps.iter_enumerated() {
         // For each dependency that we are linking to statically ...
         if *dep_format == Linkage::Static {
-            for &(symbol, info) in tcx.exported_symbols(cnum).iter() {
-                callback(symbol, info, cnum);
-            }
+            callback(tcx.exported_symbols(cnum), cnum);
         }
     }
 }
@@ -1783,12 +1784,14 @@ pub(crate) fn exported_symbols(tcx: TyCtxt<'_>, crate_type: CrateType) -> Vec<St
 fn exported_symbols_for_non_proc_macro(tcx: TyCtxt<'_>, crate_type: CrateType) -> Vec<String> {
     let mut symbols = Vec::new();
     let export_threshold = symbol_export::crates_export_threshold(&[crate_type]);
-    for_each_exported_symbols_include_dep(tcx, crate_type, |symbol, info, cnum| {
-        if info.level.is_below_threshold(export_threshold) {
-            symbols.push(symbol_export::exporting_symbol_name_for_instance_in_crate(
-                tcx, symbol, cnum,
-            ));
-            symbol_export::extend_exported_symbols(&mut symbols, tcx, symbol, cnum);
+    for_each_exported_symbols_include_dep(tcx, crate_type, |exported_symbols, cnum| {
+        for &(symbol, info) in exported_symbols {
+            if info.level.is_below_threshold(export_threshold) {
+                symbols.push(symbol_export::exporting_symbol_name_for_instance_in_crate(
+                    tcx, symbol, cnum,
+                ));
+                symbol_export::extend_exported_symbols(&mut symbols, tcx, symbol, cnum);
+            }
         }
     });
 
@@ -1808,30 +1811,84 @@ fn exported_symbols_for_proc_macro_crate(tcx: TyCtxt<'_>) -> Vec<String> {
     vec![proc_macro_decls_name, metadata_symbol_name]
 }
 
-pub(crate) fn linked_symbols(
+pub(crate) fn linked_objects(
     tcx: TyCtxt<'_>,
     crate_type: CrateType,
-) -> Vec<(String, SymbolExportKind)> {
+    linked_symbols: &mut Vec<(String, SymbolExportKind)>,
+) -> FxIndexMap<CrateNum, FxHashSet<String>> {
     match crate_type {
         CrateType::Executable | CrateType::Cdylib | CrateType::Dylib => (),
         CrateType::Staticlib | CrateType::ProcMacro | CrateType::Rlib => {
-            return Vec::new();
+            return FxIndexMap::default();
         }
     }
 
-    let mut symbols = Vec::new();
-
+    let mut objects = FxIndexMap::default();
+    let upstream_rust_objects_already_included =
+        are_upstream_rust_objects_already_included(tcx.sess);
     let export_threshold = symbol_export::crates_export_threshold(&[crate_type]);
-    for_each_exported_symbols_include_dep(tcx, crate_type, |symbol, info, cnum| {
-        if info.level.is_below_threshold(export_threshold) || info.used {
-            symbols.push((
-                symbol_export::linking_symbol_name_for_instance_in_crate(tcx, symbol, cnum),
-                info.kind,
-            ));
+    for_each_exported_symbols_include_dep(tcx, crate_type, |exported_symbols, cnum| {
+        let exported_symbols = exported_symbols.iter().filter(|(_, info)| {
+            (!matches!(crate_type, CrateType::Executable)
+                && info.level.is_below_threshold(export_threshold))
+                || info.used
+        });
+        if cnum == LOCAL_CRATE {
+            // Since the local crate is always linked directly to object files, `#[used]` works as expected,
+            // we only need add undefined symbols.
+            linked_symbols.extend(
+                exported_symbols
+                    .filter(|(symbol, _)| match symbol {
+                        ExportedSymbol::NonGeneric { cgu, .. } => cgu.is_none(),
+                        ExportedSymbol::Generic(..)
+                        | ExportedSymbol::DropGlue(..)
+                        | ExportedSymbol::AsyncDropGlueCtorShim(..) => false,
+                        ExportedSymbol::ThreadLocalShim(_def_id) => false,
+                        ExportedSymbol::NoDefId(..) => true,
+                    })
+                    .map(|&(symbol, info)| {
+                        (
+                            symbol_export::linking_symbol_name_for_instance_in_crate(
+                                tcx, symbol, cnum,
+                            ),
+                            info.kind,
+                        )
+                    }),
+            );
+            return;
+        }
+        if matches!(crate_type, CrateType::Executable) && tcx.is_compiler_builtins(cnum) {
+            return;
         }
+        let lto = upstream_rust_objects_already_included;
+        let mut cgus = FxHashSet::default();
+        for &(symbol, info) in exported_symbols {
+            match symbol {
+                ExportedSymbol::NonGeneric { cgu: Some(cgu), .. } => {
+                    if !lto {
+                        cgus.insert(cgu.as_str().to_string());
+                    }
+                }
+                ExportedSymbol::NonGeneric { cgu: None, .. } | ExportedSymbol::NoDefId(..) => {
+                    // Unresolved symbols may come from external libraries.
+                    linked_symbols.push((
+                        symbol_export::linking_symbol_name_for_instance_in_crate(tcx, symbol, cnum),
+                        info.kind,
+                    ));
+                }
+                ExportedSymbol::Generic(..)
+                | ExportedSymbol::DropGlue(..)
+                | ExportedSymbol::AsyncDropGlueCtorShim(..)
+                | ExportedSymbol::ThreadLocalShim(..) => {}
+            };
+        }
+        if cgus.is_empty() {
+            return;
+        }
+        objects.insert(cnum, cgus);
     });
 
-    symbols
+    objects
 }
 
 /// Much simplified and explicit CLI for the NVPTX linker. The linker operates

compiler/rustc_codegen_ssa/src/back/symbol_export.rs

@@ -9,10 +9,12 @@ use rustc_middle::middle::codegen_fn_attrs::CodegenFnAttrFlags;
 use rustc_middle::middle::exported_symbols::{
     ExportedSymbol, SymbolExportInfo, SymbolExportKind, SymbolExportLevel, metadata_symbol_name,
 };
+use rustc_middle::mir::mono::{CodegenUnit, MonoItem};
 use rustc_middle::query::LocalCrate;
 use rustc_middle::ty::{self, GenericArgKind, GenericArgsRef, Instance, SymbolName, Ty, TyCtxt};
 use rustc_middle::util::Providers;
-use rustc_session::config::{CrateType, OomStrategy};
+use rustc_session::config::{CrateType, OomStrategy, OutputFilenames, OutputType};
+use rustc_span::Symbol;
 use rustc_target::callconv::Conv;
 use rustc_target::spec::{SanitizerSet, TlsModel};
 use tracing::debug;
@@ -168,6 +170,36 @@ fn is_reachable_non_generic_provider_extern(tcx: TyCtxt<'_>, def_id: DefId) -> b
     tcx.reachable_non_generics(def_id.krate).contains_key(&def_id)
 }
 
+fn find_codegen_unit<'tcx>(
+    tcx: TyCtxt<'tcx>,
+    codegen_units: &'tcx [CodegenUnit<'tcx>],
+    outputs: &'tcx OutputFilenames,
+    def_id: DefId,
+) -> Option<Symbol> {
+    if !tcx.is_codegened_item(def_id) {
+        return None;
+    }
+    let item = if tcx.is_static(def_id) {
+        MonoItem::Static(def_id)
+    } else {
+        MonoItem::Fn(Instance::mono(tcx, def_id))
+    };
+    codegen_units.iter().find_map(|cgu| {
+        if cgu.contains_item(&item) {
+            Some(Symbol::intern(
+                outputs
+                    .temp_path(OutputType::Object, Some(cgu.name().as_str()))
+                    .file_name()
+                    .unwrap()
+                    .to_str()
+                    .unwrap(),
+            ))
+        } else {
+            None
+        }
+    })
+}
+
 fn exported_symbols_provider_local(
     tcx: TyCtxt<'_>,
     _: LocalCrate,
@@ -182,8 +214,20 @@ fn exported_symbols_provider_local(
         tcx.reachable_non_generics(LOCAL_CRATE).to_sorted(&hcx, true)
     });
 
-    let mut symbols: Vec<_> =
-        sorted.iter().map(|&(&def_id, &info)| (ExportedSymbol::NonGeneric(def_id), info)).collect();
+    let outputs = tcx.output_filenames(());
+    let codegen_units = tcx.collect_and_partition_mono_items(()).codegen_units;
+    let mut symbols: Vec<_> = sorted
+        .iter()
+        .map(|&(&def_id, &info)| {
+            (
+                ExportedSymbol::NonGeneric {
+                    def_id,
+                    cgu: find_codegen_unit(tcx, codegen_units, outputs, def_id),
+                },
+                info,
+            )
+        })
+        .collect();
 
     // Export TLS shims
     if !tcx.sess.target.dll_tls_export {
@@ -433,7 +477,7 @@ fn upstream_monomorphizations_provider(
                         continue;
                     }
                 }
-                ExportedSymbol::NonGeneric(..)
+                ExportedSymbol::NonGeneric { .. }
                 | ExportedSymbol::ThreadLocalShim(..)
                 | ExportedSymbol::NoDefId(..) => {
                     // These are no monomorphizations
@@ -545,7 +589,7 @@ pub(crate) fn symbol_name_for_instance_in_crate<'tcx>(
     // This is something instantiated in an upstream crate, so we have to use
     // the slower (because uncached) version of computing the symbol name.
     match symbol {
-        ExportedSymbol::NonGeneric(def_id) => {
+        ExportedSymbol::NonGeneric { def_id, .. } => {
             rustc_symbol_mangling::symbol_name_for_instance_in_crate(
                 tcx,
                 Instance::mono(tcx, def_id),
@@ -590,12 +634,12 @@ fn calling_convention_for_symbol<'tcx>(
     symbol: ExportedSymbol<'tcx>,
 ) -> (Conv, &'tcx [rustc_target::callconv::ArgAbi<'tcx, Ty<'tcx>>]) {
     let instance = match symbol {
-        ExportedSymbol::NonGeneric(def_id) | ExportedSymbol::Generic(def_id, _)
+        ExportedSymbol::NonGeneric { def_id, .. } | ExportedSymbol::Generic(def_id, _)
             if tcx.is_static(def_id) =>
         {
             None
         }
-        ExportedSymbol::NonGeneric(def_id) => Some(Instance::mono(tcx, def_id)),
+        ExportedSymbol::NonGeneric { def_id, .. } => Some(Instance::mono(tcx, def_id)),
         ExportedSymbol::Generic(def_id, args) => Some(Instance::new(def_id, args)),
         // DropGlue always use the Rust calling convention and thus follow the target's default
         // symbol decoration scheme.
@@ -711,7 +755,7 @@ fn maybe_emutls_symbol_name<'tcx>(
     undecorated: &str,
 ) -> Option<String> {
     if matches!(tcx.sess.tls_model(), TlsModel::Emulated)
-        && let ExportedSymbol::NonGeneric(def_id) = symbol
+        && let ExportedSymbol::NonGeneric { def_id, .. } = symbol
         && tcx.is_thread_local_static(def_id)
     {
         // When using emutls, LLVM will add the `__emutls_v.` prefix to thread local symbols,