link Enzyme and the offload with in-tree lld if possible by ZuseZ4 · Pull Request #162824 · rust-lang/rust

ZuseZ4 · 2026-09-15T19:57:55Z

I have multiple LLVM builds on all my computers, and often enough ended up with an older lld version on the path, which often breaks Offload builds.
I've also seen autodiff builds take forever since some of it's files are huge, and builds somehow end up using ld.

This fixes both by just using our up-to-date in-tree lld (if available).
I used an LLM to implement a couple of bootstrap fixes, including this one. I then split this one out and cleaned it up.

r? kobzol

closes: #160584

rustbot · 2026-09-15T19:57:58Z

This PR changes how LLVM is built. Consider updating src/bootstrap/download-ci-llvm-stamp.

rustbot · 2026-09-15T19:58:01Z

kobzol is not on the review rotation at the moment.
They may take a while to respond.

ZuseZ4 · 2026-09-15T20:11:52Z

This should be the last bootstrap PR for a while.
I then (might) just come with a really big one in the future to delete most of this ^^
Now that @sgasho's PR is in, we don't rely on the clang-linker-wrapper anymore and have a bit more control.
The OffloadBinary format was only needed when passing artifacts around between clang tooling, which keeps wrapping and unwrapping IR/asm. Since we're in full control now, we can delete a lot of that.

Kobzol · 2026-09-16T10:06:04Z

+    }
+
+    let lld_bin = builder.ensure(Lld { target }).join("bin");
+    ldflags.push_all(format!("-B{} -fuse-ld=lld", lld_bin.display()));


This duplicates the -fuse-ld=lld flag set on line 846, right?

View changes since the review

Yep. I don't think it's harmful, but I moved it to avoid it.

…when it is built

rustbot · 2026-09-17T04:14:13Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

Kobzol · 2026-09-18T07:16:41Z

@bors r+

rust-bors · 2026-09-18T07:16:44Z

📌 Commit 752e7c0 has been approved by Kobzol

It is now in the queue for this repository.

link Enzyme and the offload with in-tree lld if possible I have multiple LLVM builds on all my computers, and often enough ended up with an older lld version on the path, which often breaks Offload builds. I've also seen autodiff builds take forever since some of it's files are huge, and builds somehow end up using ld. This fixes both by just using our up-to-date in-tree lld (if available). I used an LLM to implement a couple of bootstrap fixes, including this one. I then split this one out and cleaned it up. r? kobzol closes: rust-lang#160584

Rollup of 9 pull requests Successful merges: - #161777 (Add Natvis visualiser and debuginfo tests for `f128`) - #162423 (Rename the src install build step to rust-src.) - #162824 (link Enzyme and the offload with in-tree lld if possible) - #161803 (Fix docs of make_ascii_lowercase/make_ascii_upercase) - #162256 (Add mentions to sync back `RELEASES.md` to the `main` branch) - #162803 (docs(num): add documentation for `NonZero::from_str`) - #162906 (Move more `rustdoc-html` tests in the right location) - #162922 (An assortment of polonius tweaks) - #162930 (Use niche length type for strlen to guarantee `isize::MAX` bound)

link Enzyme and the offload with in-tree lld if possible I have multiple LLVM builds on all my computers, and often enough ended up with an older lld version on the path, which often breaks Offload builds. I've also seen autodiff builds take forever since some of it's files are huge, and builds somehow end up using ld. This fixes both by just using our up-to-date in-tree lld (if available). I used an LLM to implement a couple of bootstrap fixes, including this one. I then split this one out and cleaned it up. r? kobzol closes: rust-lang#160584

…uwer Rollup of 15 pull requests Successful merges: - #160401 (sparc: make ABI consistent with clang) - #161777 (Add Natvis visualiser and debuginfo tests for `f128`) - #162423 (Rename the src install build step to rust-src.) - #162740 (stdarch subtree update) - #162824 (link Enzyme and the offload with in-tree lld if possible) - #162832 (add `Div` and `Mul` for `Complex<{float}>`) - #161005 (fix: unfulfilled nested dead code lint) - #161803 (Fix docs of make_ascii_lowercase/make_ascii_upercase) - #162256 (Add mentions to sync back `RELEASES.md` to the `main` branch) - #162661 (simplify `Target::GenericParam`) - #162666 (Tidy footnote in `platform-support.md`) - #162803 (docs(num): add documentation for `NonZero::from_str`) - #162906 (Move more `rustdoc-html` tests in the right location) - #162922 (An assortment of polonius tweaks) - #162930 (Use niche length type for strlen to guarantee `isize::MAX` bound)

link Enzyme and the offload with in-tree lld if possible I have multiple LLVM builds on all my computers, and often enough ended up with an older lld version on the path, which often breaks Offload builds. I've also seen autodiff builds take forever since some of it's files are huge, and builds somehow end up using ld. This fixes both by just using our up-to-date in-tree lld (if available). I used an LLM to implement a couple of bootstrap fixes, including this one. I then split this one out and cleaned it up. r? kobzol closes: rust-lang#160584

…uwer Rollup of 15 pull requests Successful merges: - #160401 (sparc: make ABI consistent with clang) - #161777 (Add Natvis visualiser and debuginfo tests for `f128`) - #162423 (Rename the src install build step to rust-src.) - #162740 (stdarch subtree update) - #162824 (link Enzyme and the offload with in-tree lld if possible) - #161005 (fix: unfulfilled nested dead code lint) - #161803 (Fix docs of make_ascii_lowercase/make_ascii_upercase) - #162256 (Add mentions to sync back `RELEASES.md` to the `main` branch) - #162661 (simplify `Target::GenericParam`) - #162666 (Tidy footnote in `platform-support.md`) - #162803 (docs(num): add documentation for `NonZero::from_str`) - #162906 (Move more `rustdoc-html` tests in the right location) - #162922 (An assortment of polonius tweaks) - #162929 (Update unicode_data to Unicode version 18.0.0) - #162930 (Use niche length type for strlen to guarantee `isize::MAX` bound)

link Enzyme and the offload with in-tree lld if possible I have multiple LLVM builds on all my computers, and often enough ended up with an older lld version on the path, which often breaks Offload builds. I've also seen autodiff builds take forever since some of it's files are huge, and builds somehow end up using ld. This fixes both by just using our up-to-date in-tree lld (if available). I used an LLM to implement a couple of bootstrap fixes, including this one. I then split this one out and cleaned it up. r? kobzol closes: rust-lang#160584

…uwer Rollup of 14 pull requests Successful merges: - #160401 (sparc: make ABI consistent with clang) - #161777 (Add Natvis visualiser and debuginfo tests for `f128`) - #162740 (stdarch subtree update) - #162824 (link Enzyme and the offload with in-tree lld if possible) - #161005 (fix: unfulfilled nested dead code lint) - #161803 (Fix docs of make_ascii_lowercase/make_ascii_upercase) - #162256 (Add mentions to sync back `RELEASES.md` to the `main` branch) - #162661 (simplify `Target::GenericParam`) - #162666 (Tidy footnote in `platform-support.md`) - #162803 (docs(num): add documentation for `NonZero::from_str`) - #162906 (Move more `rustdoc-html` tests in the right location) - #162922 (An assortment of polonius tweaks) - #162929 (Update unicode_data to Unicode version 18.0.0) - #162930 (Use niche length type for strlen to guarantee `isize::MAX` bound)

link Enzyme and the offload with in-tree lld if possible I have multiple LLVM builds on all my computers, and often enough ended up with an older lld version on the path, which often breaks Offload builds. I've also seen autodiff builds take forever since some of it's files are huge, and builds somehow end up using ld. This fixes both by just using our up-to-date in-tree lld (if available). I used an LLM to implement a couple of bootstrap fixes, including this one. I then split this one out and cleaned it up. r? kobzol closes: rust-lang#160584

…uwer Rollup of 20 pull requests Successful merges: - #160401 (sparc: make ABI consistent with clang) - #162715 (don't calculate dtors if the self ty has impossible bounds) - #162740 (stdarch subtree update) - #162824 (link Enzyme and the offload with in-tree lld if possible) - #162946 (Simplify query stack printing) - #161005 (fix: unfulfilled nested dead code lint) - #161246 (Normalize non-rigid aliases in ty_known_to_outlive) - #161803 (Fix docs of make_ascii_lowercase/make_ascii_upercase) - #162256 (Add mentions to sync back `RELEASES.md` to the `main` branch) - #162661 (simplify `Target::GenericParam`) - #162666 (Tidy footnote in `platform-support.md`) - #162803 (docs(num): add documentation for `NonZero::from_str`) - #162879 (use u64 limbs in core::num::bignum) - #162903 (PassWrapper: adapt to LLVM reading exception model from module flag) - #162905 (c-variadic: add checks for windows i686) - #162906 (Move more `rustdoc-html` tests in the right location) - #162922 (An assortment of polonius tweaks) - #162929 (Update unicode_data to Unicode version 18.0.0) - #162930 (Use niche length type for strlen to guarantee `isize::MAX` bound) - #162960 (Guard types with unstable `Allocator` params ahead of partial stabilization)

JonathanBrouwer · 2026-09-18T21:04:18Z

💔 I suspect this PR failed tests as part of a rollup
@bors r-

After fixing the problem, consider running a try job for the failed job before re-approving.

Link to failure: #162983 (comment)

rust-bors · 2026-09-18T21:04:26Z

This pull request was unapproved.

This PR was contained in a rollup (#162983), which was unapproved.

View changes since this unapproval

JonathanBrouwer · 2026-09-18T21:04:44Z

@bors try jobs=dist-x86_64-llvm-mingw

link Enzyme and the offload with in-tree lld if possible try-job: dist-x86_64-llvm-mingw

ZuseZ4 · 2026-09-18T21:17:22Z

Oh, this was aimed at and tested on x86-linux, since that's where offload is currently active, but I guess via autodiff it could have broken mingw, even though I'd find it a bit surprising. I'll have a look, sorry

rust-bors · 2026-09-18T23:17:26Z

💔 Test for f70d840 failed: CI. Failed job:

try - dist-x86_64-llvm-mingw (web logs, enhanced plaintext logs)

rust-log-analyzer · 2026-09-18T23:17:30Z

The job dist-x86_64-llvm-mingw failed! Check out the build log: (web) (plain enhanced) (plain)

Click to see the possible cause of the failure (guessed by this bot)

CMAKE_x86_64-pc-windows-gnullvm = None
CMAKE_x86_64_pc_windows_gnullvm = None
HOST_CMAKE = None
CMAKE = None
running: "cmake" "D:\\a\\rust\\rust\\src/tools/enzyme/enzyme/" "-G" "Ninja" "-DCMAKE_INSTALL_MESSAGE=LAZY" "-DCMAKE_C_COMPILER_LAUNCHER=sccache" "-DCMAKE_CXX_COMPILER_LAUNCHER=sccache" "-DCMAKE_C_COMPILER=clang.exe" "-DCMAKE_CXX_COMPILER=clang++.exe" "-DCMAKE_ASM_COMPILER=clang.exe" "-DCMAKE_C_FLAGS= -Wno-deprecated -ffunction-sections -fdata-sections -m64 --target=x86_64-pc-windows-gnu -w -fdebug-prefix-map=D:\\a\\rust\\rust=/rustc/llvm" "-DCMAKE_CXX_FLAGS= -Wno-deprecated -ffunction-sections -fdata-sections -m64 --target=x86_64-pc-windows-gnu -w -fdebug-prefix-map=D:\\a\\rust\\rust=/rustc/llvm" "-DCMAKE_SHARED_LINKER_FLAGS= -BD:\\a\\rust\\rust\\build\\x86_64-pc-windows-gnullvm\\lld\\bin -fuse-ld=lld -static-libstdc++" "-DCMAKE_MODULE_LINKER_FLAGS= -BD:\\a\\rust\\rust\\build\\x86_64-pc-windows-gnullvm\\lld\\bin -fuse-ld=lld -static-libstdc++" "-DCMAKE_EXE_LINKER_FLAGS= -BD:\\a\\rust\\rust\\build\\x86_64-pc-windows-gnullvm\\lld\\bin -fuse-ld=lld -static-libstdc++" "-DLLVM_ENABLE_ASSERTIONS=ON" "-DENZYME_EXTERNAL_SHARED_LIB=ON" "-DENZYME_BC_LOADER=OFF" "-DLLVM_DIR=D:\\a\\rust\\rust\\build\\x86_64-pc-windows-gnullvm\\llvm\\lib\\cmake\\llvm" "-DCMAKE_INSTALL_PREFIX=D:\\a\\rust\\rust\\build\\x86_64-pc-windows-gnullvm\\enzyme" "-DCMAKE_ASM_FLAGS= -ffunction-sections -fdata-sections -m64 --target=x86_64-pc-windows-gnu -w" "-DCMAKE_BUILD_TYPE=Release"
-- The C compiler identification is Clang 21.1.5
-- The CXX compiler identification is Clang 21.1.5
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - failed
-- Check for working C compiler: D:/a/rust/rust/clang64/bin/clang.exe
---
  is not able to compile a simple test program.

  It fails with the following output:

    Change Dir: 'D:/a/rust/rust/build/x86_64-pc-windows-gnullvm/enzyme/build/CMakeFiles/CMakeScratch/TryCompile-esl7lj'
    
    Run Build Command(s): D:/a/rust/rust/ninja/ninja.exe -v cmTC_548cd
    [1/2] D:\a\rust\rust\clang64\bin\clang.exe   -Wno-deprecated -ffunction-sections -fdata-sections -m64 --target=x86_64-pc-windows-gnu -w -fdebug-prefix-map=D:\a\rust\rust=/rustc/llvm -MD -MT CMakeFiles/cmTC_548cd.dir/testCCompiler.c.obj -MF CMakeFiles\cmTC_548cd.dir\testCCompiler.c.obj.d -o CMakeFiles/cmTC_548cd.dir/testCCompiler.c.obj -c D:/a/rust/rust/build/x86_64-pc-windows-gnullvm/enzyme/build/CMakeFiles/CMakeScratch/TryCompile-esl7lj/testCCompiler.c
    [2/2] C:\Windows\system32\cmd.exe /C "cd . && D:\a\rust\rust\clang64\bin\clang.exe -Wno-deprecated -ffunction-sections -fdata-sections -m64 --target=x86_64-pc-windows-gnu -w -fdebug-prefix-map=D:\a\rust\rust=/rustc/llvm -BD:\a\rust\rust\build\x86_64-pc-windows-gnullvm\lld\bin -fuse-ld=lld -static-libstdc++ CMakeFiles/cmTC_548cd.dir/testCCompiler.c.obj -o cmTC_548cd.exe -Wl,--out-implib,libcmTC_548cd.dll.a -Wl,--major-image-version,0,--minor-image-version,0  -lkernel32 -luser32 -lgdi32 -lwinspool -lshell32 -lole32 -loleaut32 -luuid -lcomdlg32 -ladvapi32 && cd ."
    FAILED: cmTC_548cd.exe 
    C:\Windows\system32\cmd.exe /C "cd . && D:\a\rust\rust\clang64\bin\clang.exe -Wno-deprecated -ffunction-sections -fdata-sections -m64 --target=x86_64-pc-windows-gnu -w -fdebug-prefix-map=D:\a\rust\rust=/rustc/llvm -BD:\a\rust\rust\build\x86_64-pc-windows-gnullvm\lld\bin -fuse-ld=lld -static-libstdc++ CMakeFiles/cmTC_548cd.dir/testCCompiler.c.obj -o cmTC_548cd.exe -Wl,--out-implib,libcmTC_548cd.dll.a -Wl,--major-image-version,0,--minor-image-version,0  -lkernel32 -luser32 -lgdi32 -lwinspool -lshell32 -lole32 -loleaut32 -luuid -lcomdlg32 -ladvapi32 && cd ."
    clang-21: error: linker command failed due to signal (use -v to see invocation)

    ninja: build stopped: subcommand failed.
    
    

---
   0: __rustc::rust_begin_unwind
             at /rustc/cbae9b4cae2b108f6a3d18cfe6075714bb739463/library\std\src/panicking.rs:679:5
   1: core::panicking::panic_fmt
             at /rustc/cbae9b4cae2b108f6a3d18cfe6075714bb739463/library\core\src/panicking.rs:80:14
   2: cmake::fail
   3: cmake::run
   4: <cmake::Config>::build
   5: <bootstrap::core::build_steps::llvm::Enzyme as bootstrap::core::builder::CommandLineStep>::run
             at .\src\bootstrap\src\core\build_steps\llvm.rs:1623:13
   6: <bootstrap::core::build_steps::llvm::Enzyme as bootstrap::core::builder::Step>::run
             at .\src\bootstrap\src\core\builder\mod.rs:134:9
   7: <bootstrap::core::builder::Builder>::ensure::<bootstrap::core::build_steps::llvm::Enzyme>
             at .\src\bootstrap\src\core\builder\mod.rs:1637:36
   8: <bootstrap::core::build_steps::dist::Enzyme as bootstrap::core::builder::CommandLineStep>::run
             at .\src\bootstrap\src\core\build_steps\dist.rs:2837:30
   9: <bootstrap::core::build_steps::dist::Enzyme as bootstrap::core::builder::Step>::run
             at .\src\bootstrap\src\core\builder\mod.rs:134:9
  10: <bootstrap::core::builder::Builder>::ensure::<bootstrap::core::build_steps::dist::Enzyme>
             at .\src\bootstrap\src\core\builder\mod.rs:1637:36
  11: <bootstrap::core::build_steps::dist::Enzyme as bootstrap::core::builder::CommandLineStep>::make_run
             at .\src\bootstrap\src\core\build_steps\dist.rs:2824:21
  12: <bootstrap::core::builder::CommandLineStepDescription>::maybe_run
             at .\src\bootstrap\src\core\builder\mod.rs:470:13
  13: bootstrap::core::builder::cli_paths::match_paths_to_steps_and_run
             at .\src\bootstrap\src\core\builder\cli_paths.rs:138:19
  14: <bootstrap::core::builder::Builder>::run_step_descriptions
             at .\src\bootstrap\src\core\builder\mod.rs:1173:9
  15: <bootstrap::core::builder::Builder>::execute_cli

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-bootstrap Relevant to the bootstrap subteam: Rust's build system (x.py and src/bootstrap) labels Sep 15, 2026

rustbot assigned Kobzol Sep 15, 2026

ZuseZ4 mentioned this pull request Sep 15, 2026

Tracking Issue for GPU-offload #131513

Open

5 tasks

Kobzol reviewed Sep 16, 2026

View reviewed changes

This comment has been minimized.

Sign in to view

ZuseZ4 force-pushed the use-lld-more branch from 3e659af to 7f7e355 Compare September 17, 2026 03:59

bootstrap: link Enzyme and the offload runtimes with the in-tree lld …

752e7c0

…when it is built

ZuseZ4 force-pushed the use-lld-more branch from 7f7e355 to 752e7c0 Compare September 17, 2026 04:14

Kobzol approved these changes Sep 18, 2026

View reviewed changes

rust-bors Bot added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Sep 18, 2026

Zalathar mentioned this pull request Sep 18, 2026

Rollup of 9 pull requests #162941

Closed

JonathanBrouwer mentioned this pull request Sep 18, 2026

Rollup of 11 pull requests #162943

Closed

JonathanBrouwer mentioned this pull request Sep 18, 2026

Rollup of 15 pull requests #162951

Closed

JonathanBrouwer mentioned this pull request Sep 18, 2026

Rollup of 15 pull requests #162954

Closed

JonathanBrouwer mentioned this pull request Sep 18, 2026

Rollup of 14 pull requests #162973

Closed

JonathanBrouwer mentioned this pull request Sep 18, 2026

Rollup of 20 pull requests #162980

Closed

JonathanBrouwer mentioned this pull request Sep 18, 2026

Rollup of 20 pull requests #162983

Closed

rust-bors Bot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Sep 18, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Sep 18, 2026

Auto merge of #162824 - ZuseZ4:use-lld-more, r=<try>

f70d840

link Enzyme and the offload with in-tree lld if possible try-job: dist-x86_64-llvm-mingw

Uh oh!

Conversation

ZuseZ4 commented Sep 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Sep 15, 2026

Uh oh!

rustbot commented Sep 15, 2026

Uh oh!

ZuseZ4 commented Sep 15, 2026

Uh oh!

Uh oh!

Kobzol Sep 16, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ZuseZ4 Sep 17, 2026

Choose a reason for hiding this comment

Uh oh!

This comment has been minimized.

This comment has been minimized.

rustbot commented Sep 17, 2026

Uh oh!

Kobzol commented Sep 18, 2026

Uh oh!

rust-bors Bot commented Sep 18, 2026

Uh oh!

JonathanBrouwer commented Sep 18, 2026

Uh oh!

rust-bors Bot commented Sep 18, 2026

Uh oh!

JonathanBrouwer commented Sep 18, 2026

Uh oh!

This comment has been minimized.

ZuseZ4 commented Sep 18, 2026

Uh oh!

rust-bors Bot commented Sep 18, 2026

Uh oh!

rust-log-analyzer commented Sep 18, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

ZuseZ4 commented Sep 15, 2026 •

edited

Loading

Kobzol Sep 16, 2026 •

edited by rustbot

Loading