Auto merge of #141272 - Shourya742:2025-05-18-modularize-config-module, r=Kobzol

modularize the config module bootstrap

Currently, our `config` module is quite large over 3,000 lines, and handles a wide range of responsibilities. This PR aims to break it down into smaller, more focused submodules to improve readability and maintainability:

* **`toml`**: Introduces a dedicated `toml` submodule within the `config` module. Its sole purpose is to define configuration-related structs along with their corresponding deserialization logic. It also contains the `parse_inner` method, which serves as the central function for extracting relevant information from the TOML structs and constructing the final configuration.

* **`rust`, `dist`, `install`, `llvm`, `build`, `gcc`, and others**: Each of these modules contains TOML subsections specific to their domain, along with the logic necessary to convert them into parts of the final configuration struct.

* **`config/mod.rs`**: Contains shared types and enums used across multiple TOML subsections.

* **`config/config.rs`**: Houses the logic that integrates all the TOML subsections into the complete configuration struct.

r? `@kobzol`

Author: bors

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

@@ -12,7 +12,8 @@ use tracing::instrument;
 use crate::core::build_steps::perf::PerfArgs;
 use crate::core::build_steps::setup::Profile;
 use crate::core::builder::{Builder, Kind};
-use crate::core::config::{Config, TargetSelectionList, target_selection_list};
+use crate::core::config::Config;
+use crate::core::config::target_selection::{TargetSelectionList, target_selection_list};
 use crate::{Build, DocTests};
 
 #[derive(Copy, Clone, Default, Debug, ValueEnum)]

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

@@ -0,0 +1,147 @@
+use std::fmt;
+
+use crate::core::config::SplitDebuginfo;
+use crate::utils::cache::{INTERNER, Interned};
+use crate::{Path, env};
+
+#[derive(Copy, Clone, Default, PartialEq, Eq, PartialOrd, Ord, Hash)]
+// N.B.: This type is used everywhere, and the entire codebase relies on it being Copy.
+// Making !Copy is highly nontrivial!
+pub struct TargetSelection {
+    pub triple: Interned<String>,
+    pub file: Option<Interned<String>>,
+    pub synthetic: bool,
+}
+
+/// Newtype over `Vec<TargetSelection>` so we can implement custom parsing logic
+#[derive(Clone, Default, PartialEq, Eq, PartialOrd, Ord, Hash, Debug)]
+pub struct TargetSelectionList(pub Vec<TargetSelection>);
+
+pub fn target_selection_list(s: &str) -> Result<TargetSelectionList, String> {
+    Ok(TargetSelectionList(
+        s.split(',').filter(|s| !s.is_empty()).map(TargetSelection::from_user).collect(),
+    ))
+}
+
+impl TargetSelection {
+    pub fn from_user(selection: &str) -> Self {
+        let path = Path::new(selection);
+
+        let (triple, file) = if path.exists() {
+            let triple = path
+                .file_stem()
+                .expect("Target specification file has no file stem")
+                .to_str()
+                .expect("Target specification file stem is not UTF-8");
+
+            (triple, Some(selection))
+        } else {
+            (selection, None)
+        };
+
+        let triple = INTERNER.intern_str(triple);
+        let file = file.map(|f| INTERNER.intern_str(f));
+
+        Self { triple, file, synthetic: false }
+    }
+
+    pub fn create_synthetic(triple: &str, file: &str) -> Self {
+        Self {
+            triple: INTERNER.intern_str(triple),
+            file: Some(INTERNER.intern_str(file)),
+            synthetic: true,
+        }
+    }
+
+    pub fn rustc_target_arg(&self) -> &str {
+        self.file.as_ref().unwrap_or(&self.triple)
+    }
+
+    pub fn contains(&self, needle: &str) -> bool {
+        self.triple.contains(needle)
+    }
+
+    pub fn starts_with(&self, needle: &str) -> bool {
+        self.triple.starts_with(needle)
+    }
+
+    pub fn ends_with(&self, needle: &str) -> bool {
+        self.triple.ends_with(needle)
+    }
+
+    // See src/bootstrap/synthetic_targets.rs
+    pub fn is_synthetic(&self) -> bool {
+        self.synthetic
+    }
+
+    pub fn is_msvc(&self) -> bool {
+        self.contains("msvc")
+    }
+
+    pub fn is_windows(&self) -> bool {
+        self.contains("windows")
+    }
+
+    pub fn is_windows_gnu(&self) -> bool {
+        self.ends_with("windows-gnu")
+    }
+
+    pub fn is_cygwin(&self) -> bool {
+        self.is_windows() &&
+        // ref. https://cygwin.com/pipermail/cygwin/2022-February/250802.html
+        env::var("OSTYPE").is_ok_and(|v| v.to_lowercase().contains("cygwin"))
+    }
+
+    pub fn needs_crt_begin_end(&self) -> bool {
+        self.contains("musl") && !self.contains("unikraft")
+    }
+
+    /// Path to the file defining the custom target, if any.
+    pub fn filepath(&self) -> Option<&Path> {
+        self.file.as_ref().map(Path::new)
+    }
+}
+
+impl fmt::Display for TargetSelection {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "{}", self.triple)?;
+        if let Some(file) = self.file {
+            write!(f, "({file})")?;
+        }
+        Ok(())
+    }
+}
+
+impl fmt::Debug for TargetSelection {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "{self}")
+    }
+}
+
+impl PartialEq<&str> for TargetSelection {
+    fn eq(&self, other: &&str) -> bool {
+        self.triple == *other
+    }
+}
+
+// Targets are often used as directory names throughout bootstrap.
+// This impl makes it more ergonomics to use them as such.
+impl AsRef<Path> for TargetSelection {
+    fn as_ref(&self) -> &Path {
+        self.triple.as_ref()
+    }
+}
+
+impl SplitDebuginfo {
+    /// Returns the default `-Csplit-debuginfo` value for the current target. See the comment for
+    /// `rust.split-debuginfo` in `bootstrap.example.toml`.
+    pub fn default_for_platform(target: TargetSelection) -> Self {
+        if target.contains("apple") {
+            SplitDebuginfo::Unpacked
+        } else if target.is_windows() {
+            SplitDebuginfo::Packed
+        } else {
+            SplitDebuginfo::Off
+        }
+    }
+}

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

@@ -10,12 +10,14 @@ use clap::CommandFactory;
 use serde::Deserialize;
 
 use super::flags::Flags;
-use super::{ChangeIdWrapper, Config, RUSTC_IF_UNCHANGED_ALLOWED_PATHS};
+use super::toml::change_id::ChangeIdWrapper;
+use super::{Config, RUSTC_IF_UNCHANGED_ALLOWED_PATHS};
 use crate::ChangeId;
 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::{LldMode, Target, TargetSelection, TomlConfig};
+use crate::core::config::toml::TomlConfig;
+use crate::core::config::{LldMode, Target, TargetSelection};
 use crate::utils::tests::git::git_test;
 
 pub(crate) fn parse(config: &str) -> Config {

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

@@ -0,0 +1,72 @@
+//! This module defines the `Build` struct, which represents the `[build]` table
+//! in the `bootstrap.toml` configuration file.
+//!
+//! The `[build]` table contains global options that influence the overall build process,
+//! such as default host and target triples, paths to tools, build directories, and
+//! various feature flags. These options apply across different stages and components
+//! unless specifically overridden by other configuration sections or command-line flags.
+
+use serde::{Deserialize, Deserializer};
+
+use crate::core::config::toml::ReplaceOpt;
+use crate::core::config::{Merge, StringOrBool};
+use crate::{HashSet, PathBuf, define_config, exit};
+
+define_config! {
+    /// TOML representation of various global build decisions.
+    #[derive(Default)]
+    struct Build {
+        build: Option<String> = "build",
+        description: Option<String> = "description",
+        host: Option<Vec<String>> = "host",
+        target: Option<Vec<String>> = "target",
+        build_dir: Option<String> = "build-dir",
+        cargo: Option<PathBuf> = "cargo",
+        rustc: Option<PathBuf> = "rustc",
+        rustfmt: Option<PathBuf> = "rustfmt",
+        cargo_clippy: Option<PathBuf> = "cargo-clippy",
+        docs: Option<bool> = "docs",
+        compiler_docs: Option<bool> = "compiler-docs",
+        library_docs_private_items: Option<bool> = "library-docs-private-items",
+        docs_minification: Option<bool> = "docs-minification",
+        submodules: Option<bool> = "submodules",
+        gdb: Option<String> = "gdb",
+        lldb: Option<String> = "lldb",
+        nodejs: Option<String> = "nodejs",
+        npm: Option<String> = "npm",
+        python: Option<String> = "python",
+        reuse: Option<String> = "reuse",
+        locked_deps: Option<bool> = "locked-deps",
+        vendor: Option<bool> = "vendor",
+        full_bootstrap: Option<bool> = "full-bootstrap",
+        bootstrap_cache_path: Option<PathBuf> = "bootstrap-cache-path",
+        extended: Option<bool> = "extended",
+        tools: Option<HashSet<String>> = "tools",
+        verbose: Option<usize> = "verbose",
+        sanitizers: Option<bool> = "sanitizers",
+        profiler: Option<bool> = "profiler",
+        cargo_native_static: Option<bool> = "cargo-native-static",
+        low_priority: Option<bool> = "low-priority",
+        configure_args: Option<Vec<String>> = "configure-args",
+        local_rebuild: Option<bool> = "local-rebuild",
+        print_step_timings: Option<bool> = "print-step-timings",
+        print_step_rusage: Option<bool> = "print-step-rusage",
+        check_stage: Option<u32> = "check-stage",
+        doc_stage: Option<u32> = "doc-stage",
+        build_stage: Option<u32> = "build-stage",
+        test_stage: Option<u32> = "test-stage",
+        install_stage: Option<u32> = "install-stage",
+        dist_stage: Option<u32> = "dist-stage",
+        bench_stage: Option<u32> = "bench-stage",
+        patch_binaries_for_nix: Option<bool> = "patch-binaries-for-nix",
+        // 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",
+        jobs: Option<u32> = "jobs",
+        compiletest_diff_tool: Option<String> = "compiletest-diff-tool",
+        compiletest_use_stage0_libtest: Option<bool> = "compiletest-use-stage0-libtest",
+        ccache: Option<StringOrBool> = "ccache",
+        exclude: Option<Vec<PathBuf>> = "exclude",
+    }
+}

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

@@ -0,0 +1,34 @@
+use serde::{Deserialize, Deserializer};
+use serde_derive::Deserialize;
+
+/// This enum is used for deserializing change IDs from TOML, allowing both numeric values and the string `"ignore"`.
+#[derive(Clone, Debug, PartialEq)]
+pub enum ChangeId {
+    Ignore,
+    Id(usize),
+}
+
+/// Since we use `#[serde(deny_unknown_fields)]` on `TomlConfig`, we need a wrapper type
+/// for the "change-id" field to parse it even if other fields are invalid. This ensures
+/// that if deserialization fails due to other fields, we can still provide the changelogs
+/// to allow developers to potentially find the reason for the failure in the logs..
+#[derive(Deserialize, Default)]
+pub(crate) struct ChangeIdWrapper {
+    #[serde(alias = "change-id", default, deserialize_with = "deserialize_change_id")]
+    pub(crate) inner: Option<ChangeId>,
+}
+
+fn deserialize_change_id<'de, D: Deserializer<'de>>(
+    deserializer: D,
+) -> Result<Option<ChangeId>, D::Error> {
+    let value = toml::Value::deserialize(deserializer)?;
+    Ok(match value {
+        toml::Value::String(s) if s == "ignore" => Some(ChangeId::Ignore),
+        toml::Value::Integer(i) => Some(ChangeId::Id(i as usize)),
+        _ => {
+            return Err(serde::de::Error::custom(
+                "expected \"ignore\" or an integer for change-id",
+            ));
+        }
+    })
+}

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

@@ -0,0 +1,52 @@
+//! This module defines the `Dist` struct, which represents the `[dist]` table
+//! in the `bootstrap.toml` configuration file.
+//!
+//! The `[dist]` table contains options related to the distribution process,
+//! including signing, uploading artifacts, source tarballs, compression settings,
+//! and inclusion of specific tools.
+
+use serde::{Deserialize, Deserializer};
+
+use crate::core::config::toml::ReplaceOpt;
+use crate::core::config::{Merge, set};
+use crate::{Config, HashSet, PathBuf, define_config, exit};
+
+define_config! {
+    struct Dist {
+        sign_folder: Option<String> = "sign-folder",
+        upload_addr: Option<String> = "upload-addr",
+        src_tarball: Option<bool> = "src-tarball",
+        compression_formats: Option<Vec<String>> = "compression-formats",
+        compression_profile: Option<String> = "compression-profile",
+        include_mingw_linker: Option<bool> = "include-mingw-linker",
+        vendor: Option<bool> = "vendor",
+    }
+}
+
+impl Config {
+    /// Applies distribution-related configuration from the `Dist` struct
+    /// to the global `Config` structure.
+    pub fn apply_dist_config(&mut self, toml_dist: Option<Dist>) {
+        if let Some(dist) = toml_dist {
+            let Dist {
+                sign_folder,
+                upload_addr,
+                src_tarball,
+                compression_formats,
+                compression_profile,
+                include_mingw_linker,
+                vendor,
+            } = dist;
+            self.dist_sign_folder = sign_folder.map(PathBuf::from);
+            self.dist_upload_addr = upload_addr;
+            self.dist_compression_formats = compression_formats;
+            set(&mut self.dist_compression_profile, compression_profile);
+            set(&mut self.rust_dist_src, src_tarball);
+            set(&mut self.dist_include_mingw_linker, include_mingw_linker);
+            self.dist_vendor = vendor.unwrap_or_else(|| {
+                // If we're building from git or tarball sources, enable it by default.
+                self.rust_info.is_managed_git_subrepository() || self.rust_info.is_from_tarball()
+            });
+        }
+    }
+}

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

@@ -0,0 +1,34 @@
+//! This module defines the `Gcc` struct, which represents the `[gcc]` table
+//! in the `bootstrap.toml` configuration file.
+//!
+//! The `[gcc]` table contains options specifically related to building or
+//! acquiring the GCC compiler for use within the Rust build process.
+
+use serde::{Deserialize, Deserializer};
+
+use crate::core::config::toml::ReplaceOpt;
+use crate::core::config::{GccCiMode, Merge};
+use crate::{Config, HashSet, PathBuf, define_config, exit};
+
+define_config! {
+    /// TOML representation of how the GCC build is configured.
+    struct Gcc {
+        download_ci_gcc: Option<bool> = "download-ci-gcc",
+    }
+}
+
+impl Config {
+    /// Applies GCC-related configuration from the `TomlGcc` struct to the
+    /// global `Config` structure.
+    pub fn apply_gcc_config(&mut self, toml_gcc: Option<Gcc>) {
+        if let Some(gcc) = toml_gcc {
+            self.gcc_ci_mode = match gcc.download_ci_gcc {
+                Some(value) => match value {
+                    true => GccCiMode::DownloadFromCi,
+                    false => GccCiMode::BuildLocally,
+                },
+                None => GccCiMode::default(),
+            };
+        }
+    }
+}

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

@@ -0,0 +1,43 @@
+//! This module defines the `Install` struct, which represents the `[install]` table
+//! in the `bootstrap.toml` configuration file.
+//!
+//! The `[install]` table contains options that specify the installation paths
+//! for various components of the Rust toolchain. These paths determine where
+//! executables, libraries, documentation, and other files will be placed
+//! during the `install` stage of the build.
+
+use serde::{Deserialize, Deserializer};
+
+use crate::core::config::toml::ReplaceOpt;
+use crate::core::config::{Merge, set};
+use crate::{Config, HashSet, PathBuf, define_config, exit};
+
+define_config! {
+    /// TOML representation of various global install decisions.
+    struct Install {
+        prefix: Option<String> = "prefix",
+        sysconfdir: Option<String> = "sysconfdir",
+        docdir: Option<String> = "docdir",
+        bindir: Option<String> = "bindir",
+        libdir: Option<String> = "libdir",
+        mandir: Option<String> = "mandir",
+        datadir: Option<String> = "datadir",
+    }
+}
+
+impl Config {
+    /// Applies installation-related configuration from the `Install` struct
+    /// to the global `Config` structure.
+    pub fn apply_install_config(&mut self, toml_install: Option<Install>) {
+        if let Some(install) = toml_install {
+            let Install { prefix, sysconfdir, docdir, bindir, libdir, mandir, datadir } = install;
+            self.prefix = prefix.map(PathBuf::from);
+            self.sysconfdir = sysconfdir.map(PathBuf::from);
+            self.datadir = datadir.map(PathBuf::from);
+            self.docdir = docdir.map(PathBuf::from);
+            set(&mut self.bindir, bindir.map(PathBuf::from));
+            self.libdir = libdir.map(PathBuf::from);
+            self.mandir = mandir.map(PathBuf::from);
+        }
+    }
+}

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

@@ -0,0 +1,284 @@
+//! This module defines the `Llvm` struct, which represents the `[llvm]` table
+//! in the `bootstrap.toml` configuration file.
+
+use serde::{Deserialize, Deserializer};
+
+use crate::core::config::toml::{Merge, ReplaceOpt, TomlConfig};
+use crate::core::config::{StringOrBool, set};
+use crate::{Config, HashMap, HashSet, PathBuf, define_config, exit};
+
+define_config! {
+    /// TOML representation of how the LLVM build is configured.
+    struct Llvm {
+        optimize: Option<bool> = "optimize",
+        thin_lto: Option<bool> = "thin-lto",
+        release_debuginfo: Option<bool> = "release-debuginfo",
+        assertions: Option<bool> = "assertions",
+        tests: Option<bool> = "tests",
+        enzyme: Option<bool> = "enzyme",
+        plugins: Option<bool> = "plugins",
+        // FIXME: Remove this field at Q2 2025, it has been replaced by build.ccache
+        ccache: Option<StringOrBool> = "ccache",
+        static_libstdcpp: Option<bool> = "static-libstdcpp",
+        libzstd: Option<bool> = "libzstd",
+        ninja: Option<bool> = "ninja",
+        targets: Option<String> = "targets",
+        experimental_targets: Option<String> = "experimental-targets",
+        link_jobs: Option<u32> = "link-jobs",
+        link_shared: Option<bool> = "link-shared",
+        version_suffix: Option<String> = "version-suffix",
+        clang_cl: Option<String> = "clang-cl",
+        cflags: Option<String> = "cflags",
+        cxxflags: Option<String> = "cxxflags",
+        ldflags: Option<String> = "ldflags",
+        use_libcxx: Option<bool> = "use-libcxx",
+        use_linker: Option<String> = "use-linker",
+        allow_old_toolchain: Option<bool> = "allow-old-toolchain",
+        offload: Option<bool> = "offload",
+        polly: Option<bool> = "polly",
+        clang: Option<bool> = "clang",
+        enable_warnings: Option<bool> = "enable-warnings",
+        download_ci_llvm: Option<StringOrBool> = "download-ci-llvm",
+        build_config: Option<HashMap<String, String>> = "build-config",
+    }
+}
+
+/// Compares the current `Llvm` options against those in the CI LLVM builder and detects any incompatible options.
+/// It does this by destructuring the `Llvm` instance to make sure every `Llvm` field is covered and not missing.
+#[cfg(not(test))]
+pub fn check_incompatible_options_for_ci_llvm(
+    current_config_toml: TomlConfig,
+    ci_config_toml: TomlConfig,
+) -> Result<(), String> {
+    macro_rules! err {
+        ($current:expr, $expected:expr) => {
+            if let Some(current) = &$current {
+                if Some(current) != $expected.as_ref() {
+                    return Err(format!(
+                        "ERROR: Setting `llvm.{}` is incompatible with `llvm.download-ci-llvm`. \
+                        Current value: {:?}, Expected value(s): {}{:?}",
+                        stringify!($expected).replace("_", "-"),
+                        $current,
+                        if $expected.is_some() { "None/" } else { "" },
+                        $expected,
+                    ));
+                };
+            };
+        };
+    }
+
+    macro_rules! warn {
+        ($current:expr, $expected:expr) => {
+            if let Some(current) = &$current {
+                if Some(current) != $expected.as_ref() {
+                    println!(
+                        "WARNING: `llvm.{}` has no effect with `llvm.download-ci-llvm`. \
+                        Current value: {:?}, Expected value(s): {}{:?}",
+                        stringify!($expected).replace("_", "-"),
+                        $current,
+                        if $expected.is_some() { "None/" } else { "" },
+                        $expected,
+                    );
+                };
+            };
+        };
+    }
+
+    let (Some(current_llvm_config), Some(ci_llvm_config)) =
+        (current_config_toml.llvm, ci_config_toml.llvm)
+    else {
+        return Ok(());
+    };
+
+    let Llvm {
+        optimize,
+        thin_lto,
+        release_debuginfo,
+        assertions: _,
+        tests: _,
+        plugins,
+        ccache: _,
+        static_libstdcpp: _,
+        libzstd,
+        ninja: _,
+        targets,
+        experimental_targets,
+        link_jobs: _,
+        link_shared: _,
+        version_suffix,
+        clang_cl,
+        cflags,
+        cxxflags,
+        ldflags,
+        use_libcxx,
+        use_linker,
+        allow_old_toolchain,
+        offload,
+        polly,
+        clang,
+        enable_warnings,
+        download_ci_llvm: _,
+        build_config,
+        enzyme,
+    } = ci_llvm_config;
+
+    err!(current_llvm_config.optimize, optimize);
+    err!(current_llvm_config.thin_lto, thin_lto);
+    err!(current_llvm_config.release_debuginfo, release_debuginfo);
+    err!(current_llvm_config.libzstd, libzstd);
+    err!(current_llvm_config.targets, targets);
+    err!(current_llvm_config.experimental_targets, experimental_targets);
+    err!(current_llvm_config.clang_cl, clang_cl);
+    err!(current_llvm_config.version_suffix, version_suffix);
+    err!(current_llvm_config.cflags, cflags);
+    err!(current_llvm_config.cxxflags, cxxflags);
+    err!(current_llvm_config.ldflags, ldflags);
+    err!(current_llvm_config.use_libcxx, use_libcxx);
+    err!(current_llvm_config.use_linker, use_linker);
+    err!(current_llvm_config.allow_old_toolchain, allow_old_toolchain);
+    err!(current_llvm_config.offload, offload);
+    err!(current_llvm_config.polly, polly);
+    err!(current_llvm_config.clang, clang);
+    err!(current_llvm_config.build_config, build_config);
+    err!(current_llvm_config.plugins, plugins);
+    err!(current_llvm_config.enzyme, enzyme);
+
+    warn!(current_llvm_config.enable_warnings, enable_warnings);
+
+    Ok(())
+}
+
+impl Config {
+    pub fn apply_llvm_config(
+        &mut self,
+        toml_llvm: Option<Llvm>,
+        ccache: &mut Option<StringOrBool>,
+    ) {
+        let mut llvm_tests = None;
+        let mut llvm_enzyme = None;
+        let mut llvm_offload = None;
+        let mut llvm_plugins = None;
+
+        if let Some(llvm) = toml_llvm {
+            let Llvm {
+                optimize: optimize_toml,
+                thin_lto,
+                release_debuginfo,
+                assertions: _,
+                tests,
+                enzyme,
+                plugins,
+                ccache: llvm_ccache,
+                static_libstdcpp,
+                libzstd,
+                ninja,
+                targets,
+                experimental_targets,
+                link_jobs,
+                link_shared,
+                version_suffix,
+                clang_cl,
+                cflags,
+                cxxflags,
+                ldflags,
+                use_libcxx,
+                use_linker,
+                allow_old_toolchain,
+                offload,
+                polly,
+                clang,
+                enable_warnings,
+                download_ci_llvm,
+                build_config,
+            } = llvm;
+            if llvm_ccache.is_some() {
+                eprintln!("Warning: llvm.ccache is deprecated. Use build.ccache instead.");
+            }
+
+            if ccache.is_none() {
+                *ccache = llvm_ccache;
+            }
+            set(&mut self.ninja_in_file, ninja);
+            llvm_tests = tests;
+            llvm_enzyme = enzyme;
+            llvm_offload = offload;
+            llvm_plugins = plugins;
+            set(&mut self.llvm_optimize, optimize_toml);
+            set(&mut self.llvm_thin_lto, thin_lto);
+            set(&mut self.llvm_release_debuginfo, release_debuginfo);
+            set(&mut self.llvm_static_stdcpp, static_libstdcpp);
+            set(&mut self.llvm_libzstd, libzstd);
+            if let Some(v) = link_shared {
+                self.llvm_link_shared.set(Some(v));
+            }
+            self.llvm_targets.clone_from(&targets);
+            self.llvm_experimental_targets.clone_from(&experimental_targets);
+            self.llvm_link_jobs = link_jobs;
+            self.llvm_version_suffix.clone_from(&version_suffix);
+            self.llvm_clang_cl.clone_from(&clang_cl);
+
+            self.llvm_cflags.clone_from(&cflags);
+            self.llvm_cxxflags.clone_from(&cxxflags);
+            self.llvm_ldflags.clone_from(&ldflags);
+            set(&mut self.llvm_use_libcxx, use_libcxx);
+            self.llvm_use_linker.clone_from(&use_linker);
+            self.llvm_allow_old_toolchain = allow_old_toolchain.unwrap_or(false);
+            self.llvm_offload = offload.unwrap_or(false);
+            self.llvm_polly = polly.unwrap_or(false);
+            self.llvm_clang = clang.unwrap_or(false);
+            self.llvm_enable_warnings = enable_warnings.unwrap_or(false);
+            self.llvm_build_config = build_config.clone().unwrap_or(Default::default());
+
+            self.llvm_from_ci = self.parse_download_ci_llvm(download_ci_llvm, self.llvm_assertions);
+
+            if self.llvm_from_ci {
+                let warn = |option: &str| {
+                    println!(
+                        "WARNING: `{option}` will only be used on `compiler/rustc_llvm` build, not for the LLVM build."
+                    );
+                    println!(
+                        "HELP: To use `{option}` for LLVM builds, set `download-ci-llvm` option to false."
+                    );
+                };
+
+                if static_libstdcpp.is_some() {
+                    warn("static-libstdcpp");
+                }
+
+                if link_shared.is_some() {
+                    warn("link-shared");
+                }
+
+                // FIXME(#129153): instead of all the ad-hoc `download-ci-llvm` checks that follow,
+                // use the `builder-config` present in tarballs since #128822 to compare the local
+                // config to the ones used to build the LLVM artifacts on CI, and only notify users
+                // if they've chosen a different value.
+
+                if libzstd.is_some() {
+                    println!(
+                        "WARNING: when using `download-ci-llvm`, the local `llvm.libzstd` option, \
+                        like almost all `llvm.*` options, will be ignored and set by the LLVM CI \
+                        artifacts builder config."
+                    );
+                    println!(
+                        "HELP: To use `llvm.libzstd` for LLVM/LLD builds, set `download-ci-llvm` option to false."
+                    );
+                }
+            }
+
+            if !self.llvm_from_ci && self.llvm_thin_lto && link_shared.is_none() {
+                // If we're building with ThinLTO on, by default we want to link
+                // to LLVM shared, to avoid re-doing ThinLTO (which happens in
+                // the link step) with each stage.
+                self.llvm_link_shared.set(Some(true));
+            }
+        } else {
+            self.llvm_from_ci = self.parse_download_ci_llvm(None, false);
+        }
+
+        self.llvm_tests = llvm_tests.unwrap_or(false);
+        self.llvm_enzyme = llvm_enzyme.unwrap_or(false);
+        self.llvm_offload = llvm_offload.unwrap_or(false);
+        self.llvm_plugins = llvm_plugins.unwrap_or(false);
+    }
+}

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

@@ -0,0 +1,184 @@
+//! This module defines the structures that directly mirror the `bootstrap.toml`
+//! file's format. These types are used for `serde` deserialization.
+//!
+//! Crucially, this module also houses the core logic for loading, parsing, and merging
+//! these raw TOML configurations from various sources (the main `bootstrap.toml`,
+//! included files, profile defaults, and command-line overrides). This processed
+//! TOML data then serves as an intermediate representation, which is further
+//! transformed and applied to the final [`Config`] struct.
+
+use serde::Deserialize;
+use serde_derive::Deserialize;
+pub mod build;
+pub mod change_id;
+pub mod dist;
+pub mod gcc;
+pub mod install;
+pub mod llvm;
+pub mod rust;
+pub mod target;
+
+use build::Build;
+use change_id::{ChangeId, ChangeIdWrapper};
+use dist::Dist;
+use gcc::Gcc;
+use install::Install;
+use llvm::Llvm;
+use rust::Rust;
+use target::TomlTarget;
+
+use crate::core::config::{Merge, ReplaceOpt};
+use crate::{Config, HashMap, HashSet, Path, PathBuf, exit, fs, t};
+
+/// Structure of the `bootstrap.toml` file that configuration is read from.
+///
+/// This structure uses `Decodable` to automatically decode a TOML configuration
+/// file into this format, and then this is traversed and written into the above
+/// `Config` structure.
+#[derive(Deserialize, Default)]
+#[serde(deny_unknown_fields, rename_all = "kebab-case")]
+pub(crate) struct TomlConfig {
+    #[serde(flatten)]
+    pub(crate) change_id: ChangeIdWrapper,
+    pub(super) build: Option<Build>,
+    pub(super) install: Option<Install>,
+    pub(super) llvm: Option<Llvm>,
+    pub(super) gcc: Option<Gcc>,
+    pub(super) rust: Option<Rust>,
+    pub(super) target: Option<HashMap<String, TomlTarget>>,
+    pub(super) dist: Option<Dist>,
+    pub(super) profile: Option<String>,
+    pub(super) include: Option<Vec<PathBuf>>,
+}
+
+impl Merge for TomlConfig {
+    fn merge(
+        &mut self,
+        parent_config_path: Option<PathBuf>,
+        included_extensions: &mut HashSet<PathBuf>,
+        TomlConfig { build, install, llvm, gcc, rust, dist, target, profile, change_id, include }: Self,
+        replace: ReplaceOpt,
+    ) {
+        fn do_merge<T: Merge>(x: &mut Option<T>, y: Option<T>, replace: ReplaceOpt) {
+            if let Some(new) = y {
+                if let Some(original) = x {
+                    original.merge(None, &mut Default::default(), new, replace);
+                } else {
+                    *x = Some(new);
+                }
+            }
+        }
+
+        self.change_id.inner.merge(None, &mut Default::default(), change_id.inner, replace);
+        self.profile.merge(None, &mut Default::default(), profile, replace);
+
+        do_merge(&mut self.build, build, replace);
+        do_merge(&mut self.install, install, replace);
+        do_merge(&mut self.llvm, llvm, replace);
+        do_merge(&mut self.gcc, gcc, replace);
+        do_merge(&mut self.rust, rust, replace);
+        do_merge(&mut self.dist, dist, replace);
+
+        match (self.target.as_mut(), target) {
+            (_, None) => {}
+            (None, Some(target)) => self.target = Some(target),
+            (Some(original_target), Some(new_target)) => {
+                for (triple, new) in new_target {
+                    if let Some(original) = original_target.get_mut(&triple) {
+                        original.merge(None, &mut Default::default(), new, replace);
+                    } else {
+                        original_target.insert(triple, new);
+                    }
+                }
+            }
+        }
+
+        let parent_dir = parent_config_path
+            .as_ref()
+            .and_then(|p| p.parent().map(ToOwned::to_owned))
+            .unwrap_or_default();
+
+        // `include` handled later since we ignore duplicates using `ReplaceOpt::IgnoreDuplicate` to
+        // keep the upper-level configuration to take precedence.
+        for include_path in include.clone().unwrap_or_default().iter().rev() {
+            let include_path = parent_dir.join(include_path);
+            let include_path = include_path.canonicalize().unwrap_or_else(|e| {
+                eprintln!("ERROR: Failed to canonicalize '{}' path: {e}", include_path.display());
+                exit!(2);
+            });
+
+            let included_toml = Config::get_toml_inner(&include_path).unwrap_or_else(|e| {
+                eprintln!("ERROR: Failed to parse '{}': {e}", include_path.display());
+                exit!(2);
+            });
+
+            assert!(
+                included_extensions.insert(include_path.clone()),
+                "Cyclic inclusion detected: '{}' is being included again before its previous inclusion was fully processed.",
+                include_path.display()
+            );
+
+            self.merge(
+                Some(include_path.clone()),
+                included_extensions,
+                included_toml,
+                // Ensures that parent configuration always takes precedence
+                // over child configurations.
+                ReplaceOpt::IgnoreDuplicate,
+            );
+
+            included_extensions.remove(&include_path);
+        }
+    }
+}
+
+/// This file is embedded in the overlay directory of the tarball sources. It is
+/// useful in scenarios where developers want to see how the tarball sources were
+/// generated.
+///
+/// We also use this file to compare the host's bootstrap.toml against the CI rustc builder
+/// configuration to detect any incompatible options.
+pub const BUILDER_CONFIG_FILENAME: &str = "builder-config";
+
+impl Config {
+    pub(crate) fn get_builder_toml(&self, build_name: &str) -> Result<TomlConfig, toml::de::Error> {
+        if self.dry_run() {
+            return Ok(TomlConfig::default());
+        }
+
+        let builder_config_path =
+            self.out.join(self.build.triple).join(build_name).join(BUILDER_CONFIG_FILENAME);
+        Self::get_toml(&builder_config_path)
+    }
+
+    pub(crate) fn get_toml(file: &Path) -> Result<TomlConfig, toml::de::Error> {
+        #[cfg(test)]
+        return Ok(TomlConfig::default());
+
+        #[cfg(not(test))]
+        Self::get_toml_inner(file)
+    }
+
+    pub(crate) fn get_toml_inner(file: &Path) -> Result<TomlConfig, toml::de::Error> {
+        let contents =
+            t!(fs::read_to_string(file), format!("config file {} not found", file.display()));
+        // Deserialize to Value and then TomlConfig to prevent the Deserialize impl of
+        // TomlConfig and sub types to be monomorphized 5x by toml.
+        toml::from_str(&contents)
+            .and_then(|table: toml::Value| TomlConfig::deserialize(table))
+            .inspect_err(|_| {
+                if let Ok(ChangeIdWrapper { inner: Some(ChangeId::Id(id)) }) =
+                    toml::from_str::<toml::Value>(&contents)
+                        .and_then(|table: toml::Value| ChangeIdWrapper::deserialize(table))
+                {
+                    let changes = crate::find_recent_config_change_ids(id);
+                    if !changes.is_empty() {
+                        println!(
+                            "WARNING: There have been changes to x.py since you last updated:\n{}",
+                            crate::human_readable_changes(changes)
+                        );
+                    }
+                }
+            })
+    }
+}

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

@@ -0,0 +1,174 @@
+//! This module defines the structures and logic for handling target-specific configuration
+//! within the `bootstrap.toml` file. This allows you to customize build settings, tools,
+//! and flags for individual compilation targets.
+//!
+//! It includes:
+//!
+//! * [`TomlTarget`]: This struct directly mirrors the `[target.<triple>]` sections in your
+//!   `bootstrap.toml`. It's used for deserializing raw TOML data for a specific target.
+//! * [`Target`]: This struct represents the processed and validated configuration for a
+//!   build target, which is is stored in the main [`Config`] structure.
+//! * [`Config::apply_target_config`]: This method processes the `TomlTarget` data and
+//!   applies it to the global [`Config`], ensuring proper path resolution, validation,
+//!   and integration with other build settings.
+
+use std::collections::HashMap;
+
+use serde::{Deserialize, Deserializer};
+
+use crate::core::build_steps::compile::CODEGEN_BACKEND_PREFIX;
+use crate::core::config::{LlvmLibunwind, Merge, ReplaceOpt, SplitDebuginfo, StringOrBool};
+use crate::{Config, HashSet, PathBuf, TargetSelection, define_config, exit};
+
+define_config! {
+    /// TOML representation of how each build target is configured.
+    struct TomlTarget {
+        cc: Option<String> = "cc",
+        cxx: Option<String> = "cxx",
+        ar: Option<String> = "ar",
+        ranlib: Option<String> = "ranlib",
+        default_linker: Option<PathBuf> = "default-linker",
+        linker: Option<String> = "linker",
+        split_debuginfo: Option<String> = "split-debuginfo",
+        llvm_config: Option<String> = "llvm-config",
+        llvm_has_rust_patches: Option<bool> = "llvm-has-rust-patches",
+        llvm_filecheck: Option<String> = "llvm-filecheck",
+        llvm_libunwind: Option<String> = "llvm-libunwind",
+        sanitizers: Option<bool> = "sanitizers",
+        profiler: Option<StringOrBool> = "profiler",
+        rpath: Option<bool> = "rpath",
+        crt_static: Option<bool> = "crt-static",
+        musl_root: Option<String> = "musl-root",
+        musl_libdir: Option<String> = "musl-libdir",
+        wasi_root: Option<String> = "wasi-root",
+        qemu_rootfs: Option<String> = "qemu-rootfs",
+        no_std: Option<bool> = "no-std",
+        codegen_backends: Option<Vec<String>> = "codegen-backends",
+        runner: Option<String> = "runner",
+        optimized_compiler_builtins: Option<bool> = "optimized-compiler-builtins",
+        jemalloc: Option<bool> = "jemalloc",
+    }
+}
+
+/// Per-target configuration stored in the global configuration structure.
+#[derive(Debug, Default, Clone, PartialEq, Eq)]
+pub struct Target {
+    /// Some(path to llvm-config) if using an external LLVM.
+    pub llvm_config: Option<PathBuf>,
+    pub llvm_has_rust_patches: Option<bool>,
+    /// Some(path to FileCheck) if one was specified.
+    pub llvm_filecheck: Option<PathBuf>,
+    pub llvm_libunwind: Option<LlvmLibunwind>,
+    pub cc: Option<PathBuf>,
+    pub cxx: Option<PathBuf>,
+    pub ar: Option<PathBuf>,
+    pub ranlib: Option<PathBuf>,
+    pub default_linker: Option<PathBuf>,
+    pub linker: Option<PathBuf>,
+    pub split_debuginfo: Option<SplitDebuginfo>,
+    pub sanitizers: Option<bool>,
+    pub profiler: Option<StringOrBool>,
+    pub rpath: Option<bool>,
+    pub crt_static: Option<bool>,
+    pub musl_root: Option<PathBuf>,
+    pub musl_libdir: Option<PathBuf>,
+    pub wasi_root: Option<PathBuf>,
+    pub qemu_rootfs: Option<PathBuf>,
+    pub runner: Option<String>,
+    pub no_std: bool,
+    pub codegen_backends: Option<Vec<String>>,
+    pub optimized_compiler_builtins: Option<bool>,
+    pub jemalloc: Option<bool>,
+}
+
+impl Target {
+    pub fn from_triple(triple: &str) -> Self {
+        let mut target: Self = Default::default();
+        if triple.contains("-none") || triple.contains("nvptx") || triple.contains("switch") {
+            target.no_std = true;
+        }
+        if triple.contains("emscripten") {
+            target.runner = Some("node".into());
+        }
+        target
+    }
+}
+
+impl Config {
+    pub fn apply_target_config(&mut self, toml_target: Option<HashMap<String, TomlTarget>>) {
+        if let Some(t) = toml_target {
+            for (triple, cfg) in t {
+                let mut target = Target::from_triple(&triple);
+
+                if let Some(ref s) = cfg.llvm_config {
+                    if self.download_rustc_commit.is_some() && triple == *self.build.triple {
+                        panic!(
+                            "setting llvm_config for the host is incompatible with download-rustc"
+                        );
+                    }
+                    target.llvm_config = Some(self.src.join(s));
+                }
+                if let Some(patches) = cfg.llvm_has_rust_patches {
+                    assert!(
+                        self.submodules == Some(false) || cfg.llvm_config.is_some(),
+                        "use of `llvm-has-rust-patches` is restricted to cases where either submodules are disabled or llvm-config been provided"
+                    );
+                    target.llvm_has_rust_patches = Some(patches);
+                }
+                if let Some(ref s) = cfg.llvm_filecheck {
+                    target.llvm_filecheck = Some(self.src.join(s));
+                }
+                target.llvm_libunwind = cfg.llvm_libunwind.as_ref().map(|v| {
+                    v.parse().unwrap_or_else(|_| {
+                        panic!("failed to parse target.{triple}.llvm-libunwind")
+                    })
+                });
+                if let Some(s) = cfg.no_std {
+                    target.no_std = s;
+                }
+                target.cc = cfg.cc.map(PathBuf::from);
+                target.cxx = cfg.cxx.map(PathBuf::from);
+                target.ar = cfg.ar.map(PathBuf::from);
+                target.ranlib = cfg.ranlib.map(PathBuf::from);
+                target.linker = cfg.linker.map(PathBuf::from);
+                target.crt_static = cfg.crt_static;
+                target.musl_root = cfg.musl_root.map(PathBuf::from);
+                target.musl_libdir = cfg.musl_libdir.map(PathBuf::from);
+                target.wasi_root = cfg.wasi_root.map(PathBuf::from);
+                target.qemu_rootfs = cfg.qemu_rootfs.map(PathBuf::from);
+                target.runner = cfg.runner;
+                target.sanitizers = cfg.sanitizers;
+                target.profiler = cfg.profiler;
+                target.rpath = cfg.rpath;
+                target.optimized_compiler_builtins = cfg.optimized_compiler_builtins;
+                target.jemalloc = cfg.jemalloc;
+
+                if let Some(ref backends) = cfg.codegen_backends {
+                    let available_backends = ["llvm", "cranelift", "gcc"];
+
+                    target.codegen_backends = Some(backends.iter().map(|s| {
+                        if let Some(backend) = s.strip_prefix(CODEGEN_BACKEND_PREFIX) {
+                            if available_backends.contains(&backend) {
+                                panic!("Invalid value '{s}' for 'target.{triple}.codegen-backends'. Instead, please use '{backend}'.");
+                            } else {
+                                println!("HELP: '{s}' for 'target.{triple}.codegen-backends' might fail. \
+                                    Codegen backends are mostly defined without the '{CODEGEN_BACKEND_PREFIX}' prefix. \
+                                    In this case, it would be referred to as '{backend}'.");
+                            }
+                        }
+
+                        s.clone()
+                    }).collect());
+                }
+
+                target.split_debuginfo = cfg.split_debuginfo.as_ref().map(|v| {
+                    v.parse().unwrap_or_else(|_| {
+                        panic!("invalid value for target.{triple}.split-debuginfo")
+                    })
+                });
+
+                self.target_config.insert(TargetSelection::from_user(&triple), target);
+            }
+        }
+    }
+}

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

@@ -727,7 +727,7 @@ download-rustc = false
         use build_helper::git::PathFreshness;
 
         use crate::core::build_steps::llvm::detect_llvm_freshness;
-        use crate::core::config::check_incompatible_options_for_ci_llvm;
+        use crate::core::config::toml::llvm::check_incompatible_options_for_ci_llvm;
 
         if !self.llvm_from_ci {
             return;