Support -Zembed-source in debuginfo

  - honour the session DWARF version, enabling v5 when embed-source is requested and plumb the flag through the debug context
  - attach embedded source bytes to line table FileInfo records alongside existing MD5 hashes
  - stop excluding rustc’s embed-source-dwarf run-make test so CG_CLIF exercises the feature

Author: kehrazy

scripts/test_rustc_tests.sh

@@ -60,7 +60,6 @@ rm tests/ui/asm/global-asm-mono-sym-fn.rs # same
 rm tests/ui/asm/naked-asm-mono-sym-fn.rs # same
 rm tests/ui/asm/x86_64/goto.rs # inline asm labels not supported
 rm tests/ui/simd/simd-bitmask-notpow2.rs # non-pow-of-2 simd vector sizes
-rm -r tests/run-make/embed-source-dwarf # embedding sources in debuginfo
 rm -r tests/run-make/used-proc-macro # used(linker) isn't supported yet
 rm tests/ui/linking/no-gc-encapsulation-symbols.rs # same
 rm tests/ui/attributes/fn-align-dyn.rs # per-function alignment not supported

src/debuginfo/line_info.rs

@@ -6,9 +6,7 @@ use std::path::{Component, Path};
 use cranelift_codegen::MachSrcLoc;
 use cranelift_codegen::binemit::CodeOffset;
 use gimli::write::{AttributeValue, FileId, FileInfo, LineProgram, LineString, LineStringTable};
-use rustc_span::{
-    FileName, Pos, SourceFile, SourceFileAndLine, SourceFileHash, SourceFileHashAlgorithm, hygiene,
-};
+use rustc_span::{FileName, Pos, SourceFile, SourceFileAndLine, SourceFileHashAlgorithm, hygiene};
 
 use crate::debuginfo::FunctionDebugContext;
 use crate::debuginfo::emit::address_for_func;
@@ -44,21 +42,27 @@ fn osstr_as_utf8_bytes(path: &OsStr) -> &[u8] {
     }
 }
 
-const MD5_LEN: usize = 16;
-
-fn make_file_info(hash: SourceFileHash) -> Option<FileInfo> {
-    if hash.kind == SourceFileHashAlgorithm::Md5 {
-        let mut buf = [0u8; MD5_LEN];
-        buf.copy_from_slice(hash.hash_bytes());
-        Some(FileInfo {
-            timestamp: 0,
-            size: 0,
-            md5: buf,
-            source: None, // FIXME implement -Zembed-source
-        })
-    } else {
-        None
+fn make_file_info(source_file: &SourceFile, embed_source: bool) -> Option<FileInfo> {
+    let has_md5 = source_file.src_hash.kind == SourceFileHashAlgorithm::Md5;
+    let has_source = embed_source && source_file.src.is_some();
+
+    if !has_md5 && !has_source {
+        return None;
+    }
+
+    let mut info = FileInfo::default();
+
+    if has_md5 {
+        info.md5.copy_from_slice(source_file.src_hash.hash_bytes());
     }
+
+    if embed_source {
+        if let Some(src) = &source_file.src {
+            info.source = Some(LineString::String(src.as_bytes().to_vec()));
+        }
+    }
+
+    Some(info)
 }
 
 impl DebugContext {
@@ -105,9 +109,10 @@ impl DebugContext {
                     let file_name =
                         LineString::new(file_name, line_program.encoding(), line_strings);
 
-                    let info = make_file_info(source_file.src_hash);
+                    let info = make_file_info(source_file, self.embed_source);
 
-                    line_program.file_has_md5 &= info.is_some();
+                    let has_md5 = source_file.src_hash.kind == SourceFileHashAlgorithm::Md5;
+                    line_program.file_has_md5 &= has_md5;
                     line_program.add_file(file_name, dir_id, info)
                 }
                 filename => {

src/debuginfo/mod.rs

@@ -45,6 +45,7 @@ pub(crate) struct DebugContext {
     array_size_type: UnitEntryId,
 
     filename_display_preference: FileNameDisplayPreference,
+    embed_source: bool,
 }
 
 pub(crate) struct FunctionDebugContext {
@@ -67,21 +68,24 @@ impl DebugContext {
             return None;
         }
 
+        let mut requested_dwarf_version = tcx.sess.dwarf_version();
+        if tcx.sess.target.is_like_darwin && requested_dwarf_version > 4 {
+            // Apple’s shipped debuggers still expect DWARF <= 4 by default.
+            // Stay on v4 unless the user explicitly opts into a feature that
+            // only works with v5 (e.g. -Zembed-source).
+            if !tcx.sess.opts.unstable_opts.embed_source {
+                requested_dwarf_version = 4;
+            }
+        }
+
         let encoding = Encoding {
             format: Format::Dwarf32,
-            // FIXME this should be configurable
-            // macOS doesn't seem to support DWARF > 3
-            // 5 version is required for md5 file hash
-            version: if tcx.sess.target.is_like_darwin {
-                3
-            } else {
-                // FIXME change to version 5 once the gdb and lldb shipping with the latest debian
-                // support it.
-                4
-            },
+            version: requested_dwarf_version as u16,
             address_size: isa.frontend_config().pointer_bytes(),
         };
 
+        let embed_source = tcx.sess.opts.unstable_opts.embed_source && encoding.version >= 5;
+
         let endian = match isa.endianness() {
             Endianness::Little => RunTimeEndian::Little,
             Endianness::Big => RunTimeEndian::Big,
@@ -125,6 +129,9 @@ impl DebugContext {
             file_info,
         );
         line_program.file_has_md5 = file_has_md5;
+        if embed_source {
+            line_program.file_has_source = true;
+        }
 
         dwarf.unit.line_program = line_program;
 
@@ -169,6 +176,7 @@ impl DebugContext {
             namespace_map: DefIdMap::default(),
             array_size_type,
             filename_display_preference,
+            embed_source,
         })
     }