Rollup merge of #155036 - bjorn3:lto_refactors16, r=TaKO8Ki

Store a PathBuf rather than SerializedModule for cached modules

In cg_gcc `ModuleBuffer` already only contains a path anyway. And for moving LTO into `-Zlink-only` we will need to serialize `MaybeLtoModules`. By storing a path cached modules we avoid writing them to the disk a second time during serialization of `MaybeLtoModules`.

Some further improvements will require changes to cg_gcc that I would prefer landing in the cg_gcc repo to actually test the LTO changes in CI.

Part of https://github.com/rust-lang/compiler-team/issues/908

compiler/rustc_codegen_gcc/src/back/lto.rs

@@ -144,9 +144,12 @@ fn fat_lto(
     for module in modules {
         match module {
             FatLtoInput::InMemory(m) => in_memory.push(m),
-            FatLtoInput::Serialized { name, buffer } => {
+            FatLtoInput::Serialized { name, bitcode_path } => {
                 info!("pushing serialized module {:?}", name);
-                serialized_modules.push((buffer, CString::new(name).unwrap()));
+                serialized_modules.push((
+                    SerializedModule::from_file(&bitcode_path),
+                    CString::new(name).unwrap(),
+                ));
             }
         }
     }

compiler/rustc_codegen_llvm/src/back/lto.rs

@@ -223,9 +223,12 @@ fn fat_lto(
     for module in modules {
         match module {
             FatLtoInput::InMemory(m) => in_memory.push(m),
-            FatLtoInput::Serialized { name, buffer } => {
+            FatLtoInput::Serialized { name, bitcode_path } => {
                 info!("pushing serialized module {:?}", name);
-                serialized_modules.push((buffer, CString::new(name).unwrap()));
+                serialized_modules.push((
+                    SerializedModule::from_file(&bitcode_path),
+                    CString::new(name).unwrap(),
+                ));
             }
         }
     }
@@ -396,7 +399,9 @@ fn thin_lto(
         for (i, module) in modules.into_iter().enumerate() {
             let (name, buffer) = match module {
                 ThinLtoInput::Red { name, buffer } => (name, buffer),
-                ThinLtoInput::Green { wp, buffer } => (wp.cgu_name, buffer),
+                ThinLtoInput::Green { wp, bitcode_path } => {
+                    (wp.cgu_name, SerializedModule::from_file(&bitcode_path))
+                }
             };
             info!("local module: {} - {}", i, name);
             let cname = CString::new(name.as_bytes()).unwrap();

compiler/rustc_codegen_ssa/src/back/lto.rs

@@ -1,4 +1,6 @@
 use std::ffi::CString;
+use std::fs;
+use std::path::Path;
 use std::sync::Arc;
 
 use rustc_data_structures::memmap::Mmap;
@@ -49,6 +51,19 @@ pub enum SerializedModule<M: ModuleBufferMethods> {
 }
 
 impl<M: ModuleBufferMethods> SerializedModule<M> {
+    pub fn from_file(bc_path: &Path) -> Self {
+        let file = fs::File::open(&bc_path).unwrap_or_else(|e| {
+            panic!("failed to open LTO bitcode file `{}`: {}", bc_path.display(), e)
+        });
+
+        let mmap = unsafe {
+            Mmap::map(file).unwrap_or_else(|e| {
+                panic!("failed to mmap LTO bitcode file `{}`: {}", bc_path.display(), e)
+            })
+        };
+        SerializedModule::FromUncompressedFile(mmap)
+    }
+
     pub fn data(&self) -> &[u8] {
         match *self {
             SerializedModule::Local(ref m) => m.data(),

compiler/rustc_codegen_ssa/src/back/write.rs

@@ -8,7 +8,6 @@
 use rustc_abi::Size;
 use rustc_data_structures::fx::FxIndexMap;
 use rustc_data_structures::jobserver::{self, Acquired};
-use rustc_data_structures::memmap::Mmap;
 use rustc_data_structures::profiling::{SelfProfilerRef, VerboseTimingGuard};
 use rustc_errors::emitter::Emitter;
 use rustc_errors::{
@@ -35,9 +34,8 @@
 use rustc_target::spec::{MergeFunctions, SanitizerSet};
 use tracing::debug;
 
-use super::link::{self, ensure_removed};
+use crate::back::link::{self, ensure_removed};
-use super::lto::{self, SerializedModule};
+use crate::back::lto::{self, SerializedModule, check_lto_allowed};
-use crate::back::lto::check_lto_allowed;
 use crate::errors::ErrorCreatingRemarkDir;
 use crate::traits::*;
 use crate::{
@@ -774,13 +772,13 @@ pub(crate) enum WorkItemResult<B: WriteBackendMethods> {
 }
 
 pub enum FatLtoInput<B: WriteBackendMethods> {
-    Serialized { name: String, buffer: SerializedModule<B::ModuleBuffer> },
+    Serialized { name: String, bitcode_path: PathBuf },
     InMemory(ModuleCodegen<B::Module>),
 }
 
 pub enum ThinLtoInput<B: WriteBackendMethods> {
     Red { name: String, buffer: SerializedModule<B::ModuleBuffer> },
-    Green { wp: WorkProduct, buffer: SerializedModule<B::ModuleBuffer> },
+    Green { wp: WorkProduct, bitcode_path: PathBuf },
 }
 
 /// Actual LTO type we end up choosing based on multiple factors.
@@ -866,7 +864,7 @@ fn execute_optimize_work_item<B: WriteBackendMethods>(
                 });
                 WorkItemResult::NeedsFatLto(FatLtoInput::Serialized {
                     name: module.name,
-                    buffer: SerializedModule::Local(buffer),
+                    bitcode_path: path,
                 })
             }
             None => WorkItemResult::NeedsFatLto(FatLtoInput::InMemory(module)),
@@ -1166,10 +1164,7 @@ pub(crate) enum Message<B: WriteBackendMethods> {
 
     /// Similar to `CodegenDone`, but for reusing a pre-LTO artifact
     /// Sent from the main thread.
-    AddImportOnlyModule {
+    AddImportOnlyModule { bitcode_path: PathBuf, work_product: WorkProduct },
-        module_data: SerializedModule<B::ModuleBuffer>,
-        work_product: WorkProduct,
-    },
 
     /// The frontend has finished generating everything for all codegen units.
     /// Sent from the main thread.
@@ -1729,10 +1724,10 @@ enum CodegenState {
                     }
                 }
 
-                Message::AddImportOnlyModule { module_data, work_product } => {
+                Message::AddImportOnlyModule { bitcode_path, work_product } => {
                     assert_eq!(codegen_state, Ongoing);
                     assert_eq!(main_thread_state, MainThreadState::Codegenning);
-                    lto_import_only_modules.push((module_data, work_product));
+                    lto_import_only_modules.push((bitcode_path, work_product));
                     main_thread_state = MainThreadState::Idle;
                 }
             }
@@ -1758,8 +1753,8 @@ enum CodegenState {
                 needs_fat_lto.push(FatLtoInput::InMemory(allocator_module));
             }
 
-            for (module, wp) in lto_import_only_modules {
+            for (bitcode_path, wp) in lto_import_only_modules {
-                needs_fat_lto.push(FatLtoInput::Serialized { name: wp.cgu_name, buffer: module })
+                needs_fat_lto.push(FatLtoInput::Serialized { name: wp.cgu_name, bitcode_path })
             }
 
             return Ok(MaybeLtoModules::FatLto {
@@ -1772,8 +1767,8 @@ enum CodegenState {
             assert!(compiled_modules.is_empty());
             assert!(needs_fat_lto.is_empty());
 
-            for (buffer, wp) in lto_import_only_modules {
+            for (bitcode_path, wp) in lto_import_only_modules {
-                needs_thin_lto.push(ThinLtoInput::Green { wp, buffer })
+                needs_thin_lto.push(ThinLtoInput::Green { wp, bitcode_path })
             }
 
             if cgcx.lto == Lto::ThinLocal {
@@ -2133,14 +2128,14 @@ fn drop(&mut self) {
 }
 
 pub struct OngoingCodegen<B: WriteBackendMethods> {
-    pub backend: B,
+    backend: B,
-    pub output_filenames: Arc<OutputFilenames>,
+    output_filenames: Arc<OutputFilenames>,
     // Field order below is intended to terminate the coordinator thread before two fields below
     // drop and prematurely close channels used by coordinator thread. See `Coordinator`'s
     // `Drop` implementation for more info.
-    pub coordinator: Coordinator<B>,
+    pub(crate) coordinator: Coordinator<B>,
-    pub codegen_worker_receive: Receiver<CguMessage>,
+    codegen_worker_receive: Receiver<CguMessage>,
-    pub shared_emitter_main: SharedEmitterMain,
+    shared_emitter_main: SharedEmitterMain,
 }
 
 impl<B: WriteBackendMethods> OngoingCodegen<B> {
@@ -2285,20 +2280,13 @@ pub(crate) fn submit_pre_lto_module_to_llvm<B: WriteBackendMethods>(
     module: CachedModuleCodegen,
 ) {
     let filename = pre_lto_bitcode_filename(&module.name);
-    let bc_path = in_incr_comp_dir_sess(tcx.sess, &filename);
+    let bitcode_path = in_incr_comp_dir_sess(tcx.sess, &filename);
-    let file = fs::File::open(&bc_path)
-        .unwrap_or_else(|e| panic!("failed to open bitcode file `{}`: {}", bc_path.display(), e));
-
-    let mmap = unsafe {
-        Mmap::map(file).unwrap_or_else(|e| {
-            panic!("failed to mmap bitcode file `{}`: {}", bc_path.display(), e)
-        })
-    };
     // Schedule the module to be loaded
-    drop(coordinator.sender.send(Message::AddImportOnlyModule::<B> {
+    drop(
-        module_data: SerializedModule::FromUncompressedFile(mmap),
+        coordinator
-        work_product: module.source,
+            .sender
-    }));
+            .send(Message::AddImportOnlyModule::<B> { bitcode_path, work_product: module.source }),
+    );
 }
 
 fn pre_lto_bitcode_filename(module_name: &str) -> String {