Introduce -Zsplit-metadata to split metadata out of rlibs/dylibs

compiler/rustc_codegen_ssa/src/back/link.rs

@@ -296,7 +296,7 @@ fn link_rlib<'a>(
             let (metadata, metadata_position) = create_wrapper_file(
                 sess,
                 ".rmeta".to_string(),
-                codegen_results.metadata.raw_data(),
+                codegen_results.metadata.stub_or_full(),
             );
             let metadata = emit_wrapper_file(sess, &metadata, tmpdir, METADATA_FILENAME);
             match metadata_position {

compiler/rustc_codegen_ssa/src/back/metadata.rs

@@ -570,8 +570,8 @@ pub fn create_compressed_metadata_file(
     symbol_name: &str,
 ) -> Vec<u8> {
     let mut packed_metadata = rustc_metadata::METADATA_HEADER.to_vec();
-    packed_metadata.write_all(&(metadata.raw_data().len() as u64).to_le_bytes()).unwrap();
-    packed_metadata.extend(metadata.raw_data());
+    packed_metadata.write_all(&(metadata.stub_or_full().len() as u64).to_le_bytes()).unwrap();
+    packed_metadata.extend(metadata.stub_or_full());
 
     let Some(mut file) = create_object_file(sess) else {
         if sess.target.is_like_wasm {

compiler/rustc_interface/src/tests.rs

@@ -859,6 +859,7 @@ fn test_unstable_options_tracking_hash() {
     tracked!(simulate_remapped_rust_src_base, Some(PathBuf::from("/rustc/abc")));
     tracked!(small_data_threshold, Some(16));
     tracked!(split_lto_unit, Some(true));
+    tracked!(split_metadata, true);
     tracked!(src_hash_algorithm, Some(SourceFileHashAlgorithm::Sha1));
     tracked!(stack_protector, StackProtector::All);
     tracked!(teach, true);

compiler/rustc_metadata/src/fs.rs

@@ -3,7 +3,7 @@ use std::{fs, io};
 
 use rustc_data_structures::temp_dir::MaybeTempDir;
 use rustc_middle::ty::TyCtxt;
-use rustc_session::config::{OutFileName, OutputType};
+use rustc_session::config::{CrateType, OutFileName, OutputType};
 use rustc_session::output::filename_for_metadata;
 use rustc_session::{MetadataKind, Session};
 use tempfile::Builder as TempFileBuilder;
@@ -50,7 +50,14 @@ pub fn encode_and_write_metadata(tcx: TyCtxt<'_>) -> (EncodedMetadata, bool) {
         .tempdir_in(out_filename.parent().unwrap_or_else(|| Path::new("")))
         .unwrap_or_else(|err| tcx.dcx().emit_fatal(FailedCreateTempdir { err }));
     let metadata_tmpdir = MaybeTempDir::new(metadata_tmpdir, tcx.sess.opts.cg.save_temps);
-    let metadata_filename = metadata_tmpdir.as_ref().join(METADATA_FILENAME);
+    let metadata_filename = metadata_tmpdir.as_ref().join("full.rmeta");
+    let metadata_stub_filename = if tcx.sess.opts.unstable_opts.split_metadata
+        && !tcx.crate_types().contains(&CrateType::ProcMacro)
+    {
+        Some(metadata_tmpdir.as_ref().join("stub.rmeta"))
+    } else {
+        None
+    };
 
     // Always create a file at `metadata_filename`, even if we have nothing to write to it.
     // This simplifies the creation of the output `out_filename` when requested.
@@ -60,9 +67,15 @@ pub fn encode_and_write_metadata(tcx: TyCtxt<'_>) -> (EncodedMetadata, bool) {
             std::fs::File::create(&metadata_filename).unwrap_or_else(|err| {
                 tcx.dcx().emit_fatal(FailedCreateFile { filename: &metadata_filename, err });
             });
+            if let Some(metadata_stub_filename) = &metadata_stub_filename {
+                std::fs::File::create(metadata_stub_filename).unwrap_or_else(|err| {
+                    tcx.dcx()
+                        .emit_fatal(FailedCreateFile { filename: &metadata_stub_filename, err });
+                });
+            }
         }
         MetadataKind::Uncompressed | MetadataKind::Compressed => {
-            encode_metadata(tcx, &metadata_filename);
+            encode_metadata(tcx, &metadata_filename, metadata_stub_filename.as_deref())
         }
     };
 
@@ -100,9 +113,10 @@ pub fn encode_and_write_metadata(tcx: TyCtxt<'_>) -> (EncodedMetadata, bool) {
 
     // Load metadata back to memory: codegen may need to include it in object files.
     let metadata =
-        EncodedMetadata::from_path(metadata_filename, metadata_tmpdir).unwrap_or_else(|err| {
-            tcx.dcx().emit_fatal(FailedCreateEncodedMetadata { err });
-        });
+        EncodedMetadata::from_path(metadata_filename, metadata_stub_filename, metadata_tmpdir)
+            .unwrap_or_else(|err| {
+                tcx.dcx().emit_fatal(FailedCreateEncodedMetadata { err });
+            });
 
     let need_metadata_module = metadata_kind == MetadataKind::Compressed;
 

compiler/rustc_metadata/src/locator.rs

@@ -247,6 +247,7 @@ pub(crate) struct CrateLocator<'a> {
 
     // Immutable per-search configuration.
     crate_name: Symbol,
+    // Dependency paths passed through --extern
     exact_paths: Vec<CanonicalizedPath>,
     pub hash: Option<Svh>,
     extra_filename: Option<&'a str>,
@@ -511,6 +512,8 @@ impl<'a> CrateLocator<'a> {
             rlib: self.extract_one(rlibs, CrateFlavor::Rlib, &mut slot)?,
             dylib: self.extract_one(dylibs, CrateFlavor::Dylib, &mut slot)?,
         };
+        // Question: check if we have no rmeta, but rlib/dylib with is_stub, and in that case
+        // invoke `find_library`?
         Ok(slot.map(|(svh, metadata, _)| (svh, Library { source, metadata })))
     }
 
@@ -728,37 +731,41 @@ impl<'a> CrateLocator<'a> {
             let Some(file) = loc_orig.file_name().and_then(|s| s.to_str()) else {
                 return Err(CrateError::ExternLocationNotFile(self.crate_name, loc_orig.clone()));
             };
-            // FnMut cannot return reference to captured value, so references
-            // must be taken outside the closure.
-            let rlibs = &mut rlibs;
-            let rmetas = &mut rmetas;
-            let dylibs = &mut dylibs;
-            let type_via_filename = (|| {
-                if file.starts_with("lib") {
-                    if file.ends_with(".rlib") {
-                        return Some(rlibs);
+            if file.starts_with("lib") {
+                if file.ends_with(".rlib") {
+                    // In case the .rlib contains only stub metadata, we will most likely
+                    // need to load a corresponding .rmeta file. Since it will often be
+                    // located right next to the .rlib path, directly add it to the candidate
+                    // paths as a "fast-path".
+                    let possible_rmeta_path = loc_orig.with_extension("rmeta");
+                    if let Ok(rmeta_path) = try_canonicalize(possible_rmeta_path) {
+                        rmetas.insert(rmeta_path, PathKind::ExternFlag);
                     }
-                    if file.ends_with(".rmeta") {
-                        return Some(rmetas);
-                    }
-                }
-                let dll_prefix = self.target.dll_prefix.as_ref();
-                let dll_suffix = self.target.dll_suffix.as_ref();
-                if file.starts_with(dll_prefix) && file.ends_with(dll_suffix) {
-                    return Some(dylibs);
+                    rlibs.insert(loc_canon.clone(), PathKind::ExternFlag);
+                    continue;
                 }
-                None
-            })();
-            match type_via_filename {
-                Some(type_via_filename) => {
-                    type_via_filename.insert(loc_canon.clone(), PathKind::ExternFlag);
+                if file.ends_with(".rmeta") {
+                    rmetas.insert(loc_canon.clone(), PathKind::ExternFlag);
+                    continue;
                 }
-                None => {
-                    self.crate_rejections
-                        .via_filename
-                        .push(CrateMismatch { path: loc_orig.clone(), got: String::new() });
+            }
+            let dll_prefix = self.target.dll_prefix.as_ref();
+            let dll_suffix = self.target.dll_suffix.as_ref();
+            if let Some(stem) =
+                file.strip_prefix(dll_prefix).and_then(|name| name.strip_suffix(dll_suffix))
+            {
+                // See comment above about stub metadata, it applies also here for dylibs
+                let possible_rmeta_path =
+                    loc_orig.parent().unwrap().join(format!("lib{stem}.rmeta"));
+                if let Ok(rmeta_path) = try_canonicalize(possible_rmeta_path) {
+                    rmetas.insert(rmeta_path, PathKind::ExternFlag);
                 }
+                dylibs.insert(loc_canon.clone(), PathKind::ExternFlag);
+                continue;
             }
+            self.crate_rejections
+                .via_filename
+                .push(CrateMismatch { path: loc_orig.clone(), got: String::new() });
         }
 
         // Extract the dylib/rlib/rmeta triple.

compiler/rustc_metadata/src/rmeta/encoder.rs

@@ -703,6 +703,7 @@ impl<'a, 'tcx> EncodeContext<'a, 'tcx> {
                     triple: tcx.sess.opts.target_triple.clone(),
                     hash: tcx.crate_hash(LOCAL_CRATE),
                     is_proc_macro_crate: proc_macro_data.is_some(),
+                    is_stub: false,
                 },
                 extra_filename: tcx.sess.opts.cg.extra_filename.clone(),
                 stable_crate_id: tcx.def_path_hash(LOCAL_CRATE.as_def_id()).stable_crate_id(),
@@ -2235,54 +2236,75 @@ fn prefetch_mir(tcx: TyCtxt<'_>) {
 // generated regardless of trailing bytes that end up in it.
 
 pub struct EncodedMetadata {
-    // The declaration order matters because `mmap` should be dropped before `_temp_dir`.
-    mmap: Option<Mmap>,
+    // The declaration order matters because `full_metadata` should be dropped
+    // before `_temp_dir`.
+    full_metadata: Option<Mmap>,
+    // This is an optional stub metadata containing only the crate header.
+    // The header should be very small, so we load it directly into memory.
+    stub_metadata: Option<Vec<u8>>,
     // We need to carry MaybeTempDir to avoid deleting the temporary
     // directory while accessing the Mmap.
     _temp_dir: Option<MaybeTempDir>,
 }
 
 impl EncodedMetadata {
     #[inline]
-    pub fn from_path(path: PathBuf, temp_dir: Option<MaybeTempDir>) -> std::io::Result<Self> {
+    pub fn from_path(
+        path: PathBuf,
+        stub_path: Option<PathBuf>,
+        temp_dir: Option<MaybeTempDir>,
+    ) -> std::io::Result<Self> {
         let file = std::fs::File::open(&path)?;
         let file_metadata = file.metadata()?;
         if file_metadata.len() == 0 {
-            return Ok(Self { mmap: None, _temp_dir: None });
+            return Ok(Self { full_metadata: None, stub_metadata: None, _temp_dir: None });
         }
-        let mmap = unsafe { Some(Mmap::map(file)?) };
-        Ok(Self { mmap, _temp_dir: temp_dir })
+        let full_mmap = unsafe { Some(Mmap::map(file)?) };
+
+        let stub =
+            if let Some(stub_path) = stub_path { Some(std::fs::read(stub_path)?) } else { None };
+
+        Ok(Self { full_metadata: full_mmap, stub_metadata: stub, _temp_dir: temp_dir })
+    }
+
+    #[inline]
+    pub fn full(&self) -> &[u8] {
+        &self.full_metadata.as_deref().unwrap_or_default()
     }
 
     #[inline]
-    pub fn raw_data(&self) -> &[u8] {
-        self.mmap.as_deref().unwrap_or_default()
+    pub fn stub_or_full(&self) -> &[u8] {
+        self.stub_metadata.as_deref().unwrap_or(self.full())
     }
 }
 
 impl<S: Encoder> Encodable<S> for EncodedMetadata {
     fn encode(&self, s: &mut S) {
-        let slice = self.raw_data();
+        self.stub_metadata.encode(s);
+
+        let slice = self.full();
         slice.encode(s)
     }
 }
 
 impl<D: Decoder> Decodable<D> for EncodedMetadata {
     fn decode(d: &mut D) -> Self {
+        let stub = <Option<Vec<u8>>>::decode(d);
+
         let len = d.read_usize();
-        let mmap = if len > 0 {
+        let full_metadata = if len > 0 {
             let mut mmap = MmapMut::map_anon(len).unwrap();
             mmap.copy_from_slice(d.read_raw_bytes(len));
             Some(mmap.make_read_only().unwrap())
         } else {
             None
         };
 
-        Self { mmap, _temp_dir: None }
+        Self { full_metadata, stub_metadata: stub, _temp_dir: None }
     }
 }
 
-pub fn encode_metadata(tcx: TyCtxt<'_>, path: &Path) {
+pub fn encode_metadata(tcx: TyCtxt<'_>, path: &Path, ref_path: Option<&Path>) {
     let _prof_timer = tcx.prof.verbose_generic_activity("generate_crate_metadata");
 
     // Since encoding metadata is not in a query, and nothing is cached,
@@ -2296,6 +2318,42 @@ pub fn encode_metadata(tcx: TyCtxt<'_>, path: &Path) {
         join(|| prefetch_mir(tcx), || tcx.exported_symbols(LOCAL_CRATE));
     }
 
+    with_encode_metadata_header(tcx, path, |ecx| {
+        // Encode all the entries and extra information in the crate,
+        // culminating in the `CrateRoot` which points to all of it.
+        let root = ecx.encode_crate_root();
+
+        // Flush buffer to ensure backing file has the correct size.
+        ecx.opaque.flush();
+        // Record metadata size for self-profiling
+        tcx.prof.artifact_size(
+            "crate_metadata",
+            "crate_metadata",
+            ecx.opaque.file().metadata().unwrap().len(),
+        );
+
+        root.position.get()
+    });
+
+    if let Some(ref_path) = ref_path {
+        with_encode_metadata_header(tcx, ref_path, |ecx| {
+            let header: LazyValue<CrateHeader> = ecx.lazy(CrateHeader {
+                name: tcx.crate_name(LOCAL_CRATE),
+                triple: tcx.sess.opts.target_triple.clone(),
+                hash: tcx.crate_hash(LOCAL_CRATE),
+                is_proc_macro_crate: false,
+                is_stub: true,
+            });
+            header.position.get()
+        });
+    }
+}
+
+fn with_encode_metadata_header(
+    tcx: TyCtxt<'_>,
+    path: &Path,
+    f: impl FnOnce(&mut EncodeContext<'_, '_>) -> usize,
+) {
     let mut encoder = opaque::FileEncoder::new(path)
         .unwrap_or_else(|err| tcx.dcx().emit_fatal(FailCreateFileEncoder { err }));
     encoder.emit_raw_bytes(METADATA_HEADER);
@@ -2330,9 +2388,7 @@ pub fn encode_metadata(tcx: TyCtxt<'_>, path: &Path) {
     // Encode the rustc version string in a predictable location.
     rustc_version(tcx.sess.cfg_version).encode(&mut ecx);
 
-    // Encode all the entries and extra information in the crate,
-    // culminating in the `CrateRoot` which points to all of it.
-    let root = ecx.encode_crate_root();
+    let root_position = f(&mut ecx);
 
     // Make sure we report any errors from writing to the file.
     // If we forget this, compilation can succeed with an incomplete rmeta file,
@@ -2342,12 +2398,9 @@ pub fn encode_metadata(tcx: TyCtxt<'_>, path: &Path) {
     }
 
     let file = ecx.opaque.file();
-    if let Err(err) = encode_root_position(file, root.position.get()) {
+    if let Err(err) = encode_root_position(file, root_position) {
         tcx.dcx().emit_fatal(FailWriteFile { path: ecx.opaque.path(), err });
     }
-
-    // Record metadata size for self-profiling
-    tcx.prof.artifact_size("crate_metadata", "crate_metadata", file.metadata().unwrap().len());
 }
 
 fn encode_root_position(mut file: &File, pos: usize) -> Result<(), std::io::Error> {

compiler/rustc_metadata/src/rmeta/mod.rs

@@ -220,6 +220,12 @@ pub(crate) struct CrateHeader {
     /// This is separate from [`ProcMacroData`] to avoid having to update [`METADATA_VERSION`] every
     /// time ProcMacroData changes.
     pub(crate) is_proc_macro_crate: bool,
+    /// Whether this crate metadata section is just a stub.
+    /// Stubs do not contain the full metadata (it will be typically stored
+    /// in a separate rmeta file).
+    ///
+    /// This is used inside rlibs and dylibs when using `-Zsplit-metadata`.
+    pub(crate) is_stub: bool,
 }
 
 /// Serialized `.rmeta` data for a crate.

compiler/rustc_session/src/options.rs

@@ -2482,6 +2482,8 @@ written to standard error output)"),
                   by the linker"),
     split_lto_unit: Option<bool> = (None, parse_opt_bool, [TRACKED],
         "enable LTO unit splitting (default: no)"),
+    split_metadata: bool = (false, parse_bool, [TRACKED],
+        "split metadata out of libraries into .rmeta files"),
     src_hash_algorithm: Option<SourceFileHashAlgorithm> = (None, parse_src_file_hash, [TRACKED],
         "hash algorithm of source files in debug info (`md5`, `sha1`, or `sha256`)"),
     #[rustc_lint_opt_deny_field_access("use `Session::stack_protector` instead of this field")]

src/doc/unstable-book/src/compiler-flags/split_metadata.md

@@ -0,0 +1,3 @@
+## `split-metadata`
+
+This option instructs `rustc` to only include a stub metadata section in `rlib` and `dylib` crate types instead of full metadata, to reduce their size on disk. You will probably want to combine this option with `--emit=metadata` to produce the full metadata into a separate `.rmeta` file.

tests/run-make/split-metadata/dep1.rs

@@ -0,0 +1 @@
+pub fn func_dep1() {}

tests/run-make/split-metadata/foo.rs

@@ -0,0 +1,5 @@
+extern crate dep1;
+
+fn main() {
+    dep1::func_dep1();
+}