initial implementation of the darwin_objc unstable feature

6 files changed, 342 insertions, 8 deletions

diff --git a/compiler/rustc_codegen_llvm/src/base.rs b/compiler/rustc_codegen_llvm/src/base.rs
index 9cc5d8dbc21..978134cc32b 100644
--- a/compiler/rustc_codegen_llvm/src/base.rs
+++ b/compiler/rustc_codegen_llvm/src/base.rs
@@ -109,18 +109,36 @@ pub(crate) fn compile_codegen_unit(
                 attributes::apply_to_llfn(entry, llvm::AttributePlace::Function, &attrs);
             }
 
+            // Define Objective-C module info and module flags. Note, the module info will
+            // also be added to the `llvm.compiler.used` variable, created later.
+            //
+            // These are only necessary when we need the linker to do its Objective-C-specific
+            // magic. We could theoretically do it unconditionally, but at a slight cost to linker
+            // performance in the common case where it's unnecessary.
+            if !cx.objc_classrefs.borrow().is_empty() || !cx.objc_selrefs.borrow().is_empty() {
+                if cx.objc_abi_version() == 1 {
+                    cx.define_objc_module_info();
+                }
+                cx.add_objc_module_flags();
+            }
+
             // Finalize code coverage by injecting the coverage map. Note, the coverage map will
             // also be added to the `llvm.compiler.used` variable, created next.
             if cx.sess().instrument_coverage() {
                 cx.coverageinfo_finalize();
             }
 
-            // Create the llvm.used and llvm.compiler.used variables.
+            // Create the llvm.used variable.
             if !cx.used_statics.is_empty() {
                 cx.create_used_variable_impl(c"llvm.used", &cx.used_statics);
             }
-            if !cx.compiler_used_statics.is_empty() {
-                cx.create_used_variable_impl(c"llvm.compiler.used", &cx.compiler_used_statics);
+
+            // Create the llvm.compiler.used variable.
+            {
+                let compiler_used_statics = cx.compiler_used_statics.borrow();
+                if !compiler_used_statics.is_empty() {
+                    cx.create_used_variable_impl(c"llvm.compiler.used", &compiler_used_statics);
+                }
             }
 
             // Run replace-all-uses-with for statics that need it. This must
diff --git a/compiler/rustc_codegen_llvm/src/common.rs b/compiler/rustc_codegen_llvm/src/common.rs
index 11b79a7fe68..aa2df46329f 100644
--- a/compiler/rustc_codegen_llvm/src/common.rs
+++ b/compiler/rustc_codegen_llvm/src/common.rs
@@ -108,6 +108,10 @@ impl<'ll, CX: Borrow<SCx<'ll>>> GenericCx<'ll, CX> {
         bytes_in_context(self.llcx(), bytes)
     }
 
+    pub(crate) fn null_terminate_const_bytes(&self, bytes: &[u8]) -> &'ll Value {
+        null_terminate_bytes_in_context(self.llcx(), bytes)
+    }
+
     pub(crate) fn const_get_elt(&self, v: &'ll Value, idx: u64) -> &'ll Value {
         unsafe {
             let idx = c_uint::try_from(idx).expect("LLVMGetAggregateElement index overflow");
@@ -381,6 +385,16 @@ pub(crate) fn bytes_in_context<'ll>(llcx: &'ll llvm::Context, bytes: &[u8]) -> &
     }
 }
 
+pub(crate) fn null_terminate_bytes_in_context<'ll>(
+    llcx: &'ll llvm::Context,
+    bytes: &[u8],
+) -> &'ll Value {
+    unsafe {
+        let ptr = bytes.as_ptr() as *const c_char;
+        llvm::LLVMConstStringInContext2(llcx, ptr, bytes.len(), FALSE)
+    }
+}
+
 pub(crate) fn named_struct<'ll>(ty: &'ll Type, elts: &[&'ll Value]) -> &'ll Value {
     let len = c_uint::try_from(elts.len()).expect("LLVMConstStructInContext elements len overflow");
     unsafe { llvm::LLVMConstNamedStruct(ty, elts.as_ptr(), len) }
diff --git a/compiler/rustc_codegen_llvm/src/consts.rs b/compiler/rustc_codegen_llvm/src/consts.rs
index dc9bb743560..a110ecbb75d 100644
--- a/compiler/rustc_codegen_llvm/src/consts.rs
+++ b/compiler/rustc_codegen_llvm/src/consts.rs
@@ -16,6 +16,7 @@ use rustc_middle::mir::mono::MonoItem;
 use rustc_middle::ty::layout::{HasTypingEnv, LayoutOf};
 use rustc_middle::ty::{self, Instance};
 use rustc_middle::{bug, span_bug};
+use rustc_span::Symbol;
 use tracing::{debug, instrument, trace};
 
 use crate::common::CodegenCx;
@@ -331,6 +332,10 @@ impl<'ll> CodegenCx<'ll, '_> {
             }
 
             g
+        } else if let Some(classname) = fn_attrs.objc_class {
+            self.get_objc_classref(classname)
+        } else if let Some(methname) = fn_attrs.objc_selector {
+            self.get_objc_selref(methname)
         } else {
             check_and_apply_linkage(self, fn_attrs, llty, sym, def_id)
         };
@@ -543,8 +548,225 @@ impl<'ll> CodegenCx<'ll, '_> {
 
     /// Add a global value to a list to be stored in the `llvm.compiler.used` variable,
     /// an array of ptr.
-    pub(crate) fn add_compiler_used_global(&mut self, global: &'ll Value) {
-        self.compiler_used_statics.push(global);
+    pub(crate) fn add_compiler_used_global(&self, global: &'ll Value) {
+        self.compiler_used_statics.borrow_mut().push(global);
+    }
+
+    // We do our best here to match what Clang does when compiling Objective-C natively.
+    // See Clang's `CGObjCCommonMac::CreateCStringLiteral`:
+    // https://github.com/llvm/llvm-project/blob/llvmorg-20.1.8/clang/lib/CodeGen/CGObjCMac.cpp#L4134
+    fn define_objc_classname(&self, classname: &str) -> &'ll Value {
+        assert_eq!(self.objc_abi_version(), 1);
+
+        let llval = self.null_terminate_const_bytes(classname.as_bytes());
+        let llty = self.val_ty(llval);
+        let sym = self.generate_local_symbol_name("OBJC_CLASS_NAME_");
+        let g = self.define_global(&sym, llty).unwrap_or_else(|| {
+            bug!("symbol `{}` is already defined", sym);
+        });
+        set_global_alignment(self, g, self.tcx.data_layout.i8_align.abi);
+        llvm::set_initializer(g, llval);
+        llvm::set_linkage(g, llvm::Linkage::PrivateLinkage);
+        llvm::set_section(g, c"__TEXT,__cstring,cstring_literals");
+        llvm::LLVMSetGlobalConstant(g, llvm::TRUE);
+        llvm::LLVMSetUnnamedAddress(g, llvm::UnnamedAddr::Global);
+        self.add_compiler_used_global(g);
+
+        g
+    }
+
+    // We do our best here to match what Clang does when compiling Objective-C natively.
+    // See Clang's `ObjCNonFragileABITypesHelper`:
+    // https://github.com/llvm/llvm-project/blob/llvmorg-20.1.8/clang/lib/CodeGen/CGObjCMac.cpp#L6052
+    fn get_objc_class_t(&self) -> &'ll Type {
+        if let Some(class_t) = self.objc_class_t.get() {
+            return class_t;
+        }
+
+        assert_eq!(self.objc_abi_version(), 2);
+
+        // struct _class_t {
+        //     struct _class_t* isa;
+        //     struct _class_t* const superclass;
+        //     void* cache;
+        //     IMP* vtable;
+        //     struct class_ro_t* ro;
+        // }
+
+        let class_t = self.type_named_struct("struct._class_t");
+        let els = [self.type_ptr(); 5];
+        let packed = false;
+        self.set_struct_body(class_t, &els, packed);
+
+        self.objc_class_t.set(Some(class_t));
+        class_t
+    }
+
+    // We do our best here to match what Clang does when compiling Objective-C natively. We
+    // deduplicate references within a CGU, but we need a reference definition in each referencing
+    // CGU. All attempts at using external references to a single reference definition result in
+    // linker errors.
+    fn get_objc_classref(&self, classname: Symbol) -> &'ll Value {
+        let mut classrefs = self.objc_classrefs.borrow_mut();
+        if let Some(classref) = classrefs.get(&classname).copied() {
+            return classref;
+        }
+
+        let g = match self.objc_abi_version() {
+            1 => {
+                // See Clang's `CGObjCMac::EmitClassRefFromId`:
+                // https://github.com/llvm/llvm-project/blob/llvmorg-20.1.8/clang/lib/CodeGen/CGObjCMac.cpp#L5205
+                let llval = self.define_objc_classname(classname.as_str());
+                let llty = self.type_ptr();
+                let sym = self.generate_local_symbol_name("OBJC_CLASS_REFERENCES_");
+                let g = self.define_global(&sym, llty).unwrap_or_else(|| {
+                    bug!("symbol `{}` is already defined", sym);
+                });
+                set_global_alignment(self, g, self.tcx.data_layout.pointer_align().abi);
+                llvm::set_initializer(g, llval);
+                llvm::set_linkage(g, llvm::Linkage::PrivateLinkage);
+                llvm::set_section(g, c"__OBJC,__cls_refs,literal_pointers,no_dead_strip");
+                self.add_compiler_used_global(g);
+                g
+            }
+            2 => {
+                // See Clang's `CGObjCNonFragileABIMac::EmitClassRefFromId`:
+                // https://github.com/llvm/llvm-project/blob/llvmorg-20.1.8/clang/lib/CodeGen/CGObjCMac.cpp#L7423
+                let llval = {
+                    let extern_sym = format!("OBJC_CLASS_$_{}", classname.as_str());
+                    let extern_llty = self.get_objc_class_t();
+                    self.declare_global(&extern_sym, extern_llty)
+                };
+                let llty = self.type_ptr();
+                let sym = self.generate_local_symbol_name("OBJC_CLASSLIST_REFERENCES_$_");
+                let g = self.define_global(&sym, llty).unwrap_or_else(|| {
+                    bug!("symbol `{}` is already defined", sym);
+                });
+                set_global_alignment(self, g, self.tcx.data_layout.pointer_align().abi);
+                llvm::set_initializer(g, llval);
+                llvm::set_linkage(g, llvm::Linkage::InternalLinkage);
+                llvm::set_section(g, c"__DATA,__objc_classrefs,regular,no_dead_strip");
+                self.add_compiler_used_global(g);
+                g
+            }
+            _ => unreachable!(),
+        };
+
+        classrefs.insert(classname, g);
+        g
+    }
+
+    // We do our best here to match what Clang does when compiling Objective-C natively. We
+    // deduplicate references within a CGU, but we need a reference definition in each referencing
+    // CGU. All attempts at using external references to a single reference definition result in
+    // linker errors.
+    //
+    // Newer versions of Apple Clang generate calls to `@"objc_msgSend$methname"` selector stub
+    // functions. We don't currently do that. The code we generate is closer to what Apple Clang
+    // generates with the `-fno-objc-msgsend-selector-stubs` option.
+    fn get_objc_selref(&self, methname: Symbol) -> &'ll Value {
+        let mut selrefs = self.objc_selrefs.borrow_mut();
+        if let Some(selref) = selrefs.get(&methname).copied() {
+            return selref;
+        }
+
+        let abi_version = self.objc_abi_version();
+
+        // See Clang's `CGObjCCommonMac::CreateCStringLiteral`:
+        // https://github.com/llvm/llvm-project/blob/llvmorg-20.1.8/clang/lib/CodeGen/CGObjCMac.cpp#L4134
+        let methname_llval = self.null_terminate_const_bytes(methname.as_str().as_bytes());
+        let methname_llty = self.val_ty(methname_llval);
+        let methname_sym = self.generate_local_symbol_name("OBJC_METH_VAR_NAME_");
+        let methname_g = self.define_global(&methname_sym, methname_llty).unwrap_or_else(|| {
+            bug!("symbol `{}` is already defined", methname_sym);
+        });
+        set_global_alignment(self, methname_g, self.tcx.data_layout.i8_align.abi);
+        llvm::set_initializer(methname_g, methname_llval);
+        llvm::set_linkage(methname_g, llvm::Linkage::PrivateLinkage);
+        llvm::set_section(
+            methname_g,
+            match abi_version {
+                1 => c"__TEXT,__cstring,cstring_literals",
+                2 => c"__TEXT,__objc_methname,cstring_literals",
+                _ => unreachable!(),
+            },
+        );
+        llvm::LLVMSetGlobalConstant(methname_g, llvm::TRUE);
+        llvm::LLVMSetUnnamedAddress(methname_g, llvm::UnnamedAddr::Global);
+        self.add_compiler_used_global(methname_g);
+
+        // See Clang's `CGObjCMac::EmitSelectorAddr`:
+        // https://github.com/llvm/llvm-project/blob/llvmorg-20.1.8/clang/lib/CodeGen/CGObjCMac.cpp#L5243
+        // And Clang's `CGObjCNonFragileABIMac::EmitSelectorAddr`:
+        // https://github.com/llvm/llvm-project/blob/llvmorg-20.1.8/clang/lib/CodeGen/CGObjCMac.cpp#L7586
+        let selref_llval = methname_g;
+        let selref_llty = self.type_ptr();
+        let selref_sym = self.generate_local_symbol_name("OBJC_SELECTOR_REFERENCES_");
+        let selref_g = self.define_global(&selref_sym, selref_llty).unwrap_or_else(|| {
+            bug!("symbol `{}` is already defined", selref_sym);
+        });
+        set_global_alignment(self, selref_g, self.tcx.data_layout.pointer_align().abi);
+        llvm::set_initializer(selref_g, selref_llval);
+        llvm::set_externally_initialized(selref_g, true);
+        llvm::set_linkage(
+            selref_g,
+            match abi_version {
+                1 => llvm::Linkage::PrivateLinkage,
+                2 => llvm::Linkage::InternalLinkage,
+                _ => unreachable!(),
+            },
+        );
+        llvm::set_section(
+            selref_g,
+            match abi_version {
+                1 => c"__OBJC,__message_refs,literal_pointers,no_dead_strip",
+                2 => c"__DATA,__objc_selrefs,literal_pointers,no_dead_strip",
+                _ => unreachable!(),
+            },
+        );
+        self.add_compiler_used_global(selref_g);
+
+        selrefs.insert(methname, selref_g);
+        selref_g
+    }
+
+    // We do our best here to match what Clang does when compiling Objective-C natively.
+    // See Clang's `ObjCTypesHelper`:
+    // https://github.com/llvm/llvm-project/blob/llvmorg-20.1.8/clang/lib/CodeGen/CGObjCMac.cpp#L5936
+    // And Clang's `CGObjCMac::EmitModuleInfo`:
+    // https://github.com/llvm/llvm-project/blob/llvmorg-20.1.8/clang/lib/CodeGen/CGObjCMac.cpp#L5151
+    pub(crate) fn define_objc_module_info(&mut self) {
+        assert_eq!(self.objc_abi_version(), 1);
+
+        // struct _objc_module {
+        //     long version;                // Hardcoded to 7 in Clang.
+        //     long size;                   // sizeof(struct _objc_module)
+        //     char* name;                  // Hardcoded to classname "" in Clang.
+        //     struct _objc_symtab* symtab; // Null without class or category definitions.
+        //  }
+
+        let llty = self.type_named_struct("struct._objc_module");
+        let i32_llty = self.type_i32();
+        let ptr_llty = self.type_ptr();
+        let packed = false;
+        self.set_struct_body(llty, &[i32_llty, i32_llty, ptr_llty, ptr_llty], packed);
+
+        let version = self.const_uint(i32_llty, 7);
+        let size = self.const_uint(i32_llty, 16);
+        let name = self.define_objc_classname("");
+        let symtab = self.const_null(ptr_llty);
+        let llval = crate::common::named_struct(llty, &[version, size, name, symtab]);
+
+        let sym = "OBJC_MODULES";
+        let g = self.define_global(&sym, llty).unwrap_or_else(|| {
+            bug!("symbol `{}` is already defined", sym);
+        });
+        set_global_alignment(self, g, self.tcx.data_layout.pointer_align().abi);
+        llvm::set_initializer(g, llval);
+        llvm::set_linkage(g, llvm::Linkage::PrivateLinkage);
+        llvm::set_section(g, c"__OBJC,__module_info,regular,no_dead_strip");
+
+        self.add_compiler_used_global(g);
     }
 }
 
diff --git a/compiler/rustc_codegen_llvm/src/context.rs b/compiler/rustc_codegen_llvm/src/context.rs
index a69fa54a54a..2f309d4694c 100644
--- a/compiler/rustc_codegen_llvm/src/context.rs
+++ b/compiler/rustc_codegen_llvm/src/context.rs
@@ -26,7 +26,7 @@ use rustc_session::config::{
     BranchProtection, CFGuard, CFProtection, CrateType, DebugInfo, FunctionReturn, PAuthKey, PacRet,
 };
 use rustc_span::source_map::Spanned;
-use rustc_span::{DUMMY_SP, Span};
+use rustc_span::{DUMMY_SP, Span, Symbol};
 use rustc_symbol_mangling::mangle_internal_symbol;
 use rustc_target::spec::{HasTargetSpec, RelocModel, SmallDataThresholdSupport, Target, TlsModel};
 use smallvec::SmallVec;
@@ -119,7 +119,7 @@ pub(crate) struct FullCx<'ll, 'tcx> {
 
     /// Statics that will be placed in the llvm.compiler.used variable
     /// See <https://llvm.org/docs/LangRef.html#the-llvm-compiler-used-global-variable> for details
-    pub compiler_used_statics: Vec<&'ll Value>,
+    pub compiler_used_statics: RefCell<Vec<&'ll Value>>,
 
     /// Mapping of non-scalar types to llvm types.
     pub type_lowering: RefCell<FxHashMap<(Ty<'tcx>, Option<VariantIdx>), &'ll Type>>,
@@ -146,6 +146,15 @@ pub(crate) struct FullCx<'ll, 'tcx> {
     /// `global_asm!` needs to be able to find this new global so that it can
     /// compute the correct mangled symbol name to insert into the asm.
     pub renamed_statics: RefCell<FxHashMap<DefId, &'ll Value>>,
+
+    /// Cached Objective-C class type
+    pub objc_class_t: Cell<Option<&'ll Type>>,
+
+    /// Cache of Objective-C class references
+    pub objc_classrefs: RefCell<FxHashMap<Symbol, &'ll Value>>,
+
+    /// Cache of Objective-C selector references
+    pub objc_selrefs: RefCell<FxHashMap<Symbol, &'ll Value>>,
 }
 
 fn to_llvm_tls_model(tls_model: TlsModel) -> llvm::ThreadLocalMode {
@@ -644,7 +653,7 @@ impl<'ll, 'tcx> CodegenCx<'ll, 'tcx> {
                 const_globals: Default::default(),
                 statics_to_rauw: RefCell::new(Vec::new()),
                 used_statics: Vec::new(),
-                compiler_used_statics: Vec::new(),
+                compiler_used_statics: Default::default(),
                 type_lowering: Default::default(),
                 scalar_lltypes: Default::default(),
                 coverage_cx,
@@ -655,6 +664,9 @@ impl<'ll, 'tcx> CodegenCx<'ll, 'tcx> {
                 intrinsics: Default::default(),
                 local_gen_sym_counter: Cell::new(0),
                 renamed_statics: Default::default(),
+                objc_class_t: Cell::new(None),
+                objc_classrefs: Default::default(),
+                objc_selrefs: Default::default(),
             },
             PhantomData,
         )
@@ -679,6 +691,69 @@ impl<'ll, 'tcx> CodegenCx<'ll, 'tcx> {
         llvm::set_linkage(g, llvm::Linkage::AppendingLinkage);
         llvm::set_section(g, c"llvm.metadata");
     }
+
+    /// The Objective-C ABI that is used.
+    ///
+    /// This corresponds to the `-fobjc-abi-version=` flag in Clang / GCC.
+    pub(crate) fn objc_abi_version(&self) -> u32 {
+        assert!(self.tcx.sess.target.is_like_darwin);
+        if self.tcx.sess.target.arch == "x86" && self.tcx.sess.target.os == "macos" {
+            // 32-bit x86 macOS uses ABI version 1 (a.k.a. the "fragile ABI").
+            1
+        } else {
+            // All other Darwin-like targets we support use ABI version 2
+            // (a.k.a the "non-fragile ABI").
+            2
+        }
+    }
+
+    // We do our best here to match what Clang does when compiling Objective-C natively.
+    // See Clang's `CGObjCCommonMac::EmitImageInfo`:
+    // https://github.com/llvm/llvm-project/blob/llvmorg-20.1.8/clang/lib/CodeGen/CGObjCMac.cpp#L5085
+    pub(crate) fn add_objc_module_flags(&self) {
+        let abi_version = self.objc_abi_version();
+
+        llvm::add_module_flag_u32(
+            self.llmod,
+            llvm::ModuleFlagMergeBehavior::Error,
+            "Objective-C Version",
+            abi_version,
+        );
+
+        llvm::add_module_flag_u32(
+            self.llmod,
+            llvm::ModuleFlagMergeBehavior::Error,
+            "Objective-C Image Info Version",
+            0,
+        );
+
+        llvm::add_module_flag_str(
+            self.llmod,
+            llvm::ModuleFlagMergeBehavior::Error,
+            "Objective-C Image Info Section",
+            match abi_version {
+                1 => "__OBJC,__image_info,regular",
+                2 => "__DATA,__objc_imageinfo,regular,no_dead_strip",
+                _ => unreachable!(),
+            },
+        );
+
+        if self.tcx.sess.target.env == "sim" {
+            llvm::add_module_flag_u32(
+                self.llmod,
+                llvm::ModuleFlagMergeBehavior::Error,
+                "Objective-C Is Simulated",
+                1 << 5,
+            );
+        }
+
+        llvm::add_module_flag_u32(
+            self.llmod,
+            llvm::ModuleFlagMergeBehavior::Error,
+            "Objective-C Class Properties",
+            1 << 6,
+        );
+    }
 }
 impl<'ll> SimpleCx<'ll> {
     pub(crate) fn get_type_of_global(&self, val: &'ll Value) -> &'ll Type {
diff --git a/compiler/rustc_codegen_llvm/src/llvm/ffi.rs b/compiler/rustc_codegen_llvm/src/llvm/ffi.rs
index 0679f55ab7f..e125aba6afc 100644
--- a/compiler/rustc_codegen_llvm/src/llvm/ffi.rs
+++ b/compiler/rustc_codegen_llvm/src/llvm/ffi.rs
@@ -1237,6 +1237,7 @@ unsafe extern "C" {
     pub(crate) safe fn LLVMSetGlobalConstant(GlobalVar: &Value, IsConstant: Bool);
     pub(crate) safe fn LLVMSetTailCall(CallInst: &Value, IsTailCall: Bool);
     pub(crate) safe fn LLVMSetTailCallKind(CallInst: &Value, kind: TailCallKind);
+    pub(crate) safe fn LLVMSetExternallyInitialized(GlobalVar: &Value, IsExtInit: Bool);
 
     // Operations on attributes
     pub(crate) fn LLVMCreateStringAttribute(
diff --git a/compiler/rustc_codegen_llvm/src/llvm/mod.rs b/compiler/rustc_codegen_llvm/src/llvm/mod.rs
index d6974e22c85..1115d82fa85 100644
--- a/compiler/rustc_codegen_llvm/src/llvm/mod.rs
+++ b/compiler/rustc_codegen_llvm/src/llvm/mod.rs
@@ -258,6 +258,10 @@ pub(crate) fn set_alignment(llglobal: &Value, align: Align) {
     }
 }
 
+pub(crate) fn set_externally_initialized(llglobal: &Value, is_ext_init: bool) {
+    LLVMSetExternallyInitialized(llglobal, is_ext_init.to_llvm_bool());
+}
+
 /// Get the `name`d comdat from `llmod` and assign it to `llglobal`.
 ///
 /// Inserts the comdat into `llmod` if it does not exist.