Unrolled build for #143689

Rollup merge of #143689 - pmur:murp/external-rt-optimized-compiler-builtins, r=Kobzol,tgross35

Allow linking a prebuilt optimized compiler-rt builtins library

Extend the <target>.optimized-compiler-builtins bootstrap option to accept a path to a prebuilt compiler-rt builtins library, and update compiler-builtins to enable optimized builtins without building compiler-rt builtins.

Author: rust-timer

bootstrap.example.toml

@@ -407,8 +407,11 @@
 #build.profiler = false
 
 # Use the optimized LLVM C intrinsics for `compiler_builtins`, rather than Rust intrinsics.
-# Requires the LLVM submodule to be managed by bootstrap (i.e. not external) so that `compiler-rt`
-# sources are available.
+# Choosing true requires the LLVM submodule to be managed by bootstrap (i.e. not external)
+# so that `compiler-rt` sources are available.
+#
+# Setting this to a path removes the requirement for a C toolchain, but requires setting the
+# path to an existing library containing the builtins library from LLVM's compiler-rt.
 #
 # Setting this to `false` generates slower code, but removes the requirement for a C toolchain in
 # order to run `x check`.
@@ -1041,13 +1044,15 @@
 #runner = <none> (string)
 
 # Use the optimized LLVM C intrinsics for `compiler_builtins`, rather than Rust intrinsics
-# on this target.
-# Requires the LLVM submodule to be managed by bootstrap (i.e. not external) so that `compiler-rt`
-# sources are available.
+# on this target. Choosing true requires the LLVM submodule to be managed by bootstrap
+# (i.e. not external) so that `compiler-rt` sources are available.
+#
+# Setting this to a path removes the requirement for a C toolchain, but requires setting the
+# path to an existing library containing the builtins library from LLVM's compiler-rt.
 #
 # Setting this to `false` generates slower code, but removes the requirement for a C toolchain in
 # order to run `x check`.
-#optimized-compiler-builtins = build.optimized-compiler-builtins (bool)
+#optimized-compiler-builtins = build.optimized-compiler-builtins (bool or path)
 
 # Link the compiler and LLVM against `jemalloc` instead of the default libc allocator.
 # This overrides the global `rust.jemalloc` option. See that option for more info.

library/compiler-builtins/compiler-builtins/README.md

@@ -10,6 +10,16 @@ to be added as an explicit dependency in `Cargo.toml`.
 
 [`compiler-rt`]: https://github.com/llvm/llvm-project/tree/1b1dc505057322f4fa1110ef4f53c44347f52986/compiler-rt
 
+## Configuration
+
+`compiler-builtins` can be configured with the following environment variables when the `c` feature
+is enabled:
+
+- `LLVM_COMPILER_RT_LIB`
+- `RUST_COMPILER_RT_ROOT`
+
+See `build.rs` for details.
+
 ## Contributing
 
 See [CONTRIBUTING.md](CONTRIBUTING.md).

library/compiler-builtins/compiler-builtins/build.rs

@@ -540,20 +540,28 @@ mod c {
             sources.extend(&[("__emutls_get_address", "emutls.c")]);
         }
 
+        // Optionally, link against a prebuilt llvm compiler-rt containing the builtins
+        // library. Only the builtins library is required. On many platforms, this is
+        // available as a library named libclang_rt.builtins.a.
+        let link_against_prebuilt_rt = env::var_os("LLVM_COMPILER_RT_LIB").is_some();
+
         // When compiling the C code we require the user to tell us where the
         // source code is, and this is largely done so when we're compiling as
         // part of rust-lang/rust we can use the same llvm-project repository as
         // rust-lang/rust.
         let root = match env::var_os("RUST_COMPILER_RT_ROOT") {
             Some(s) => PathBuf::from(s),
+            // If a prebuild libcompiler-rt is provided, set a valid
+            // path to simplify later logic. Nothing should be compiled.
+            None if link_against_prebuilt_rt => PathBuf::new(),
             None => {
                 panic!(
                     "RUST_COMPILER_RT_ROOT is not set. You may need to run \
                     `ci/download-compiler-rt.sh`."
                 );
             }
         };
-        if !root.exists() {
+        if !link_against_prebuilt_rt && !root.exists() {
             panic!("RUST_COMPILER_RT_ROOT={} does not exist", root.display());
         }
 
@@ -569,7 +577,7 @@ mod c {
         let src_dir = root.join("lib/builtins");
         if target.arch == "aarch64" && target.env != "msvc" && target.os != "uefi" {
             // See below for why we're building these as separate libraries.
-            build_aarch64_out_of_line_atomics_libraries(&src_dir, cfg);
+            build_aarch64_out_of_line_atomics_libraries(&src_dir, cfg, link_against_prebuilt_rt);
 
             // Some run-time CPU feature detection is necessary, as well.
             let cpu_model_src = if src_dir.join("cpu_model.c").exists() {
@@ -583,20 +591,45 @@ mod c {
         let mut added_sources = HashSet::new();
         for (sym, src) in sources.map.iter() {
             let src = src_dir.join(src);
-            if added_sources.insert(src.clone()) {
+            if !link_against_prebuilt_rt && added_sources.insert(src.clone()) {
                 cfg.file(&src);
                 println!("cargo:rerun-if-changed={}", src.display());
             }
             println!("cargo:rustc-cfg={}=\"optimized-c\"", sym);
         }
 
-        cfg.compile("libcompiler-rt.a");
+        if link_against_prebuilt_rt {
+            let rt_builtins_ext = PathBuf::from(env::var_os("LLVM_COMPILER_RT_LIB").unwrap());
+            if !rt_builtins_ext.exists() {
+                panic!(
+                    "LLVM_COMPILER_RT_LIB={} does not exist",
+                    rt_builtins_ext.display()
+                );
+            }
+            if let Some(dir) = rt_builtins_ext.parent() {
+                println!("cargo::rustc-link-search=native={}", dir.display());
+            }
+            if let Some(lib) = rt_builtins_ext.file_name() {
+                println!(
+                    "cargo::rustc-link-lib=static:+verbatim={}",
+                    lib.to_str().unwrap()
+                );
+            }
+        } else {
+            cfg.compile("libcompiler-rt.a");
+        }
     }
 
-    fn build_aarch64_out_of_line_atomics_libraries(builtins_dir: &Path, cfg: &mut cc::Build) {
+    fn build_aarch64_out_of_line_atomics_libraries(
+        builtins_dir: &Path,
+        cfg: &mut cc::Build,
+        link_against_prebuilt_rt: bool,
+    ) {
         let out_dir = PathBuf::from(env::var("OUT_DIR").unwrap());
         let outlined_atomics_file = builtins_dir.join("aarch64").join("lse.S");
-        println!("cargo:rerun-if-changed={}", outlined_atomics_file.display());
+        if !link_against_prebuilt_rt {
+            println!("cargo:rerun-if-changed={}", outlined_atomics_file.display());
+        }
 
         cfg.include(&builtins_dir);
 
@@ -609,6 +642,13 @@ mod c {
                 for (model_number, model_name) in
                     &[(1, "relax"), (2, "acq"), (3, "rel"), (4, "acq_rel")]
                 {
+                    let sym = format!("__aarch64_{}{}_{}", instruction_type, size, model_name);
+                    println!("cargo:rustc-cfg={}=\"optimized-c\"", sym);
+
+                    if link_against_prebuilt_rt {
+                        continue;
+                    }
+
                     // The original compiler-rt build system compiles the same
                     // source file multiple times with different compiler
                     // options. Here we do something slightly different: we
@@ -632,9 +672,6 @@ mod c {
                     .unwrap();
                     drop(file);
                     cfg.file(path);
-
-                    let sym = format!("__aarch64_{}{}_{}", instruction_type, size, model_name);
-                    println!("cargo:rustc-cfg={}=\"optimized-c\"", sym);
                 }
             }
         }

src/bootstrap/src/core/build_steps/compile.rs

@@ -26,7 +26,9 @@ use crate::core::builder;
 use crate::core::builder::{
     Builder, Cargo, Kind, RunConfig, ShouldRun, Step, StepMetadata, crate_description,
 };
-use crate::core::config::{DebuginfoLevel, LlvmLibunwind, RustcLto, TargetSelection};
+use crate::core::config::{
+    CompilerBuiltins, DebuginfoLevel, LlvmLibunwind, RustcLto, TargetSelection,
+};
 use crate::utils::build_stamp;
 use crate::utils::build_stamp::BuildStamp;
 use crate::utils::exec::command;
@@ -574,29 +576,36 @@ pub fn std_cargo(builder: &Builder<'_>, target: TargetSelection, cargo: &mut Car
     // If `compiler-rt` is available ensure that the `c` feature of the
     // `compiler-builtins` crate is enabled and it's configured to learn where
     // `compiler-rt` is located.
-    let compiler_builtins_c_feature = if builder.config.optimized_compiler_builtins(target) {
-        // NOTE: this interacts strangely with `llvm-has-rust-patches`. In that case, we enforce `submodules = false`, so this is a no-op.
-        // But, the user could still decide to manually use an in-tree submodule.
-        //
-        // NOTE: if we're using system llvm, we'll end up building a version of `compiler-rt` that doesn't match the LLVM we're linking to.
-        // That's probably ok? At least, the difference wasn't enforced before. There's a comment in
-        // the compiler_builtins build script that makes me nervous, though:
-        // https://github.com/rust-lang/compiler-builtins/blob/31ee4544dbe47903ce771270d6e3bea8654e9e50/build.rs#L575-L579
-        builder.require_submodule(
-            "src/llvm-project",
-            Some(
-                "The `build.optimized-compiler-builtins` config option \
-                 requires `compiler-rt` sources from LLVM.",
-            ),
-        );
-        let compiler_builtins_root = builder.src.join("src/llvm-project/compiler-rt");
-        assert!(compiler_builtins_root.exists());
-        // The path to `compiler-rt` is also used by `profiler_builtins` (above),
-        // so if you're changing something here please also change that as appropriate.
-        cargo.env("RUST_COMPILER_RT_ROOT", &compiler_builtins_root);
-        " compiler-builtins-c"
-    } else {
-        ""
+    let compiler_builtins_c_feature = match builder.config.optimized_compiler_builtins(target) {
+        CompilerBuiltins::LinkLLVMBuiltinsLib(path) => {
+            cargo.env("LLVM_COMPILER_RT_LIB", path);
+            " compiler-builtins-c"
+        }
+        CompilerBuiltins::BuildLLVMFuncs => {
+            // NOTE: this interacts strangely with `llvm-has-rust-patches`. In that case, we enforce
+            // `submodules = false`, so this is a no-op. But, the user could still decide to
+            //  manually use an in-tree submodule.
+            //
+            // NOTE: if we're using system llvm, we'll end up building a version of `compiler-rt`
+            // that doesn't match the LLVM we're linking to. That's probably ok? At least, the
+            // difference wasn't enforced before. There's a comment in the compiler_builtins build
+            // script that makes me nervous, though:
+            // https://github.com/rust-lang/compiler-builtins/blob/31ee4544dbe47903ce771270d6e3bea8654e9e50/build.rs#L575-L579
+            builder.require_submodule(
+                "src/llvm-project",
+                Some(
+                    "The `build.optimized-compiler-builtins` config option \
+                     requires `compiler-rt` sources from LLVM.",
+                ),
+            );
+            let compiler_builtins_root = builder.src.join("src/llvm-project/compiler-rt");
+            assert!(compiler_builtins_root.exists());
+            // The path to `compiler-rt` is also used by `profiler_builtins` (above),
+            // so if you're changing something here please also change that as appropriate.
+            cargo.env("RUST_COMPILER_RT_ROOT", &compiler_builtins_root);
+            " compiler-builtins-c"
+        }
+        CompilerBuiltins::BuildRustOnly => "",
     };
 
     // `libtest` uses this to know whether or not to support

src/bootstrap/src/core/config/config.rs

@@ -46,8 +46,8 @@ use crate::core::config::toml::rust::{
 };
 use crate::core::config::toml::target::Target;
 use crate::core::config::{
-    DebuginfoLevel, DryRun, GccCiMode, LlvmLibunwind, Merge, ReplaceOpt, RustcLto, SplitDebuginfo,
-    StringOrBool, threads_from_config,
+    CompilerBuiltins, DebuginfoLevel, DryRun, GccCiMode, LlvmLibunwind, Merge, ReplaceOpt,
+    RustcLto, SplitDebuginfo, StringOrBool, threads_from_config,
 };
 use crate::core::download::{
     DownloadContext, download_beta_toolchain, is_download_ci_available, maybe_download_rustfmt,
@@ -121,8 +121,7 @@ pub struct Config {
     pub patch_binaries_for_nix: Option<bool>,
     pub stage0_metadata: build_helper::stage0_parser::Stage0,
     pub android_ndk: Option<PathBuf>,
-    /// Whether to use the `c` feature of the `compiler_builtins` crate.
-    pub optimized_compiler_builtins: bool,
+    pub optimized_compiler_builtins: CompilerBuiltins,
 
     pub stdout_is_tty: bool,
     pub stderr_is_tty: bool,
@@ -1109,7 +1108,11 @@ impl Config {
         let rustfmt_info = git_info(&exec_ctx, omit_git_hash, &src.join("src/tools/rustfmt"));
 
         let optimized_compiler_builtins =
-            build_optimized_compiler_builtins.unwrap_or(channel != "dev");
+            build_optimized_compiler_builtins.unwrap_or(if channel == "dev" {
+                CompilerBuiltins::BuildRustOnly
+            } else {
+                CompilerBuiltins::BuildLLVMFuncs
+            });
         let vendor = build_vendor.unwrap_or(
             rust_info.is_from_tarball()
                 && src.join("vendor").exists()
@@ -1672,11 +1675,11 @@ impl Config {
         self.target_config.get(&target).and_then(|t| t.rpath).unwrap_or(self.rust_rpath)
     }
 
-    pub fn optimized_compiler_builtins(&self, target: TargetSelection) -> bool {
+    pub fn optimized_compiler_builtins(&self, target: TargetSelection) -> &CompilerBuiltins {
         self.target_config
             .get(&target)
-            .and_then(|t| t.optimized_compiler_builtins)
-            .unwrap_or(self.optimized_compiler_builtins)
+            .and_then(|t| t.optimized_compiler_builtins.as_ref())
+            .unwrap_or(&self.optimized_compiler_builtins)
     }
 
     pub fn llvm_enabled(&self, target: TargetSelection) -> bool {

src/bootstrap/src/core/config/mod.rs

@@ -218,6 +218,33 @@ impl<T> Merge for Option<T> {
     }
 }
 
+#[derive(Clone, Debug, Default, Eq, PartialEq)]
+pub enum CompilerBuiltins {
+    #[default]
+    // Only build native rust intrinsic compiler functions.
+    BuildRustOnly,
+    // Some intrinsic functions have a C implementation provided by LLVM's
+    // compiler-rt builtins library. Build them from the LLVM source included
+    // with Rust.
+    BuildLLVMFuncs,
+    // Similar to BuildLLVMFuncs, but specify a path to an existing library
+    // containing LLVM's compiler-rt builtins instead of compiling them.
+    LinkLLVMBuiltinsLib(String),
+}
+
+impl<'de> Deserialize<'de> for CompilerBuiltins {
+    fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
+    where
+        D: Deserializer<'de>,
+    {
+        Ok(match Deserialize::deserialize(deserializer)? {
+            StringOrBool::Bool(false) => Self::BuildRustOnly,
+            StringOrBool::Bool(true) => Self::BuildLLVMFuncs,
+            StringOrBool::String(path) => Self::LinkLLVMBuiltinsLib(path),
+        })
+    }
+}
+
 #[derive(Copy, Clone, Default, Debug, Eq, PartialEq)]
 pub enum DebuginfoLevel {
     #[default]

src/bootstrap/src/core/config/tests.rs

@@ -17,7 +17,7 @@ use crate::core::build_steps::clippy::{LintConfig, get_clippy_rules_in_order};
 use crate::core::build_steps::llvm;
 use crate::core::build_steps::llvm::LLVM_INVALIDATION_PATHS;
 use crate::core::config::toml::TomlConfig;
-use crate::core::config::{LldMode, Target, TargetSelection};
+use crate::core::config::{CompilerBuiltins, LldMode, StringOrBool, Target, TargetSelection};
 use crate::utils::tests::git::git_test;
 
 pub(crate) fn parse(config: &str) -> Config {
@@ -183,7 +183,11 @@ runner = "x86_64-runner"
     );
     assert_eq!(config.gdb, Some("bar".into()), "setting string value with quotes");
     assert!(!config.deny_warnings, "setting boolean value");
-    assert!(config.optimized_compiler_builtins, "setting boolean value");
+    assert_eq!(
+        config.optimized_compiler_builtins,
+        CompilerBuiltins::BuildLLVMFuncs,
+        "setting boolean value"
+    );
     assert_eq!(
         config.tools,
         Some(["cargo".to_string()].into_iter().collect()),
@@ -212,7 +216,7 @@ runner = "x86_64-runner"
     let darwin = TargetSelection::from_user("aarch64-apple-darwin");
     let darwin_values = Target {
         runner: Some("apple".into()),
-        optimized_compiler_builtins: Some(false),
+        optimized_compiler_builtins: Some(CompilerBuiltins::BuildRustOnly),
         ..Default::default()
     };
     assert_eq!(

src/bootstrap/src/core/config/toml/build.rs

@@ -11,7 +11,7 @@ use std::collections::HashMap;
 use serde::{Deserialize, Deserializer};
 
 use crate::core::config::toml::ReplaceOpt;
-use crate::core::config::{Merge, StringOrBool};
+use crate::core::config::{CompilerBuiltins, Merge, StringOrBool};
 use crate::{HashSet, PathBuf, define_config, exit};
 
 define_config! {
@@ -65,7 +65,7 @@ define_config! {
         // NOTE: only parsed by bootstrap.py, `--feature build-metrics` enables metrics unconditionally
         metrics: Option<bool> = "metrics",
         android_ndk: Option<PathBuf> = "android-ndk",
-        optimized_compiler_builtins: Option<bool> = "optimized-compiler-builtins",
+        optimized_compiler_builtins: Option<CompilerBuiltins> = "optimized-compiler-builtins",
         jobs: Option<u32> = "jobs",
         compiletest_diff_tool: Option<String> = "compiletest-diff-tool",
         compiletest_allow_stage0: Option<bool> = "compiletest-allow-stage0",

src/bootstrap/src/core/config/toml/rust.rs

@@ -269,9 +269,9 @@ pub fn check_incompatible_options_for_ci_rustc(
     err!(current_profiler, profiler, "build");
 
     let current_optimized_compiler_builtins =
-        current_config_toml.build.as_ref().and_then(|b| b.optimized_compiler_builtins);
+        current_config_toml.build.as_ref().and_then(|b| b.optimized_compiler_builtins.clone());
     let optimized_compiler_builtins =
-        ci_config_toml.build.as_ref().and_then(|b| b.optimized_compiler_builtins);
+        ci_config_toml.build.as_ref().and_then(|b| b.optimized_compiler_builtins.clone());
     err!(current_optimized_compiler_builtins, optimized_compiler_builtins, "build");
 
     // We always build the in-tree compiler on cross targets, so we only care