Rollup merge of #77795 - bjorn3:codegen_backend_interface_refactor, r=oli-obk

Codegen backend interface refactor

This moves several things away from the codegen backend to rustc_interface. There are a few behavioral changes where previously the incremental cache (incorrectly) wouldn't get finalized, but now it does. See the individual commit messages.

compiler/rustc_codegen_llvm/src/attributes.rs

@@ -6,7 +6,7 @@ use rustc_codegen_ssa::traits::*;
 use rustc_data_structures::const_cstr;
 use rustc_data_structures::fx::FxHashMap;
 use rustc_data_structures::small_c_str::SmallCStr;
-use rustc_hir::def_id::{DefId, LOCAL_CRATE};
+use rustc_hir::def_id::DefId;
 use rustc_middle::middle::codegen_fn_attrs::CodegenFnAttrFlags;
 use rustc_middle::ty::layout::HasTyCtxt;
 use rustc_middle::ty::query::Providers;
@@ -367,23 +367,7 @@ pub fn from_fn_attrs(cx: &CodegenCx<'ll, 'tcx>, llfn: &'ll Value, instance: ty::
     }
 }
 
-pub fn provide(providers: &mut Providers) {
+pub fn provide_both(providers: &mut Providers) {
-    use rustc_codegen_ssa::target_features::{all_known_features, supported_target_features};
-    providers.supported_target_features = |tcx, cnum| {
-        assert_eq!(cnum, LOCAL_CRATE);
-        if tcx.sess.opts.actually_rustdoc {
-            // rustdoc needs to be able to document functions that use all the features, so
-            // provide them all.
-            all_known_features().map(|(a, b)| (a.to_string(), b)).collect()
-        } else {
-            supported_target_features(tcx.sess).iter().map(|&(a, b)| (a.to_string(), b)).collect()
-        }
-    };
-
-    provide_extern(providers);
-}
-
-pub fn provide_extern(providers: &mut Providers) {
     providers.wasm_import_module_map = |tcx, cnum| {
         // Build up a map from DefId to a `NativeLib` structure, where
         // `NativeLib` internally contains information about

compiler/rustc_codegen_llvm/src/lib.rs

@@ -23,18 +23,17 @@ use rustc_codegen_ssa::back::write::{CodegenContext, FatLTOInput, ModuleConfig};
 use rustc_codegen_ssa::traits::*;
 use rustc_codegen_ssa::ModuleCodegen;
 use rustc_codegen_ssa::{CodegenResults, CompiledModule};
+use rustc_data_structures::fx::FxHashMap;
 use rustc_errors::{ErrorReported, FatalError, Handler};
-use rustc_middle::dep_graph::{DepGraph, WorkProduct};
+use rustc_middle::dep_graph::{WorkProduct, WorkProductId};
 use rustc_middle::middle::cstore::{EncodedMetadata, MetadataLoaderDyn};
 use rustc_middle::ty::{self, TyCtxt};
-use rustc_serialize::json;
+use rustc_session::config::{OptLevel, OutputFilenames, PrintRequest};
-use rustc_session::config::{self, OptLevel, OutputFilenames, PrintRequest};
 use rustc_session::Session;
 use rustc_span::symbol::Symbol;
 
 use std::any::Any;
 use std::ffi::CStr;
-use std::fs;
 use std::sync::Arc;
 
 mod back {
@@ -252,11 +251,11 @@ impl CodegenBackend for LlvmCodegenBackend {
     }
 
     fn provide(&self, providers: &mut ty::query::Providers) {
-        attributes::provide(providers);
+        attributes::provide_both(providers);
     }
 
     fn provide_extern(&self, providers: &mut ty::query::Providers) {
-        attributes::provide_extern(providers);
+        attributes::provide_both(providers);
     }
 
     fn codegen_crate<'tcx>(
@@ -277,47 +276,27 @@ impl CodegenBackend for LlvmCodegenBackend {
         &self,
         ongoing_codegen: Box<dyn Any>,
         sess: &Session,
-        dep_graph: &DepGraph,
+    ) -> Result<(CodegenResults, FxHashMap<WorkProductId, WorkProduct>), ErrorReported> {
-    ) -> Result<Box<dyn Any>, ErrorReported> {
         let (codegen_results, work_products) = ongoing_codegen
             .downcast::<rustc_codegen_ssa::back::write::OngoingCodegen<LlvmCodegenBackend>>()
             .expect("Expected LlvmCodegenBackend's OngoingCodegen, found Box<Any>")
             .join(sess);
-        if sess.opts.debugging_opts.incremental_info {
-            rustc_codegen_ssa::back::write::dump_incremental_data(&codegen_results);
-        }
 
-        sess.time("serialize_work_products", move || {
+        sess.time("llvm_dump_timing_file", || {
-            rustc_incremental::save_work_product_index(sess, &dep_graph, work_products)
+            if sess.opts.debugging_opts.llvm_time_trace {
+                llvm_util::time_trace_profiler_finish("llvm_timings.json");
+            }
         });
 
-        sess.compile_status()?;
+        Ok((codegen_results, work_products))
-
-        Ok(Box::new(codegen_results))
     }
 
     fn link(
         &self,
         sess: &Session,
-        codegen_results: Box<dyn Any>,
+        codegen_results: CodegenResults,
         outputs: &OutputFilenames,
     ) -> Result<(), ErrorReported> {
-        let codegen_results = codegen_results
-            .downcast::<CodegenResults>()
-            .expect("Expected CodegenResults, found Box<Any>");
-
-        if sess.opts.debugging_opts.no_link {
-            // FIXME: use a binary format to encode the `.rlink` file
-            let rlink_data = json::encode(&codegen_results).map_err(|err| {
-                sess.fatal(&format!("failed to encode rlink: {}", err));
-            })?;
-            let rlink_file = outputs.with_extension(config::RLINK_EXT);
-            fs::write(&rlink_file, rlink_data).map_err(|err| {
-                sess.fatal(&format!("failed to write file {}: {}", rlink_file.display(), err));
-            })?;
-            return Ok(());
-        }
-
         // Run the linker on any artifacts that resulted from the LLVM run.
         // This should produce either a finished executable or library.
         sess.time("link_crate", || {
@@ -334,16 +313,6 @@ impl CodegenBackend for LlvmCodegenBackend {
             );
         });
 
-        // Now that we won't touch anything in the incremental compilation directory
-        // any more, we can finalize it (which involves renaming it)
-        rustc_incremental::finalize_session_directory(sess, codegen_results.crate_hash);
-
-        sess.time("llvm_dump_timing_file", || {
-            if sess.opts.debugging_opts.llvm_time_trace {
-                llvm_util::time_trace_profiler_finish("llvm_timings.json");
-            }
-        });
-
         Ok(())
     }
 }

compiler/rustc_codegen_ssa/src/back/write.rs

@@ -13,7 +13,6 @@ use rustc_data_structures::fx::FxHashMap;
 use rustc_data_structures::profiling::SelfProfilerRef;
 use rustc_data_structures::profiling::TimingGuard;
 use rustc_data_structures::profiling::VerboseTimingGuard;
-use rustc_data_structures::svh::Svh;
 use rustc_data_structures::sync::Lrc;
 use rustc_errors::emitter::Emitter;
 use rustc_errors::{DiagnosticId, FatalError, Handler, Level};
@@ -414,7 +413,6 @@ pub fn start_async_codegen<B: ExtraBackendMethods>(
     let sess = tcx.sess;
 
     let crate_name = tcx.crate_name(LOCAL_CRATE);
-    let crate_hash = tcx.crate_hash(LOCAL_CRATE);
     let no_builtins = tcx.sess.contains_name(&tcx.hir().krate().item.attrs, sym::no_builtins);
     let is_compiler_builtins =
         tcx.sess.contains_name(&tcx.hir().krate().item.attrs, sym::compiler_builtins);
@@ -463,7 +461,6 @@ pub fn start_async_codegen<B: ExtraBackendMethods>(
     OngoingCodegen {
         backend,
         crate_name,
-        crate_hash,
         metadata,
         windows_subsystem,
         linker_info,
@@ -658,15 +655,6 @@ fn produce_final_output_artifacts(
     // These are used in linking steps and will be cleaned up afterward.
 }
 
-pub fn dump_incremental_data(_codegen_results: &CodegenResults) {
-    // FIXME(mw): This does not work at the moment because the situation has
-    //            become more complicated due to incremental LTO. Now a CGU
-    //            can have more than two caching states.
-    // println!("[incremental] Re-using {} out of {} modules",
-    //           codegen_results.modules.iter().filter(|m| m.pre_existing).count(),
-    //           codegen_results.modules.len());
-}
-
 pub enum WorkItem<B: WriteBackendMethods> {
     /// Optimize a newly codegened, totally unoptimized module.
     Optimize(ModuleCodegen<B::Module>),
@@ -1720,7 +1708,6 @@ impl SharedEmitterMain {
 pub struct OngoingCodegen<B: ExtraBackendMethods> {
     pub backend: B,
     pub crate_name: Symbol,
-    pub crate_hash: Svh,
     pub metadata: EncodedMetadata,
     pub windows_subsystem: Option<String>,
     pub linker_info: LinkerInfo,
@@ -1766,7 +1753,6 @@ impl<B: ExtraBackendMethods> OngoingCodegen<B> {
         (
             CodegenResults {
                 crate_name: self.crate_name,
-                crate_hash: self.crate_hash,
                 metadata: self.metadata,
                 windows_subsystem: self.windows_subsystem,
                 linker_info: self.linker_info,

compiler/rustc_codegen_ssa/src/lib.rs

@@ -21,7 +21,6 @@ extern crate tracing;
 extern crate rustc_middle;
 
 use rustc_data_structures::fx::{FxHashMap, FxHashSet};
-use rustc_data_structures::svh::Svh;
 use rustc_data_structures::sync::Lrc;
 use rustc_hir::def_id::CrateNum;
 use rustc_hir::LangItem;
@@ -134,7 +133,6 @@ pub struct CodegenResults {
     pub modules: Vec<CompiledModule>,
     pub allocator_module: Option<CompiledModule>,
     pub metadata_module: Option<CompiledModule>,
-    pub crate_hash: Svh,
     pub metadata: rustc_middle::middle::cstore::EncodedMetadata,
     pub windows_subsystem: Option<String>,
     pub linker_info: back::linker::LinkerInfo,
@@ -144,6 +142,7 @@ pub struct CodegenResults {
 pub fn provide(providers: &mut Providers) {
     crate::back::symbol_export::provide(providers);
     crate::base::provide_both(providers);
+    crate::target_features::provide(providers);
 }
 
 pub fn provide_extern(providers: &mut Providers) {

compiler/rustc_codegen_ssa/src/target_features.rs

@@ -1,3 +1,5 @@
+use rustc_hir::def_id::LOCAL_CRATE;
+use rustc_middle::ty::query::Providers;
 use rustc_session::Session;
 use rustc_span::symbol::sym;
 use rustc_span::symbol::Symbol;
@@ -148,3 +150,16 @@ pub fn supported_target_features(sess: &Session) -> &'static [(&'static str, Opt
         _ => &[],
     }
 }
+
+pub(crate) fn provide(providers: &mut Providers) {
+    providers.supported_target_features = |tcx, cnum| {
+        assert_eq!(cnum, LOCAL_CRATE);
+        if tcx.sess.opts.actually_rustdoc {
+            // rustdoc needs to be able to document functions that use all the features, so
+            // whitelist them all
+            all_known_features().map(|(a, b)| (a.to_string(), b)).collect()
+        } else {
+            supported_target_features(tcx.sess).iter().map(|&(a, b)| (a.to_string(), b)).collect()
+        }
+    };
+}

compiler/rustc_codegen_ssa/src/traits/backend.rs

@@ -1,10 +1,11 @@
 use super::write::WriteBackendMethods;
 use super::CodegenObject;
-use crate::ModuleCodegen;
+use crate::{CodegenResults, ModuleCodegen};
 
 use rustc_ast::expand::allocator::AllocatorKind;
+use rustc_data_structures::fx::FxHashMap;
 use rustc_errors::ErrorReported;
-use rustc_middle::dep_graph::DepGraph;
+use rustc_middle::dep_graph::{WorkProduct, WorkProductId};
 use rustc_middle::middle::cstore::{EncodedMetadata, MetadataLoaderDyn};
 use rustc_middle::ty::layout::{HasTyCtxt, TyAndLayout};
 use rustc_middle::ty::query::Providers;
@@ -80,8 +81,7 @@ pub trait CodegenBackend {
         &self,
         ongoing_codegen: Box<dyn Any>,
         sess: &Session,
-        dep_graph: &DepGraph,
+    ) -> Result<(CodegenResults, FxHashMap<WorkProductId, WorkProduct>), ErrorReported>;
-    ) -> Result<Box<dyn Any>, ErrorReported>;
 
     /// This is called on the returned `Box<dyn Any>` from `join_codegen`
     ///
@@ -91,7 +91,7 @@ pub trait CodegenBackend {
     fn link(
         &self,
         sess: &Session,
-        codegen_results: Box<dyn Any>,
+        codegen_results: CodegenResults,
         outputs: &OutputFilenames,
     ) -> Result<(), ErrorReported>;
 }

compiler/rustc_driver/src/lib.rs

@@ -642,7 +642,7 @@ impl RustcDefaultCalls {
             let codegen_results: CodegenResults = json::decode(&rlink_data).unwrap_or_else(|err| {
                 sess.fatal(&format!("failed to decode rlink: {}", err));
             });
-            compiler.codegen_backend().link(&sess, Box::new(codegen_results), &outputs)
+            compiler.codegen_backend().link(&sess, codegen_results, &outputs)
         } else {
             sess.fatal("rlink must be a file")
         }

compiler/rustc_interface/src/queries.rs

@@ -3,6 +3,7 @@ use crate::passes::{self, BoxedResolver, QueryContext};
 
 use rustc_ast as ast;
 use rustc_codegen_ssa::traits::CodegenBackend;
+use rustc_data_structures::svh::Svh;
 use rustc_data_structures::sync::{Lrc, OnceCell, WorkerLocal};
 use rustc_errors::ErrorReported;
 use rustc_hir::def_id::LOCAL_CRATE;
@@ -13,7 +14,8 @@ use rustc_middle::arena::Arena;
 use rustc_middle::dep_graph::DepGraph;
 use rustc_middle::ty::steal::Steal;
 use rustc_middle::ty::{GlobalCtxt, ResolverOutputs, TyCtxt};
-use rustc_session::config::{OutputFilenames, OutputType};
+use rustc_serialize::json;
+use rustc_session::config::{self, OutputFilenames, OutputType};
 use rustc_session::{output::find_crate_name, Session};
 use rustc_span::symbol::sym;
 use std::any::Any;
@@ -331,6 +333,7 @@ impl<'tcx> Queries<'tcx> {
     pub fn linker(&'tcx self) -> Result<Linker> {
         let dep_graph = self.dep_graph()?;
         let prepare_outputs = self.prepare_outputs()?;
+        let crate_hash = self.global_ctxt()?.peek_mut().enter(|tcx| tcx.crate_hash(LOCAL_CRATE));
         let ongoing_codegen = self.ongoing_codegen()?;
 
         let sess = self.session().clone();
@@ -340,6 +343,7 @@ impl<'tcx> Queries<'tcx> {
             sess,
             dep_graph: dep_graph.peek().clone(),
             prepare_outputs: prepare_outputs.take(),
+            crate_hash,
             ongoing_codegen: ongoing_codegen.take(),
             codegen_backend,
         })
@@ -350,18 +354,31 @@ pub struct Linker {
     sess: Lrc<Session>,
     dep_graph: DepGraph,
     prepare_outputs: OutputFilenames,
+    crate_hash: Svh,
     ongoing_codegen: Box<dyn Any>,
     codegen_backend: Lrc<Box<dyn CodegenBackend>>,
 }
 
 impl Linker {
     pub fn link(self) -> Result<()> {
-        let codegen_results =
+        let (codegen_results, work_products) =
-            self.codegen_backend.join_codegen(self.ongoing_codegen, &self.sess, &self.dep_graph)?;
+            self.codegen_backend.join_codegen(self.ongoing_codegen, &self.sess)?;
-        let prof = self.sess.prof.clone();
+
+        self.sess.compile_status()?;
+
+        let sess = &self.sess;
         let dep_graph = self.dep_graph;
+        sess.time("serialize_work_products", || {
+            rustc_incremental::save_work_product_index(&sess, &dep_graph, work_products)
+        });
+
+        let prof = self.sess.prof.clone();
         prof.generic_activity("drop_dep_graph").run(move || drop(dep_graph));
 
+        // Now that we won't touch anything in the incremental compilation directory
+        // any more, we can finalize it (which involves renaming it)
+        rustc_incremental::finalize_session_directory(&self.sess, self.crate_hash);
+
         if !self
             .sess
             .opts
@@ -371,6 +388,19 @@ impl Linker {
         {
             return Ok(());
         }
+
+        if sess.opts.debugging_opts.no_link {
+            // FIXME: use a binary format to encode the `.rlink` file
+            let rlink_data = json::encode(&codegen_results).map_err(|err| {
+                sess.fatal(&format!("failed to encode rlink: {}", err));
+            })?;
+            let rlink_file = self.prepare_outputs.with_extension(config::RLINK_EXT);
+            std::fs::write(&rlink_file, rlink_data).map_err(|err| {
+                sess.fatal(&format!("failed to write file {}: {}", rlink_file.display(), err));
+            })?;
+            return Ok(());
+        }
+
         self.codegen_backend.link(&self.sess, codegen_results, &self.prepare_outputs)
     }
 }

src/test/run-make-fulldeps/hotplug_codegen_backend/the_backend.rs

@@ -3,7 +3,6 @@
 extern crate rustc_codegen_ssa;
 extern crate rustc_errors;
 extern crate rustc_middle;
-#[macro_use]
 extern crate rustc_data_structures;
 extern crate rustc_driver;
 extern crate rustc_hir;
@@ -12,17 +11,19 @@ extern crate rustc_span;
 extern crate rustc_symbol_mangling;
 extern crate rustc_target;
 
+use rustc_codegen_ssa::back::linker::LinkerInfo;
 use rustc_codegen_ssa::traits::CodegenBackend;
-use rustc_data_structures::owning_ref::OwningRef;
+use rustc_codegen_ssa::{CodegenResults, CrateInfo};
+use rustc_data_structures::fx::FxHashMap;
 use rustc_data_structures::sync::MetadataRef;
 use rustc_errors::ErrorReported;
 use rustc_middle::dep_graph::DepGraph;
+use rustc_middle::dep_graph::{WorkProduct, WorkProductId};
 use rustc_middle::middle::cstore::{EncodedMetadata, MetadataLoader, MetadataLoaderDyn};
 use rustc_middle::ty::query::Providers;
 use rustc_middle::ty::TyCtxt;
 use rustc_session::config::OutputFilenames;
 use rustc_session::Session;
-use rustc_span::symbol::Symbol;
 use rustc_target::spec::Target;
 use std::any::Any;
 use std::path::Path;
@@ -31,14 +32,11 @@ pub struct NoLlvmMetadataLoader;
 
 impl MetadataLoader for NoLlvmMetadataLoader {
     fn get_rlib_metadata(&self, _: &Target, filename: &Path) -> Result<MetadataRef, String> {
-        let buf =
+        unreachable!("some_crate.rs shouldn't depend on any external crates");
-            std::fs::read(filename).map_err(|e| format!("metadata file open err: {:?}", e))?;
-        let buf: OwningRef<Vec<u8>, [u8]> = OwningRef::new(buf);
-        Ok(rustc_erase_owner!(buf.map_owner_box()))
     }
 
     fn get_dylib_metadata(&self, target: &Target, filename: &Path) -> Result<MetadataRef, String> {
-        self.get_rlib_metadata(target, filename)
+        unreachable!("some_crate.rs shouldn't depend on any external crates");
     }
 }
 
@@ -49,53 +47,49 @@ impl CodegenBackend for TheBackend {
         Box::new(NoLlvmMetadataLoader)
     }
 
-    fn provide(&self, providers: &mut Providers) {
+    fn provide(&self, providers: &mut Providers) {}
-        rustc_symbol_mangling::provide(providers);
+    fn provide_extern(&self, providers: &mut Providers) {}
-
-        providers.supported_target_features = |tcx, _cnum| {
-            Default::default() // Just a dummy
-        };
-        providers.is_reachable_non_generic = |_tcx, _defid| true;
-        providers.exported_symbols = |_tcx, _crate| &[];
-    }
-
-    fn provide_extern(&self, providers: &mut Providers) {
-        providers.is_reachable_non_generic = |_tcx, _defid| true;
-    }
 
     fn codegen_crate<'a, 'tcx>(
         &self,
         tcx: TyCtxt<'tcx>,
-        _metadata: EncodedMetadata,
+        metadata: EncodedMetadata,
         _need_metadata_module: bool,
     ) -> Box<dyn Any> {
         use rustc_hir::def_id::LOCAL_CRATE;
 
-        Box::new(tcx.crate_name(LOCAL_CRATE) as Symbol)
+        Box::new(CodegenResults {
+            crate_name: tcx.crate_name(LOCAL_CRATE),
+            modules: vec![],
+            allocator_module: None,
+            metadata_module: None,
+            metadata,
+            windows_subsystem: None,
+            linker_info: LinkerInfo::new(tcx),
+            crate_info: CrateInfo::new(tcx),
+        })
     }
 
     fn join_codegen(
         &self,
         ongoing_codegen: Box<dyn Any>,
         _sess: &Session,
-        _dep_graph: &DepGraph,
+    ) -> Result<(CodegenResults, FxHashMap<WorkProductId, WorkProduct>), ErrorReported> {
-    ) -> Result<Box<dyn Any>, ErrorReported> {
+        let codegen_results = ongoing_codegen
-        let crate_name = ongoing_codegen
+            .downcast::<CodegenResults>()
-            .downcast::<Symbol>()
+            .expect("in join_codegen: ongoing_codegen is not a CodegenResults");
-            .expect("in join_codegen: ongoing_codegen is not a Symbol");
+        Ok((*codegen_results, FxHashMap::default()))
-        Ok(crate_name)
     }
 
     fn link(
         &self,
         sess: &Session,
-        codegen_results: Box<dyn Any>,
+        codegen_results: CodegenResults,
         outputs: &OutputFilenames,
     ) -> Result<(), ErrorReported> {
         use rustc_session::{config::CrateType, output::out_filename};
         use std::io::Write;
-        let crate_name =
+        let crate_name = codegen_results.crate_name;
-            codegen_results.downcast::<Symbol>().expect("in link: codegen_results is not a Symbol");
         for &crate_type in sess.opts.crate_types.iter() {
             if crate_type != CrateType::Rlib {
                 sess.fatal(&format!("Crate type is {:?}", crate_type));