Auto merge of #98100 - bjorn3:use_object_for_bitcode_reading, r=wesleywiser

Use object instead of LLVM for reading bitcode from rlibs

Together with changes I plan to make as part of #97485 this will allow entirely removing usage of LLVM's archive reader and thus allow removing `archive_ro.rs` and `ArchiveWrapper.cpp`.

Author: bors

Cargo.lock

@@ -3334,6 +3334,7 @@ dependencies = [
  "libc",
  "libloading",
  "measureme",
+ "object 0.29.0",
  "rustc-demangle",
  "rustc_ast",
  "rustc_attr",

compiler/rustc_codegen_llvm/Cargo.toml

@@ -13,6 +13,7 @@ cstr = "0.2"
 libc = "0.2"
 libloading = "0.7.1"
 measureme = "10.0.0"
+object = { version = "0.29.0", default-features = false, features = ["std", "read_core", "archive", "coff", "elf", "macho", "pe"] }
 tracing = "0.1"
 rustc_middle = { path = "../rustc_middle" }
 rustc-demangle = "0.1.21"

compiler/rustc_codegen_llvm/src/back/lto.rs

@@ -1,15 +1,16 @@
 use crate::back::write::{
     self, save_temp_bitcode, to_llvm_opt_settings, with_llvm_pmb, DiagnosticHandlers,
 };
-use crate::llvm::archive_ro::ArchiveRO;
 use crate::llvm::{self, build_string, False, True};
 use crate::{llvm_util, LlvmCodegenBackend, ModuleLlvm};
+use object::read::archive::ArchiveFile;
 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, TargetMachineFactoryConfig};
 use rustc_codegen_ssa::traits::*;
 use rustc_codegen_ssa::{looks_like_rust_object_file, ModuleCodegen, ModuleKind};
 use rustc_data_structures::fx::FxHashMap;
+use rustc_data_structures::memmap::Mmap;
 use rustc_errors::{FatalError, Handler};
 use rustc_hir::def_id::LOCAL_CRATE;
 use rustc_middle::bug;
@@ -107,14 +108,24 @@ fn prepare_lto(
                     .extend(exported_symbols[&cnum].iter().filter_map(symbol_filter));
             }
 
-            let archive = ArchiveRO::open(path).expect("wanted an rlib");
+            let archive_data = unsafe {
+                Mmap::map(std::fs::File::open(&path).expect("couldn't open rlib"))
+                    .expect("couldn't map rlib")
+            };
+            let archive = ArchiveFile::parse(&*archive_data).expect("wanted an rlib");
             let obj_files = archive
-                .iter()
-                .filter_map(|child| child.ok().and_then(|c| c.name().map(|name| (name, c))))
+                .members()
+                .filter_map(|child| {
+                    child.ok().and_then(|c| {
+                        std::str::from_utf8(c.name()).ok().map(|name| (name.trim(), c))
+                    })
+                })
                 .filter(|&(name, _)| looks_like_rust_object_file(name));
             for (name, child) in obj_files {
                 info!("adding bitcode from {}", name);
-                match get_bitcode_slice_from_object_data(child.data()) {
+                match get_bitcode_slice_from_object_data(
+                    child.data(&*archive_data).expect("corrupt rlib"),
+                ) {
                     Ok(data) => {
                         let module = SerializedModule::FromRlib(data.to_vec());
                         upstream_modules.push((module, CString::new(name).unwrap()));

compiler/rustc_codegen_llvm/src/llvm/archive_ro.rs

@@ -83,17 +83,6 @@ impl<'a> Child<'a> {
             }
         }
     }
-
-    pub fn data(&self) -> &'a [u8] {
-        unsafe {
-            let mut data_len = 0;
-            let data_ptr = super::LLVMRustArchiveChildData(self.raw, &mut data_len);
-            if data_ptr.is_null() {
-                panic!("failed to read data from archive child");
-            }
-            slice::from_raw_parts(data_ptr as *const u8, data_len as usize)
-        }
-    }
 }
 
 impl<'a> Drop for Child<'a> {

compiler/rustc_codegen_llvm/src/llvm/ffi.rs

@@ -2389,7 +2389,6 @@ extern "C" {
         AIR: &ArchiveIterator<'a>,
     ) -> Option<&'a mut ArchiveChild<'a>>;
     pub fn LLVMRustArchiveChildName(ACR: &ArchiveChild<'_>, size: &mut size_t) -> *const c_char;
-    pub fn LLVMRustArchiveChildData(ACR: &ArchiveChild<'_>, size: &mut size_t) -> *const c_char;
     pub fn LLVMRustArchiveChildFree<'a>(ACR: &'a mut ArchiveChild<'a>);
     pub fn LLVMRustArchiveIteratorFree<'a>(AIR: &'a mut ArchiveIterator<'a>);
     pub fn LLVMRustDestroyArchive(AR: &'static mut Archive);

compiler/rustc_llvm/llvm-wrapper/ArchiveWrapper.cpp

@@ -154,19 +154,6 @@ LLVMRustArchiveChildName(LLVMRustArchiveChildConstRef Child, size_t *Size) {
   return Name.data();
 }
 
-extern "C" const char *LLVMRustArchiveChildData(LLVMRustArchiveChildRef Child,
-                                                size_t *Size) {
-  StringRef Buf;
-  Expected<StringRef> BufOrErr = Child->getBuffer();
-  if (!BufOrErr) {
-    LLVMRustSetLastError(toString(BufOrErr.takeError()).c_str());
-    return nullptr;
-  }
-  Buf = BufOrErr.get();
-  *Size = Buf.size();
-  return Buf.data();
-}
-
 extern "C" LLVMRustArchiveMemberRef
 LLVMRustArchiveMemberNew(char *Filename, char *Name,
                          LLVMRustArchiveChildRef Child) {