Auto merge of #115964 - bjorn3:cgu_reuse_tracker_global_state, r=cjgillot

Remove cgu_reuse_tracker from Session

This removes a bit of global mutable state.

It will now miss post-lto cgu reuse when ThinLTO determines that a cgu doesn't get changed, but there weren't any tests for this anyway and a test for it would be fragile to the exact implementation of ThinLTO in LLVM.

Author: bors

Cargo.lock

@@ -3650,6 +3650,7 @@ dependencies = [
  "serde_json",
  "smallvec",
  "tempfile",
+ "thin-vec",
  "thorin-dwp",
  "tracing",
  "windows",

compiler/rustc_codegen_cranelift/src/driver/aot.rs

@@ -7,14 +7,15 @@ use std::sync::Arc;
 use std::thread::JoinHandle;
 
 use cranelift_object::{ObjectBuilder, ObjectModule};
+use rustc_codegen_ssa::assert_module_sources::CguReuse;
 use rustc_codegen_ssa::back::metadata::create_compressed_metadata_file;
+use rustc_codegen_ssa::base::determine_cgu_reuse;
 use rustc_codegen_ssa::{CodegenResults, CompiledModule, CrateInfo, ModuleKind};
 use rustc_data_structures::profiling::SelfProfilerRef;
 use rustc_data_structures::stable_hasher::{HashStable, StableHasher};
 use rustc_metadata::EncodedMetadata;
 use rustc_middle::dep_graph::{WorkProduct, WorkProductId};
 use rustc_middle::mir::mono::{CodegenUnit, MonoItem};
-use rustc_session::cgu_reuse_tracker::CguReuse;
 use rustc_session::config::{DebugInfo, OutputFilenames, OutputType};
 use rustc_session::Session;
 
@@ -374,43 +375,47 @@ pub(crate) fn run_aot(
         }
     }
 
+    // Calculate the CGU reuse
+    let cgu_reuse = tcx.sess.time("find_cgu_reuse", || {
+        cgus.iter().map(|cgu| determine_cgu_reuse(tcx, &cgu)).collect::<Vec<_>>()
+    });
+
+    rustc_codegen_ssa::assert_module_sources::assert_module_sources(tcx, &|cgu_reuse_tracker| {
+        for (i, cgu) in cgus.iter().enumerate() {
+            let cgu_reuse = cgu_reuse[i];
+            cgu_reuse_tracker.set_actual_reuse(cgu.name().as_str(), cgu_reuse);
+        }
+    });
+
     let global_asm_config = Arc::new(crate::global_asm::GlobalAsmConfig::new(tcx));
 
     let mut concurrency_limiter = ConcurrencyLimiter::new(tcx.sess, cgus.len());
 
     let modules = tcx.sess.time("codegen mono items", || {
         cgus.iter()
-            .map(|cgu| {
-                let cgu_reuse = if backend_config.disable_incr_cache {
-                    CguReuse::No
-                } else {
-                    determine_cgu_reuse(tcx, cgu)
-                };
-                tcx.sess.cgu_reuse_tracker.set_actual_reuse(cgu.name().as_str(), cgu_reuse);
-
-                match cgu_reuse {
-                    CguReuse::No => {
-                        let dep_node = cgu.codegen_dep_node(tcx);
-                        tcx.dep_graph
-                            .with_task(
-                                dep_node,
-                                tcx,
-                                (
-                                    backend_config.clone(),
-                                    global_asm_config.clone(),
-                                    cgu.name(),
-                                    concurrency_limiter.acquire(tcx.sess.diagnostic()),
-                                ),
-                                module_codegen,
-                                Some(rustc_middle::dep_graph::hash_result),
-                            )
-                            .0
-                    }
-                    CguReuse::PreLto => unreachable!(),
-                    CguReuse::PostLto => {
-                        concurrency_limiter.job_already_done();
-                        OngoingModuleCodegen::Sync(reuse_workproduct_for_cgu(tcx, cgu))
-                    }
+            .enumerate()
+            .map(|(i, cgu)| match cgu_reuse[i] {
+                CguReuse::No => {
+                    let dep_node = cgu.codegen_dep_node(tcx);
+                    tcx.dep_graph
+                        .with_task(
+                            dep_node,
+                            tcx,
+                            (
+                                backend_config.clone(),
+                                global_asm_config.clone(),
+                                cgu.name(),
+                                concurrency_limiter.acquire(tcx.sess.diagnostic()),
+                            ),
+                            module_codegen,
+                            Some(rustc_middle::dep_graph::hash_result),
+                        )
+                        .0
+                }
+                CguReuse::PreLto => unreachable!("LTO not yet supported"),
+                CguReuse::PostLto => {
+                    concurrency_limiter.job_already_done();
+                    OngoingModuleCodegen::Sync(reuse_workproduct_for_cgu(tcx, cgu))
                 }
             })
             .collect::<Vec<_>>()
@@ -489,32 +494,3 @@ pub(crate) fn run_aot(
         concurrency_limiter,
     })
 }
-
-// Adapted from https://github.com/rust-lang/rust/blob/303d8aff6092709edd4dbd35b1c88e9aa40bf6d8/src/librustc_codegen_ssa/base.rs#L922-L953
-fn determine_cgu_reuse<'tcx>(tcx: TyCtxt<'tcx>, cgu: &CodegenUnit<'tcx>) -> CguReuse {
-    if !tcx.dep_graph.is_fully_enabled() {
-        return CguReuse::No;
-    }
-
-    let work_product_id = &cgu.work_product_id();
-    if tcx.dep_graph.previous_work_product(work_product_id).is_none() {
-        // We don't have anything cached for this CGU. This can happen
-        // if the CGU did not exist in the previous session.
-        return CguReuse::No;
-    }
-
-    // Try to mark the CGU as green. If it we can do so, it means that nothing
-    // affecting the LLVM module has changed and we can re-use a cached version.
-    // If we compile with any kind of LTO, this means we can re-use the bitcode
-    // of the Pre-LTO stage (possibly also the Post-LTO version but we'll only
-    // know that later). If we are not doing LTO, there is only one optimized
-    // version of each module, so we re-use that.
-    let dep_node = cgu.codegen_dep_node(tcx);
-    assert!(
-        !tcx.dep_graph.dep_node_exists(&dep_node),
-        "CompileCodegenUnit dep-node for CGU `{}` already exists before marking.",
-        cgu.name()
-    );
-
-    if tcx.try_mark_green(&dep_node) { CguReuse::PostLto } else { CguReuse::No }
-}

compiler/rustc_codegen_llvm/src/back/lto.rs

@@ -19,7 +19,6 @@ use rustc_hir::def_id::LOCAL_CRATE;
 use rustc_middle::bug;
 use rustc_middle::dep_graph::WorkProduct;
 use rustc_middle::middle::exported_symbols::{SymbolExportInfo, SymbolExportLevel};
-use rustc_session::cgu_reuse_tracker::CguReuse;
 use rustc_session::config::{self, CrateType, Lto};
 
 use std::ffi::{CStr, CString};
@@ -584,7 +583,6 @@ fn thin_lto(
                     copy_jobs.push(work_product);
                     info!(" - {}: re-used", module_name);
                     assert!(cgcx.incr_comp_session_dir.is_some());
-                    cgcx.cgu_reuse_tracker.set_actual_reuse(module_name, CguReuse::PostLto);
                     continue;
                 }
             }

compiler/rustc_codegen_ssa/Cargo.toml

@@ -16,6 +16,7 @@ pathdiff = "0.2.0"
 serde_json = "1.0.59"
 smallvec = { version = "1.8.1", features = ["union", "may_dangle"] }
 regex = "1.4"
+thin-vec = "0.2.12"
 
 rustc_serialize = { path = "../rustc_serialize" }
 rustc_arena = { path = "../rustc_arena" }

compiler/rustc_codegen_ssa/messages.ftl

@@ -11,6 +11,9 @@ codegen_ssa_atomic_compare_exchange = Atomic compare-exchange intrinsic missing
 
 codegen_ssa_binary_output_to_tty = option `-o` or `--emit` is used to write binary output type `{$shorthand}` to stdout, but stdout is a tty
 
+codegen_ssa_cgu_not_recorded =
+    CGU-reuse for `{$cgu_user_name}` is (mangled: `{$cgu_name}`) was not recorded
+
 codegen_ssa_check_installed_visual_studio = please ensure that Visual Studio 2017 or later, or Build Tools for Visual Studio were installed with the Visual C++ option.
 
 codegen_ssa_copy_path = could not copy {$from} to {$to}: {$error}
@@ -39,13 +42,21 @@ codegen_ssa_failed_to_get_layout = failed to get layout for {$ty}: {$err}
 
 codegen_ssa_failed_to_write = failed to write {$path}: {$error}
 
+codegen_ssa_field_associated_value_expected = associated value expected for `{$name}`
+
 codegen_ssa_ignoring_emit_path = ignoring emit path because multiple .{$extension} files were produced
 
 codegen_ssa_ignoring_output = ignoring -o because multiple .{$extension} files were produced
 
 codegen_ssa_illegal_link_ordinal_format = illegal ordinal format in `link_ordinal`
     .note = an unsuffixed integer value, e.g., `1`, is expected
 
+codegen_ssa_incorrect_cgu_reuse_type =
+    CGU-reuse for `{$cgu_user_name}` is `{$actual_reuse}` but should be {$at_least ->
+    [one] {"at least "}
+    *[other] {""}
+    }`{$expected_reuse}`
+
 codegen_ssa_insufficient_vs_code_product = VS Code is a different product, and is not sufficient.
 
 codegen_ssa_invalid_link_ordinal_nargs = incorrect number of arguments to `#[link_ordinal]`
@@ -153,19 +164,30 @@ codegen_ssa_linker_unsupported_modifier = `as-needed` modifier not supported for
 
 codegen_ssa_linking_failed = linking with `{$linker_path}` failed: {$exit_status}
 
+codegen_ssa_malformed_cgu_name =
+    found malformed codegen unit name `{$user_path}`. codegen units names must always start with the name of the crate (`{$crate_name}` in this case).
+
 codegen_ssa_metadata_object_file_write = error writing metadata object file: {$error}
 
 codegen_ssa_missing_cpp_build_tool_component = or a necessary component may be missing from the "C++ build tools" workload
 
 codegen_ssa_missing_memory_ordering = Atomic intrinsic missing memory ordering
 
+codegen_ssa_missing_query_depgraph =
+    found CGU-reuse attribute but `-Zquery-dep-graph` was not specified
+
 codegen_ssa_msvc_missing_linker = the msvc targets depend on the msvc linker but `link.exe` was not found
 
 codegen_ssa_multiple_external_func_decl = multiple declarations of external function `{$function}` from library `{$library_name}` have different calling conventions
 
 codegen_ssa_multiple_main_functions = entry symbol `main` declared multiple times
     .help = did you use `#[no_mangle]` on `fn main`? Use `#[start]` instead
 
+codegen_ssa_no_field = no field `{$name}`
+
+codegen_ssa_no_module_named =
+    no module named `{$user_path}` (mangled: {$cgu_name}). available modules: {$cgu_names}
+
 codegen_ssa_no_natvis_directory = error enumerating natvis directory: {$error}
 
 codegen_ssa_processing_dymutil_failed = processing debug info with `dsymutil` failed: {$status}
@@ -297,6 +319,8 @@ codegen_ssa_unknown_atomic_operation = unknown atomic operation
 
 codegen_ssa_unknown_atomic_ordering = unknown ordering in atomic intrinsic
 
+codegen_ssa_unknown_reuse_kind = unknown cgu-reuse-kind `{$kind}` specified
+
 codegen_ssa_unsupported_arch = unsupported arch `{$arch}` for os `{$os}`
 
 codegen_ssa_unsupported_link_self_contained = option `-C link-self-contained` is not supported on this target