Auto merge of #86015 - jyn514:revert-revert, r=Mark-Simulacrum

Move LLVM submodule updates back to native.rs

Time to find more bugs!

The first commit is a straight revert of https://github.com/rust-lang/rust/pull/85647, the second is a fix for https://rust-lang.zulipchat.com/#narrow/stream/182449-t-compiler.2Fhelp/topic/x.2Epy.20always.20updates.20LLVM.20submodule/near/240113320 and https://github.com/rust-lang/rust/pull/82653#issuecomment-846755631. I haven't been able to replicate https://github.com/rust-lang/rust/pull/82653#issuecomment-849013698.

3 files changed, 90 insertions, 13 deletions

diff --git a/src/bootstrap/bootstrap.py b/src/bootstrap/bootstrap.py
index 7c7f162b82c..9b122bf4914 100644
--- a/src/bootstrap/bootstrap.py
+++ b/src/bootstrap/bootstrap.py
@@ -993,28 +993,20 @@ class RustBuild(object):
         ).decode(default_encoding).splitlines()]
         filtered_submodules = []
         submodules_names = []
-        llvm_checked_out = os.path.exists(os.path.join(self.rust_root, "src/llvm-project/.git"))
-        external_llvm_provided = self.get_toml('llvm-config') or self.downloading_llvm()
-        llvm_needed = not self.get_toml('codegen-backends', 'rust') \
-            or "llvm" in self.get_toml('codegen-backends', 'rust')
         for module in submodules:
+            # This is handled by native::Llvm in rustbuild, not here
             if module.endswith("llvm-project"):
-                # Don't sync the llvm-project submodule if an external LLVM was
-                # provided, if we are downloading LLVM or if the LLVM backend is
-                # not being built. Also, if the submodule has been initialized
-                # already, sync it anyways so that it doesn't mess up contributor
-                # pull requests.
-                if external_llvm_provided or not llvm_needed:
-                    if self.get_toml('lld') != 'true' and not llvm_checked_out:
-                        continue
+                continue
             check = self.check_submodule(module, slow_submodules)
             filtered_submodules.append((module, check))
             submodules_names.append(module)
         recorded = subprocess.Popen(["git", "ls-tree", "HEAD"] + submodules_names,
                                     cwd=self.rust_root, stdout=subprocess.PIPE)
         recorded = recorded.communicate()[0].decode(default_encoding).strip().splitlines()
+        # { filename: hash }
         recorded_submodules = {}
         for data in recorded:
+            # [mode, kind, hash, filename]
             data = data.split()
             recorded_submodules[data[3]] = data[2]
         for module in filtered_submodules:
diff --git a/src/bootstrap/lib.rs b/src/bootstrap/lib.rs
index 347236c655a..8e8c27b0d0d 100644
--- a/src/bootstrap/lib.rs
+++ b/src/bootstrap/lib.rs
@@ -483,6 +483,12 @@ impl Build {
             job::setup(self);
         }
 
+        // If the LLVM submodule has been initialized already, sync it unconditionally. This avoids
+        // contributors checking in a submodule change by accident.
+        if self.in_tree_llvm_info.is_git() {
+            native::update_llvm_submodule(self);
+        }
+
         if let Subcommand::Format { check, paths } = &self.config.cmd {
             return format::format(self, *check, &paths);
         }
diff --git a/src/bootstrap/native.rs b/src/bootstrap/native.rs
index 449fdb87b02..375f5d727f0 100644
--- a/src/bootstrap/native.rs
+++ b/src/bootstrap/native.rs
@@ -21,7 +21,7 @@ use build_helper::{output, t};
 use crate::builder::{Builder, RunConfig, ShouldRun, Step};
 use crate::config::TargetSelection;
 use crate::util::{self, exe};
-use crate::GitRepo;
+use crate::{Build, GitRepo};
 use build_helper::up_to_date;
 
 pub struct Meta {
@@ -91,6 +91,82 @@ pub fn prebuilt_llvm_config(
     Err(Meta { stamp, build_llvm_config, out_dir, root: root.into() })
 }
 
+// modified from `check_submodule` and `update_submodule` in bootstrap.py
+pub(crate) fn update_llvm_submodule(build: &Build) {
+    let llvm_project = &Path::new("src").join("llvm-project");
+
+    fn dir_is_empty(dir: &Path) -> bool {
+        t!(std::fs::read_dir(dir)).next().is_none()
+    }
+
+    // NOTE: The check for the empty directory is here because when running x.py
+    // the first time, the llvm submodule won't be checked out. Check it out
+    // now so we can build it.
+    if !build.in_tree_llvm_info.is_git() && !dir_is_empty(&build.config.src.join(llvm_project)) {
+        return;
+    }
+
+    // check_submodule
+    if build.config.fast_submodules {
+        let checked_out_hash = output(
+            Command::new("git")
+                .args(&["rev-parse", "HEAD"])
+                .current_dir(build.config.src.join(llvm_project)),
+        );
+        // update_submodules
+        let recorded = output(
+            Command::new("git")
+                .args(&["ls-tree", "HEAD"])
+                .arg(llvm_project)
+                .current_dir(&build.config.src),
+        );
+        let actual_hash = recorded
+            .split_whitespace()
+            .nth(2)
+            .unwrap_or_else(|| panic!("unexpected output `{}`", recorded));
+
+        // update_submodule
+        if actual_hash == checked_out_hash.trim_end() {
+            // already checked out
+            return;
+        }
+    }
+
+    println!("Updating submodule {}", llvm_project.display());
+    build.run(
+        Command::new("git")
+            .args(&["submodule", "-q", "sync"])
+            .arg(llvm_project)
+            .current_dir(&build.config.src),
+    );
+
+    // Try passing `--progress` to start, then run git again without if that fails.
+    let update = |progress: bool| {
+        let mut git = Command::new("git");
+        git.args(&["submodule", "update", "--init", "--recursive"]);
+        if progress {
+            git.arg("--progress");
+        }
+        git.arg(llvm_project).current_dir(&build.config.src);
+        git
+    };
+    // NOTE: doesn't use `try_run` because this shouldn't print an error if it fails.
+    if !update(true).status().map_or(false, |status| status.success()) {
+        build.run(&mut update(false));
+    }
+
+    build.run(
+        Command::new("git")
+            .args(&["reset", "-q", "--hard"])
+            .current_dir(build.config.src.join(llvm_project)),
+    );
+    build.run(
+        Command::new("git")
+            .args(&["clean", "-qdfx"])
+            .current_dir(build.config.src.join(llvm_project)),
+    );
+}
+
 #[derive(Debug, Copy, Clone, Hash, PartialEq, Eq)]
 pub struct Llvm {
     pub target: TargetSelection,
@@ -128,6 +204,9 @@ impl Step for Llvm {
                 Err(m) => m,
             };
 
+        if !builder.config.dry_run {
+            update_llvm_submodule(builder);
+        }
         if builder.config.llvm_link_shared
             && (target.contains("windows") || target.contains("apple-darwin"))
         {