add unstable support for outputting file checksums for use in cargo

Cargo.lock

@@ -199,6 +199,12 @@ dependencies = [
  "object 0.36.4",
 ]
 
+[[package]]
+name = "arrayref"
+version = "0.3.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6b4930d2cb77ce62f89ee5d5289b4ac049559b1c45539271f5ed4fdc7db34545"
+
 [[package]]
 name = "arrayvec"
 version = "0.7.6"
@@ -262,6 +268,19 @@ version = "2.6.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b048fb63fd8b5923fc5aa7b340d8e156aec7ec02f0c78fa8a6ddc2613f6f71de"
 
+[[package]]
+name = "blake3"
+version = "1.5.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3d08263faac5cde2a4d52b513dadb80846023aade56fcd8fc99ba73ba8050e92"
+dependencies = [
+ "arrayref",
+ "arrayvec",
+ "cc",
+ "cfg-if",
+ "constant_time_eq",
+]
+
 [[package]]
 name = "block-buffer"
 version = "0.10.4"
@@ -719,6 +738,12 @@ dependencies = [
  "windows-sys 0.52.0",
 ]
 
+[[package]]
+name = "constant_time_eq"
+version = "0.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f7144d30dcf0fafbce74250a3963025d8d52177934239851c917d29f1df280c2"
+
 [[package]]
 name = "core-foundation-sys"
 version = "0.8.7"
@@ -4374,6 +4399,7 @@ dependencies = [
 name = "rustc_span"
 version = "0.0.0"
 dependencies = [
+ "blake3",
  "derive-where",
  "indexmap",
  "itoa",

compiler/rustc_codegen_llvm/src/debuginfo/metadata.rs

@@ -630,6 +630,7 @@ fn alloc_new_file_metadata<'ll>(
             rustc_span::SourceFileHashAlgorithm::Md5 => llvm::ChecksumKind::MD5,
             rustc_span::SourceFileHashAlgorithm::Sha1 => llvm::ChecksumKind::SHA1,
             rustc_span::SourceFileHashAlgorithm::Sha256 => llvm::ChecksumKind::SHA256,
+            rustc_span::SourceFileHashAlgorithm::Blake3 => llvm::ChecksumKind::None,
         };
         let hash_value = hex_encode(source_file.src_hash.hash_bytes());
 

compiler/rustc_interface/src/interface.rs

@@ -389,12 +389,13 @@ pub fn run_compiler<R: Send>(config: Config, f: impl FnOnce(&Compiler) -> R + Se
     let file_loader = config.file_loader.unwrap_or_else(|| Box::new(RealFileLoader));
     let path_mapping = config.opts.file_path_mapping();
     let hash_kind = config.opts.unstable_opts.src_hash_algorithm(&target);
+    let checksum_hash_kind = config.opts.unstable_opts.checksum_hash_algorithm();
 
     util::run_in_thread_pool_with_globals(
         &early_dcx,
         config.opts.edition,
         config.opts.unstable_opts.threads,
-        SourceMapInputs { file_loader, path_mapping, hash_kind },
+        SourceMapInputs { file_loader, path_mapping, hash_kind, checksum_hash_kind },
         |current_gcx| {
             // The previous `early_dcx` can't be reused here because it doesn't
             // impl `Send`. Creating a new one is fine.

compiler/rustc_interface/src/lib.rs

@@ -1,6 +1,7 @@
 // tidy-alphabetical-start
 #![feature(decl_macro)]
 #![feature(file_buffered)]
+#![feature(iter_intersperse)]
 #![feature(let_chains)]
 #![feature(try_blocks)]
 #![warn(unreachable_pub)]

compiler/rustc_interface/src/passes.rs

@@ -32,8 +32,8 @@
 use rustc_session::output::{collect_crate_types, filename_for_input, find_crate_name};
 use rustc_session::search_paths::PathKind;
 use rustc_session::{Limit, Session};
-use rustc_span::FileName;
 use rustc_span::symbol::{Symbol, sym};
+use rustc_span::{FileName, SourceFileHash, SourceFileHashAlgorithm};
 use rustc_target::spec::PanicStrategy;
 use rustc_trait_selection::traits;
 use tracing::{info, instrument};
@@ -417,15 +417,23 @@ fn write_out_deps(tcx: TyCtxt<'_>, outputs: &OutputFilenames, out_filenames: &[P
     let result: io::Result<()> = try {
         // Build a list of files used to compile the output and
         // write Makefile-compatible dependency rules
-        let mut files: Vec<String> = sess
+        let mut files: Vec<(String, u64, Option<SourceFileHash>)> = sess
             .source_map()
             .files()
             .iter()
             .filter(|fmap| fmap.is_real_file())
             .filter(|fmap| !fmap.is_imported())
-            .map(|fmap| escape_dep_filename(&fmap.name.prefer_local().to_string()))
+            .map(|fmap| {
+                (
+                    escape_dep_filename(&fmap.name.prefer_local().to_string()),
+                    fmap.source_len.0 as u64,
+                    fmap.checksum_hash,
+                )
+            })
             .collect();
 
+        let checksum_hash_algo = sess.opts.unstable_opts.checksum_hash_algorithm;
+
         // Account for explicitly marked-to-track files
         // (e.g. accessed in proc macros).
         let file_depinfo = sess.psess.file_depinfo.borrow();
@@ -437,22 +445,58 @@ fn write_out_deps(tcx: TyCtxt<'_>, outputs: &OutputFilenames, out_filenames: &[P
 
         // The entries will be used to declare dependencies between files in a
         // Makefile-like output, so the iteration order does not matter.
+        fn hash_iter_files<P: AsRef<Path>>(
+            it: impl Iterator<Item = P>,
+            checksum_hash_algo: Option<SourceFileHashAlgorithm>,
+        ) -> impl Iterator<Item = (P, u64, Option<SourceFileHash>)> {
+            it.map(move |path| {
+                match checksum_hash_algo.and_then(|algo| {
+                    fs::File::open(path.as_ref())
+                        .and_then(|mut file| {
+                            SourceFileHash::new(algo, &mut file).map(|h| (file, h))
+                        })
+                        .and_then(|(file, h)| file.metadata().map(|m| (m.len(), h)))
+                        .map_err(|e| {
+                            tracing::error!(
+                                "failed to compute checksum, omitting it from dep-info {} {e}",
+                                path.as_ref().display()
+                            )
+                        })
+                        .ok()
+                }) {
+                    Some((file_len, checksum)) => (path, file_len, Some(checksum)),
+                    None => (path, 0, None),
+                }
+            })
+        }
+
         #[allow(rustc::potential_query_instability)]
-        let extra_tracked_files =
+        let extra_tracked_files = hash_iter_files(
-            file_depinfo.iter().map(|path_sym| normalize_path(PathBuf::from(path_sym.as_str())));
+            file_depinfo.iter().map(|path_sym| normalize_path(PathBuf::from(path_sym.as_str()))),
+            checksum_hash_algo,
+        );
         files.extend(extra_tracked_files);
 
         // We also need to track used PGO profile files
         if let Some(ref profile_instr) = sess.opts.cg.profile_use {
-            files.push(normalize_path(profile_instr.as_path().to_path_buf()));
+            files.extend(hash_iter_files(
+                iter::once(normalize_path(profile_instr.as_path().to_path_buf())),
+                checksum_hash_algo,
+            ));
         }
         if let Some(ref profile_sample) = sess.opts.unstable_opts.profile_sample_use {
-            files.push(normalize_path(profile_sample.as_path().to_path_buf()));
+            files.extend(hash_iter_files(
+                iter::once(normalize_path(profile_sample.as_path().to_path_buf())),
+                checksum_hash_algo,
+            ));
         }
 
         // Debugger visualizer files
         for debugger_visualizer in tcx.debugger_visualizers(LOCAL_CRATE) {
-            files.push(normalize_path(debugger_visualizer.path.clone().unwrap()));
+            files.extend(hash_iter_files(
+                iter::once(normalize_path(debugger_visualizer.path.clone().unwrap())),
+                checksum_hash_algo,
+            ));
         }
 
         if sess.binary_dep_depinfo() {
@@ -460,33 +504,54 @@ fn write_out_deps(tcx: TyCtxt<'_>, outputs: &OutputFilenames, out_filenames: &[P
                 if backend.contains('.') {
                     // If the backend name contain a `.`, it is the path to an external dynamic
                     // library. If not, it is not a path.
-                    files.push(backend.to_string());
+                    files.extend(hash_iter_files(
+                        iter::once(backend.to_string()),
+                        checksum_hash_algo,
+                    ));
                 }
             }
 
             for &cnum in tcx.crates(()) {
                 let source = tcx.used_crate_source(cnum);
                 if let Some((path, _)) = &source.dylib {
-                    files.push(escape_dep_filename(&path.display().to_string()));
+                    files.extend(hash_iter_files(
+                        iter::once(escape_dep_filename(&path.display().to_string())),
+                        checksum_hash_algo,
+                    ));
                 }
                 if let Some((path, _)) = &source.rlib {
-                    files.push(escape_dep_filename(&path.display().to_string()));
+                    files.extend(hash_iter_files(
+                        iter::once(escape_dep_filename(&path.display().to_string())),
+                        checksum_hash_algo,
+                    ));
                 }
                 if let Some((path, _)) = &source.rmeta {
-                    files.push(escape_dep_filename(&path.display().to_string()));
+                    files.extend(hash_iter_files(
+                        iter::once(escape_dep_filename(&path.display().to_string())),
+                        checksum_hash_algo,
+                    ));
                 }
             }
         }
 
         let write_deps_to_file = |file: &mut dyn Write| -> io::Result<()> {
             for path in out_filenames {
-                writeln!(file, "{}: {}\n", path.display(), files.join(" "))?;
+                writeln!(
+                    file,
+                    "{}: {}\n",
+                    path.display(),
+                    files
+                        .iter()
+                        .map(|(path, _file_len, _checksum_hash_algo)| path.as_str())
+                        .intersperse(" ")
+                        .collect::<String>()
+                )?;
             }
 
             // Emit a fake target for each input file to the compilation. This
             // prevents `make` from spitting out an error if a file is later
             // deleted. For more info see #28735
-            for path in files {
+            for (path, _file_len, _checksum_hash_algo) in &files {
                 writeln!(file, "{path}:")?;
             }
 
@@ -510,6 +575,18 @@ fn write_out_deps(tcx: TyCtxt<'_>, outputs: &OutputFilenames, out_filenames: &[P
                 }
             }
 
+            // If caller requested this information, add special comments about source file checksums.
+            // These are not necessarily the same checksums as was used in the debug files.
+            if sess.opts.unstable_opts.checksum_hash_algorithm().is_some() {
+                for (path, file_len, checksum_hash) in
+                    files.iter().filter_map(|(path, file_len, hash_algo)| {
+                        hash_algo.map(|hash_algo| (path, file_len, hash_algo))
+                    })
+                {
+                    writeln!(file, "# checksum:{checksum_hash} file_len:{file_len} {path}")?;
+                }
+            }
+
             Ok(())
         };
 

compiler/rustc_interface/src/tests.rs

@@ -44,10 +44,12 @@ fn sess_and_cfg<F>(args: &[&'static str], f: F)
     let sysroot = filesearch::materialize_sysroot(sessopts.maybe_sysroot.clone());
     let target = rustc_session::config::build_target_config(&early_dcx, &sessopts, &sysroot);
     let hash_kind = sessopts.unstable_opts.src_hash_algorithm(&target);
+    let checksum_hash_kind = sessopts.unstable_opts.checksum_hash_algorithm();
     let sm_inputs = Some(SourceMapInputs {
         file_loader: Box::new(RealFileLoader) as _,
         path_mapping: sessopts.file_path_mapping(),
         hash_kind,
+        checksum_hash_kind,
     });
 
     rustc_span::create_session_globals_then(DEFAULT_EDITION, sm_inputs, || {

compiler/rustc_metadata/src/rmeta/decoder.rs

@@ -1702,6 +1702,7 @@ fn filter<'a>(sess: &Session, path: Option<&'a Path>) -> Option<&'a Path> {
                 let rustc_span::SourceFile {
                     mut name,
                     src_hash,
+                    checksum_hash,
                     start_pos: original_start_pos,
                     source_len,
                     lines,
@@ -1752,6 +1753,7 @@ fn filter<'a>(sess: &Session, path: Option<&'a Path>) -> Option<&'a Path> {
                 let local_version = sess.source_map().new_imported_source_file(
                     name,
                     src_hash,
+                    checksum_hash,
                     stable_id,
                     source_len.to_u32(),
                     self.cnum,

compiler/rustc_query_system/src/ich/impls_syntax.rs

@@ -68,6 +68,8 @@ fn hash_stable(&self, hcx: &mut StableHashingContext<'a>, hasher: &mut StableHas
             // Do not hash the source as it is not encoded
             src: _,
             ref src_hash,
+            // Already includes src_hash, this is redundant
+            checksum_hash: _,
             external_src: _,
             start_pos: _,
             source_len: _,

compiler/rustc_session/src/config.rs

@@ -1242,6 +1242,10 @@ pub fn src_hash_algorithm(&self, target: &Target) -> SourceFileHashAlgorithm {
             }
         })
     }
+
+    pub fn checksum_hash_algorithm(&self) -> Option<SourceFileHashAlgorithm> {
+        self.checksum_hash_algorithm
+    }
 }
 
 // The type of entry function, so users can have their own entry functions

compiler/rustc_session/src/options.rs

@@ -418,7 +418,8 @@ mod desc {
         "one of: `legacy`, `v0` (RFC 2603), or `hashed`";
     pub(crate) const parse_opt_symbol_visibility: &str =
         "one of: `hidden`, `protected`, or `interposable`";
-    pub(crate) const parse_src_file_hash: &str = "either `md5` or `sha1`";
+    pub(crate) const parse_cargo_src_file_hash: &str = "one of `md5`, `sha1`, or `sha256`";
+    pub(crate) const parse_src_file_hash: &str = "one of `md5`, `sha1`, or `sha256`";
     pub(crate) const parse_relocation_model: &str =
         "one of supported relocation models (`rustc --print relocation-models`)";
     pub(crate) const parse_code_model: &str =
@@ -1288,6 +1289,23 @@ pub(crate) fn parse_src_file_hash(
         true
     }
 
+    pub(crate) fn parse_cargo_src_file_hash(
+        slot: &mut Option<SourceFileHashAlgorithm>,
+        v: Option<&str>,
+    ) -> bool {
+        match v.and_then(|s| SourceFileHashAlgorithm::from_str(s).ok()) {
+            Some(hash_kind) => {
+                if hash_kind.supported_in_cargo() {
+                    *slot = Some(hash_kind);
+                } else {
+                    return false;
+                }
+            }
+            _ => return false,
+        }
+        true
+    }
+
     pub(crate) fn parse_target_feature(slot: &mut String, v: Option<&str>) -> bool {
         match v {
             Some(s) => {
@@ -1688,6 +1706,8 @@ pub(crate) fn parse_mir_include_spans(slot: &mut MirIncludeSpans, v: Option<&str
         "instrument control-flow architecture protection"),
     check_cfg_all_expected: bool = (false, parse_bool, [UNTRACKED],
         "show all expected values in check-cfg diagnostics (default: no)"),
+    checksum_hash_algorithm: Option<SourceFileHashAlgorithm> = (None, parse_cargo_src_file_hash, [TRACKED],
+        "hash algorithm of source files used to check freshness in cargo (`sha256`)"),
     codegen_backend: Option<String> = (None, parse_opt_string, [TRACKED],
         "the backend to use"),
     combine_cgu: bool = (false, parse_bool, [TRACKED],

compiler/rustc_span/Cargo.toml

@@ -5,6 +5,7 @@ edition = "2021"
 
 [dependencies]
 # tidy-alphabetical-start
+blake3 = "1.5.2"
 derive-where = "1.2.7"
 indexmap = { version = "2.0.0" }
 itoa = "1.0"

compiler/rustc_span/src/lib.rs

@@ -75,7 +75,9 @@
 
 use std::borrow::Cow;
 use std::cmp::{self, Ordering};
+use std::fmt::Display;
 use std::hash::Hash;
+use std::io::{self, Read};
 use std::ops::{Add, Range, Sub};
 use std::path::{Path, PathBuf};
 use std::str::FromStr;
@@ -1395,6 +1397,27 @@ pub enum SourceFileHashAlgorithm {
     Md5,
     Sha1,
     Sha256,
+    Blake3,
+}
+
+impl SourceFileHashAlgorithm {
+    pub fn supported_in_cargo(&self) -> bool {
+        match self {
+            Self::Md5 | Self::Sha1 => false,
+            Self::Sha256 | Self::Blake3 => true,
+        }
+    }
+}
+
+impl Display for SourceFileHashAlgorithm {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        f.write_str(match self {
+            Self::Md5 => "md5",
+            Self::Sha1 => "sha1",
+            Self::Sha256 => "sha256",
+            Self::Blake3 => "blake3",
+        })
+    }
 }
 
 impl FromStr for SourceFileHashAlgorithm {
@@ -1405,12 +1428,13 @@ fn from_str(s: &str) -> Result<SourceFileHashAlgorithm, ()> {
             "md5" => Ok(SourceFileHashAlgorithm::Md5),
             "sha1" => Ok(SourceFileHashAlgorithm::Sha1),
             "sha256" => Ok(SourceFileHashAlgorithm::Sha256),
+            "blake3" => Ok(SourceFileHashAlgorithm::Blake3),
             _ => Err(()),
         }
     }
 }
 
-/// The hash of the on-disk source file used for debug info.
+/// The hash of the on-disk source file used for debug info and cargo freshness checks.
 #[derive(Copy, Clone, PartialEq, Eq, Debug, Hash)]
 #[derive(HashStable_Generic, Encodable, Decodable)]
 pub struct SourceFileHash {
@@ -1418,12 +1442,22 @@ pub struct SourceFileHash {
     value: [u8; 32],
 }
 
+impl Display for SourceFileHash {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "{}=", self.kind)?;
+        for byte in self.value[0..self.hash_len()].into_iter() {
+            write!(f, "{byte:02x}")?;
+        }
+        Ok(())
+    }
+}
+
 impl SourceFileHash {
-    pub fn new(kind: SourceFileHashAlgorithm, src: &str) -> SourceFileHash {
+    pub fn new_in_memory(kind: SourceFileHashAlgorithm, src: impl AsRef<[u8]>) -> SourceFileHash {
         let mut hash = SourceFileHash { kind, value: Default::default() };
         let len = hash.hash_len();
         let value = &mut hash.value[..len];
-        let data = src.as_bytes();
+        let data = src.as_ref();
         match kind {
             SourceFileHashAlgorithm::Md5 => {
                 value.copy_from_slice(&Md5::digest(data));
@@ -1434,13 +1468,94 @@ pub fn new(kind: SourceFileHashAlgorithm, src: &str) -> SourceFileHash {
             SourceFileHashAlgorithm::Sha256 => {
                 value.copy_from_slice(&Sha256::digest(data));
             }
-        }
+            SourceFileHashAlgorithm::Blake3 => value.copy_from_slice(blake3::hash(data).as_bytes()),
+        };
         hash
     }
 
+    pub fn new(kind: SourceFileHashAlgorithm, src: impl Read) -> Result<SourceFileHash, io::Error> {
+        let mut hash = SourceFileHash { kind, value: Default::default() };
+        let len = hash.hash_len();
+        let value = &mut hash.value[..len];
+        // Buffer size is the recommended amount to fully leverage SIMD instructions on AVX-512 as per
+        // blake3 documentation.
+        let mut buf = vec![0; 16 * 1024];
+
+        fn digest<T>(
+            mut hasher: T,
+            mut update: impl FnMut(&mut T, &[u8]),
+            finish: impl FnOnce(T, &mut [u8]),
+            mut src: impl Read,
+            buf: &mut [u8],
+            value: &mut [u8],
+        ) -> Result<(), io::Error> {
+            loop {
+                let bytes_read = src.read(buf)?;
+                if bytes_read == 0 {
+                    break;
+                }
+                update(&mut hasher, &buf[0..bytes_read]);
+            }
+            finish(hasher, value);
+            Ok(())
+        }
+
+        match kind {
+            SourceFileHashAlgorithm::Sha256 => {
+                digest(
+                    Sha256::new(),
+                    |h, b| {
+                        h.update(b);
+                    },
+                    |h, out| out.copy_from_slice(&h.finalize()),
+                    src,
+                    &mut buf,
+                    value,
+                )?;
+            }
+            SourceFileHashAlgorithm::Sha1 => {
+                digest(
+                    Sha1::new(),
+                    |h, b| {
+                        h.update(b);
+                    },
+                    |h, out| out.copy_from_slice(&h.finalize()),
+                    src,
+                    &mut buf,
+                    value,
+                )?;
+            }
+            SourceFileHashAlgorithm::Md5 => {
+                digest(
+                    Md5::new(),
+                    |h, b| {
+                        h.update(b);
+                    },
+                    |h, out| out.copy_from_slice(&h.finalize()),
+                    src,
+                    &mut buf,
+                    value,
+                )?;
+            }
+            SourceFileHashAlgorithm::Blake3 => {
+                digest(
+                    blake3::Hasher::new(),
+                    |h, b| {
+                        h.update(b);
+                    },
+                    |h, out| out.copy_from_slice(h.finalize().as_bytes()),
+                    src,
+                    &mut buf,
+                    value,
+                )?;
+            }
+        }
+        Ok(hash)
+    }
+
     /// Check if the stored hash matches the hash of the string.
     pub fn matches(&self, src: &str) -> bool {
-        Self::new(self.kind, src) == *self
+        Self::new_in_memory(self.kind, src.as_bytes()) == *self
     }
 
     /// The bytes of the hash.
@@ -1453,7 +1568,7 @@ fn hash_len(&self) -> usize {
         match self.kind {
             SourceFileHashAlgorithm::Md5 => 16,
             SourceFileHashAlgorithm::Sha1 => 20,
-            SourceFileHashAlgorithm::Sha256 => 32,
+            SourceFileHashAlgorithm::Sha256 | SourceFileHashAlgorithm::Blake3 => 32,
         }
     }
 }
@@ -1509,6 +1624,10 @@ pub struct SourceFile {
     pub src: Option<Lrc<String>>,
     /// The source code's hash.
     pub src_hash: SourceFileHash,
+    /// Used to enable cargo to use checksums to check if a crate is fresh rather
+    /// than mtimes. This might be the same as `src_hash`, and if the requested algorithm
+    /// is identical we won't compute it twice.
+    pub checksum_hash: Option<SourceFileHash>,
     /// The external source code (used for external crates, which will have a `None`
     /// value as `self.src`.
     pub external_src: FreezeLock<ExternalSource>,
@@ -1536,6 +1655,7 @@ fn clone(&self) -> Self {
             name: self.name.clone(),
             src: self.src.clone(),
             src_hash: self.src_hash,
+            checksum_hash: self.checksum_hash,
             external_src: self.external_src.clone(),
             start_pos: self.start_pos,
             source_len: self.source_len,
@@ -1552,6 +1672,7 @@ impl<S: SpanEncoder> Encodable<S> for SourceFile {
     fn encode(&self, s: &mut S) {
         self.name.encode(s);
         self.src_hash.encode(s);
+        self.checksum_hash.encode(s);
         // Do not encode `start_pos` as it's global state for this session.
         self.source_len.encode(s);
 
@@ -1625,6 +1746,7 @@ impl<D: SpanDecoder> Decodable<D> for SourceFile {
     fn decode(d: &mut D) -> SourceFile {
         let name: FileName = Decodable::decode(d);
         let src_hash: SourceFileHash = Decodable::decode(d);
+        let checksum_hash: Option<SourceFileHash> = Decodable::decode(d);
         let source_len: RelativeBytePos = Decodable::decode(d);
         let lines = {
             let num_lines: u32 = Decodable::decode(d);
@@ -1650,6 +1772,7 @@ fn decode(d: &mut D) -> SourceFile {
             source_len,
             src: None,
             src_hash,
+            checksum_hash,
             // Unused - the metadata decoder will construct
             // a new SourceFile, filling in `external_src` properly
             external_src: FreezeLock::frozen(ExternalSource::Unneeded),
@@ -1733,9 +1856,17 @@ pub fn new(
         name: FileName,
         mut src: String,
         hash_kind: SourceFileHashAlgorithm,
+        checksum_hash_kind: Option<SourceFileHashAlgorithm>,
     ) -> Result<Self, OffsetOverflowError> {
         // Compute the file hash before any normalization.
-        let src_hash = SourceFileHash::new(hash_kind, &src);
+        let src_hash = SourceFileHash::new_in_memory(hash_kind, src.as_bytes());
+        let checksum_hash = checksum_hash_kind.map(|checksum_hash_kind| {
+            if checksum_hash_kind == hash_kind {
+                src_hash
+            } else {
+                SourceFileHash::new_in_memory(checksum_hash_kind, src.as_bytes())
+            }
+        });
         let normalized_pos = normalize_src(&mut src);
 
         let stable_id = StableSourceFileId::from_filename_in_current_crate(&name);
@@ -1748,6 +1879,7 @@ pub fn new(
             name,
             src: Some(Lrc::new(src)),
             src_hash,
+            checksum_hash,
             external_src: FreezeLock::frozen(ExternalSource::Unneeded),
             start_pos: BytePos::from_u32(0),
             source_len: RelativeBytePos::from_u32(source_len),

compiler/rustc_span/src/source_map.rs

@@ -175,6 +175,7 @@ pub struct SourceMapInputs {
     pub file_loader: Box<dyn FileLoader + Send + Sync>,
     pub path_mapping: FilePathMapping,
     pub hash_kind: SourceFileHashAlgorithm,
+    pub checksum_hash_kind: Option<SourceFileHashAlgorithm>,
 }
 
 pub struct SourceMap {
@@ -187,6 +188,12 @@ pub struct SourceMap {
 
     /// The algorithm used for hashing the contents of each source file.
     hash_kind: SourceFileHashAlgorithm,
+
+    /// Similar to `hash_kind`, however this algorithm is used for checksums to determine if a crate is fresh.
+    /// `cargo` is the primary user of these.
+    ///
+    /// If this is equal to `hash_kind` then the checksum won't be computed twice.
+    checksum_hash_kind: Option<SourceFileHashAlgorithm>,
 }
 
 impl SourceMap {
@@ -195,17 +202,19 @@ pub fn new(path_mapping: FilePathMapping) -> SourceMap {
             file_loader: Box::new(RealFileLoader),
             path_mapping,
             hash_kind: SourceFileHashAlgorithm::Md5,
+            checksum_hash_kind: None,
         })
     }
 
     pub fn with_inputs(
-        SourceMapInputs { file_loader, path_mapping, hash_kind }: SourceMapInputs,
+        SourceMapInputs { file_loader, path_mapping, hash_kind, checksum_hash_kind }: SourceMapInputs,
     ) -> SourceMap {
         SourceMap {
             files: Default::default(),
             file_loader: IntoDynSyncSend(file_loader),
             path_mapping,
             hash_kind,
+            checksum_hash_kind,
         }
     }
 
@@ -307,7 +316,8 @@ fn try_new_source_file(
         match self.source_file_by_stable_id(stable_id) {
             Some(lrc_sf) => Ok(lrc_sf),
             None => {
-                let source_file = SourceFile::new(filename, src, self.hash_kind)?;
+                let source_file =
+                    SourceFile::new(filename, src, self.hash_kind, self.checksum_hash_kind)?;
 
                 // Let's make sure the file_id we generated above actually matches
                 // the ID we generate for the SourceFile we just created.
@@ -326,6 +336,7 @@ pub fn new_imported_source_file(
         &self,
         filename: FileName,
         src_hash: SourceFileHash,
+        checksum_hash: Option<SourceFileHash>,
         stable_id: StableSourceFileId,
         source_len: u32,
         cnum: CrateNum,
@@ -340,6 +351,7 @@ pub fn new_imported_source_file(
             name: filename,
             src: None,
             src_hash,
+            checksum_hash,
             external_src: FreezeLock::new(ExternalSource::Foreign {
                 kind: ExternalSourceKind::AbsentOk,
                 metadata_index,

compiler/rustc_span/src/source_map/tests.rs

@@ -229,6 +229,7 @@ fn t10() {
     let SourceFile {
         name,
         src_hash,
+        checksum_hash,
         source_len,
         lines,
         multibyte_chars,
@@ -240,6 +241,7 @@ fn t10() {
     let imported_src_file = sm.new_imported_source_file(
         name,
         src_hash,
+        checksum_hash,
         stable_id,
         source_len.to_u32(),
         CrateNum::ZERO,

compiler/rustc_span/src/tests.rs

@@ -3,9 +3,13 @@
 #[test]
 fn test_lookup_line() {
     let source = "abcdefghijklm\nabcdefghij\n...".to_owned();
-    let mut sf =
+    let mut sf = SourceFile::new(
-        SourceFile::new(FileName::Anon(Hash64::ZERO), source, SourceFileHashAlgorithm::Sha256)
+        FileName::Anon(Hash64::ZERO),
-            .unwrap();
+        source,
+        SourceFileHashAlgorithm::Sha256,
+        Some(SourceFileHashAlgorithm::Sha256),
+    )
+    .unwrap();
     sf.start_pos = BytePos(3);
     assert_eq!(sf.lines(), &[RelativeBytePos(0), RelativeBytePos(14), RelativeBytePos(25)]);
 

src/tools/tidy/src/deps.rs

@@ -88,7 +88,10 @@
 const EXCEPTIONS: ExceptionList = &[
     // tidy-alphabetical-start
     ("ar_archive_writer", "Apache-2.0 WITH LLVM-exception"), // rustc
+    ("arrayref", "BSD-2-Clause"),                            // rustc
+    ("blake3", "CC0-1.0 OR Apache-2.0 OR Apache-2.0 WITH LLVM-exception"),  // rustc
     ("colored", "MPL-2.0"),                                  // rustfmt
+    ("constant_time_eq", "CC0-1.0 OR MIT-0 OR Apache-2.0"),  // rustc
     ("dissimilar", "Apache-2.0"),                            // rustdoc, rustc_lexer (few tests) via expect-test, (dev deps)
     ("fluent-langneg", "Apache-2.0"),                        // rustc (fluent translations)
     ("instant", "BSD-3-Clause"),                             // rustc_driver/tracing-subscriber/parking_lot
@@ -249,14 +252,17 @@
     "annotate-snippets",
     "anstyle",
     "ar_archive_writer",
+    "arrayref",
     "arrayvec",
     "autocfg",
     "bitflags",
+    "blake3",
     "block-buffer",
     "byteorder", // via ruzstd in object in thorin-dwp
     "cc",
     "cfg-if",
     "cfg_aliases",
+    "constant_time_eq",
     "cpufeatures",
     "crc32fast",
     "crossbeam-channel",