rust: Increase QEMU size to 1024 MB
PATCH v5 1/3 rust: Increase QEMU size to 1024 MB
From: Deepesh Varatharajan <[email protected]>
Rightnow rust depends on llvm instead of rust-llvm
Setup in rust-llvm
CFLAGS:remove = "-g"
CXXFLAGS:remove = "-g"
Setup in llvmDEBUG_LEVELFLAG = "-g1"
As a result, the stage1 compiler crate binaries include debug symbols, increasing their size. These binaries are used to run tests inside QEMU. To accommodate this, increase the QEMU RAM allocation to 1024 MB.
--- meta/lib/oeqa/selftest/cases/rust.py | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
meta/lib/oeqa/selftest/cases/rust.py +3 -2
@@ -102,8 +102,9 @@ class RustSelfTestSystemEmulated(OESelftestTestCase, OEPTestResultTestCase): testargs = exclude_fail_tests + " --no-doc --no-fail-fast --bless" # wrap the execution with a qemu instance. - # Tests are run with 512 tasks in parallel to execute all tests very quickly - with runqemu("core-image-minimal", runqemuparams = "nographic", qemuparams = "-m 512") as qemu: + # Set QEMU RAM to 1024MB to support running unit tests for the compiler crate, including larger + # test cases and parallel execution in the test environment. + with runqemu("core-image-minimal", runqemuparams = "nographic", qemuparams = "-m 1024") as qemu: # Copy remote-test-server to image through scp host_sys = get_bb_var("RUST_BUILD_SYS", "rust") ssh = SSHControl(ip=qemu.ip, logfile=qemu.sshlog, user="root")
Signature · 1 line
2.49.0
PATCH v5 2/3 rust: Use llvm instead of rust-llvm
From: Deepesh Varatharajan <[email protected]>
Updated the Rust build to depend on llvm instead.
*Summary of discussion with the rust upstream about using latest LLVM instead of Rust maintained LLVM fork.
https://internals.rust-lang.org/t/can-we-use-proper-clang-instead-of-llvm-fork-what-rust-uses/23489
*Upstream LLVM is generally compatible:
- Rust does support building with upstream (vanilla) LLVM, especially the latest major release and the one or two preceding ones.
https://rustc-dev-guide.rust-lang.org/backend/updating-llvm.html#updating-llvm
*Impact on Yocto Rust upgrades:
- Rust upgrades shall always check for updates on rust forked llvm and backport the relevant patches to llvm.
*Regarding the rust forked llvm local patches:
- There are no local patches on rust forked llvm other than the backported fixes from llvm master.
*We are copying the natively built llvm-config binary into the target sysroot and running it. However, this llvm-config has compile time dependencies on various other arch's LLVM libraries because native-llvm is built for all oe-core supported targets.
Attempting to work around this by symlinking the missing libraries from the native sysroot into the target sysroot leads to mixed architectures in the final .rlib. Specifically, the object files extracted from those symlinked libraries within librustc_llvm-<hash>.rlib are built for the host, while others are correctly built for the target This results in linker failures due to file format not recognized.
To resolve this, we now build llvm-target also for all oe-core supported architectures in addition to the native-llvm build. This ensures that llvm-config and all associated libraries are built for the correct target, eliminating cross-architecture contamination and linker issues.
*We are enabling -DLLVM_INSTALL_UTILS=ON to ensure essential LLVM utilities like FileCheck are available, as they are required by the Rust build.
Without this, the build fails with an error as below:
| thread 'main' panicked at src/bootstrap/src/core/sanity.rs:315:21:
| FileCheck executable "poky/build/tmp/work/x86_64-linux/rust-native/1.90.0/recipe-sysroot -native/usr/bin/FileCheck" does not exist
*We now add these flags "-Clink-arg=-lz -Clink-arg=-lzstd" because of this following diff otherwise we will get errors during link time.
Setup in rust-llvm
-DLLVM_ENABLE_ZLIB=OFF \
-DLLVM_ENABLE_ZSTD=OFF \
-DLLVM_ENABLE_FFI=OFF \
Setup in llvm
-DLLVM_ENABLE_FFI=ON \
*When multilibs enabled:
llvm-config expects static libraries to be located in the lib directory rather than lib64. However, since we are copying the natively built llvm-config to target sysroot and running it and llvm-config doesn't know anything about lib64 existence. To accommodate this without breaking multilib behavior, we are creating a symlink from 'lib' to 'lib64' directory.
Previously, when we depended on rust-llvm, this worked because we specified:
-DCMAKE_INSTALL_PREFIX:PATH=${libdir}/llvm-rust
With this setup, llvm-config was installed inside ${libdir}/llvm-rust, which included its own bin and lib directories. Thus, llvm-config located in bin would correctly find the libraries in the adjacent lib directory.
Even when multilib was enabled or not, llvm-config would still look for libraries under lib in this structure, so everything functioned as expected.
--- meta/recipes-devtools/clang/llvm_git.bb | 2 +- meta/recipes-devtools/rust/rust_1.90.0.bb | 39 ++++++++++++++++++++--- 2 files changed, 36 insertions(+), 5 deletions(-)
meta/recipes-devtools/clang/llvm_git.bb +1 -1
@@ -27,7 +27,6 @@ OECMAKE_SOURCEPATH = "${S}/llvm" # https://github.com/llvm/llvm-project/blob/main/llvm/CMakeLists.txt LLVM_TARGETS_GPU ?= "${@bb.utils.contains_any('DISTRO_FEATURES', 'opencl opengl vulkan', 'AMDGPU;NVPTX;SPIRV', '', d)}" LLVM_TARGETS_TO_BUILD ?= "AArch64;ARM;BPF;Mips;PowerPC;RISCV;X86;LoongArch;${LLVM_TARGETS_GPU}" -LLVM_TARGETS_TO_BUILD:class-target ?= "${@get_clang_host_arch(bb, d)};BPF;${LLVM_TARGETS_GPU}" LLVM_EXPERIMENTAL_TARGETS_TO_BUILD ?= ""
@@ -37,6 +36,7 @@ HF[vardepvalue] = "${HF}" EXTRA_OECMAKE += "-DCMAKE_BUILD_TYPE=MinSizeRel \ -DLLVM_ENABLE_BINDINGS=OFF \ + -DLLVM_INSTALL_UTILS=ON \ -DLLVM_ENABLE_FFI=ON \ -DLLVM_ENABLE_RTTI=ON \ -DLLVM_TARGETS_TO_BUILD='${LLVM_TARGETS_TO_BUILD}' \
meta/recipes-devtools/rust/rust_1.90.0.bb +35 -4
@@ -7,7 +7,7 @@ LIC_FILES_CHKSUM = "file://COPYRIGHT;md5=11a3899825f4376896e438c8c753f8dc" inherit rust inherit cargo_common -DEPENDS += "rust-llvm" +DEPENDS += "llvm" # native rust uses cargo/rustc from binary snapshots to bootstrap # but everything else should use our native builds DEPENDS:append:class-target = " cargo-native rust-native"
@@ -28,8 +28,8 @@ PV .= "${@bb.utils.contains('RUST_CHANNEL', 'stable', '', '-${RUST_CHANNEL}', d) export FORCE_CRATE_HASH = "${BB_TASKHASH}" -RUST_ALTERNATE_EXE_PATH ?= "${STAGING_LIBDIR}/llvm-rust/bin/llvm-config" -RUST_ALTERNATE_EXE_PATH_NATIVE = "${STAGING_LIBDIR_NATIVE}/llvm-rust/bin/llvm-config" +RUST_ALTERNATE_EXE_PATH ?= "${STAGING_BINDIR}/llvm-config" +RUST_ALTERNATE_EXE_PATH_NATIVE = "${STAGING_BINDIR_NATIVE}/llvm-config" # We don't want to use bitbakes vendoring because the rust sources do their # own vendoring.
@@ -188,6 +188,37 @@ python do_configure() { bb.build.exec_func("setup_cargo_environment", d) } +# llvm-config expects static libraries to be in the 'lib' directory rather than 'lib64' when +# multilibs enabled. Since we are copying the natively built llvm-config into the target sysroot +# and executing it there, it will default to searching in 'lib', as it is unaware of the 'lib64' +# directory. To ensure llvm-config can locate the necessary libraries, create a symlink from 'lib' +do_compile:append:class-target() { + # Ensure llvm-config can find static libraries in multilib setup + lib64_dir="${STAGING_DIR_TARGET}/usr/lib64" + lib_dir="${STAGING_DIR_TARGET}/usr/lib" + + if [ -d "$lib64_dir" ]; then + # If lib does not exist, symlink it to lib64 + if [ ! -e "$lib_dir" ]; then + ln -s lib64 "$lib_dir" + fi + + # Only do per-file symlinking if lib is a real directory (not symlink) + if [ -d "$lib_dir" ] && [ ! -L "$lib_dir" ]; then + for lib64_file in "${lib64_dir}"/libLLVM*.a; do + if [ -e "$lib64_file" ]; then + lib_name=$(basename "${lib64_file}") + target_link="${lib_dir}/${lib_name}" + + if [ ! -e "${target_link}" ]; then + ln -s "../lib64/${lib_name}" "${target_link}" + fi + fi + done + fi + fi +} + rust_runx () { echo "COMPILE ${PN}" "$@"
@@ -199,7 +230,7 @@ rust_runx () { unset CXXFLAGS unset CPPFLAGS - export RUSTFLAGS="${RUST_DEBUG_REMAP}" + export RUSTFLAGS="${RUST_DEBUG_REMAP} -Clink-arg=-lz -Clink-arg=-lzstd" # Copy the natively built llvm-config into the target so we can run it. Horrible, # but works!
Signature · 1 line
2.49.0
Re: [OE-core] [PATCH V5 2/3] rust: Use llvm instead of rust-llvm
I am seeing a build regression when using clang as default compiler
https://errors.yoctoproject.org/Errors/Details/885430/
PREFERRED_TOOLCHAIN_TARGET ?= "clang"PREFERRED_TOOLCHAIN_SDK ?= "clang"
TOOLCHAIN_NATIVE:pn-rust ?= "clang"
TOOLCHAIN_NATIVE:pn-rust-native ?= "clang"
TOOLCHAIN_NATIVE:pn-nativesdk-rust ?= "clang"
Quoted 182 lines · Deepesh via lists.openembedded.org
From: Deepesh Varatharajan <[email protected]>
Updated the Rust build to depend on llvm instead.
*Summary of discussion with the rust upstream about using latest LLVM instead of Rust maintained LLVM fork.
https://internals.rust-lang.org/t/can-we-use-proper-clang-instead-of-llvm-fork-what-rust-uses/23489*Upstream LLVM is generally compatible:
- Rust does support building with upstream (vanilla) LLVM, especially the latest major release and the one or two preceding ones.
https://rustc-dev-guide.rust-lang.org/backend/updating-llvm.html#updating-llvm*Impact on Yocto Rust upgrades:
- Rust upgrades shall always check for updates on rust forked llvm and backport the relevant patches to llvm.*Regarding the rust forked llvm local patches:
- There are no local patches on rust forked llvm other than the backported fixes from llvm master.*We are copying the natively built
llvm-configbinary into the target sysroot and running it. However, thisllvm-confighas compile time dependencies on various other arch's LLVM libraries because native-llvm is built for all oe-core supported targets.Attempting to work around this by symlinking the missing libraries from the native sysroot into the target sysroot leads to mixed architectures in the final
.rlib. Specifically, the object files extracted from those symlinked libraries withinlibrustc_llvm-<hash>.rlibare built for the host, while others are correctly built for the target This results in linker failures due to file format not recognized.To resolve this, we now build llvm-target also for all oe-core supported architectures in addition to the native-llvm build. This ensures that
llvm-configand all associated libraries are built for the correct target, eliminating cross-architecture contamination and linker issues.*We are enabling -DLLVM_INSTALL_UTILS=ON to ensure essential LLVM utilities like FileCheck are available, as they are required by the Rust build.
Without this, the build fails with an error as below:
| thread 'main' panicked atsrc/bootstrap/src/core/sanity.rs:315:21:
| FileCheck executable "poky/build/tmp/work/x86_64-linux/rust-native/1.90.0/recipe-sysroot -native/usr/bin/FileCheck" does not exist*We now add these flags "-Clink-arg=-lz -Clink-arg=-lzstd" because of this following diff otherwise we will get errors during link time.
Setup in rust-llvm
(quoted diff) +0
-3-DLLVM_ENABLE_ZLIB=OFF \ -DLLVM_ENABLE_ZSTD=OFF \ -DLLVM_ENABLE_FFI=OFF \Setup in llvm
-DLLVM_ENABLE_FFI=ON \*When multilibs enabled:
llvm-config expects static libraries to be located in the lib directory rather than lib64. However, since we are copying the natively built llvm-config to target sysroot and running it and llvm-config doesn't know anything about lib64 existence. To accommodate this without breaking multilib behavior, we are creating a symlink from 'lib' to 'lib64' directory.
Previously, when we depended on rust-llvm, this worked because we specified:
-DCMAKE_INSTALL_PREFIX:PATH=${libdir}/llvm-rustWith this setup, llvm-config was installed inside ${libdir}/llvm-rust, which included its own bin and lib directories. Thus, llvm-config located in bin would correctly find the libraries in the adjacent lib directory.
Even when multilib was enabled or not, llvm-config would still look for libraries under lib in this structure, so everything functioned as expected.
Signed-off-by Deepesh Varatharajan <[email protected]>--- meta/recipes-devtools/clang/llvm_git.bb | 2 +- meta/recipes-devtools/rust/rust_1.90.0.bb | 39 ++++++++++++++++++++--- 2 files changed, 36 insertions(+), 5 deletions(-)meta/recipes-devtools/clang/llvm_git.bb +1
-1@@ -27,7 +27,6 @@ OECMAKE_SOURCEPATH = "${S}/llvm" # https://github.com/llvm/llvm-project/blob/main/llvm/CMakeLists.txt LLVM_TARGETS_GPU ?= "${@bb.utils.contains_any('DISTRO_FEATURES', 'opencl opengl vulkan', 'AMDGPU;NVPTX;SPIRV', '', d)}" LLVM_TARGETS_TO_BUILD ?= "AArch64;ARM;BPF;Mips;PowerPC;RISCV;X86;LoongArch;${LLVM_TARGETS_GPU}" -LLVM_TARGETS_TO_BUILD:class-target ?= "${@get_clang_host_arch(bb, d)};BPF;${LLVM_TARGETS_GPU}" LLVM_EXPERIMENTAL_TARGETS_TO_BUILD ?= ""@@ -37,6 +36,7 @@ HF[vardepvalue] = "${HF}" EXTRA_OECMAKE += "-DCMAKE_BUILD_TYPE=MinSizeRel \ -DLLVM_ENABLE_BINDINGS=OFF \ + -DLLVM_INSTALL_UTILS=ON \ -DLLVM_ENABLE_FFI=ON \ -DLLVM_ENABLE_RTTI=ON \ -DLLVM_TARGETS_TO_BUILD='${LLVM_TARGETS_TO_BUILD}' \meta/recipes-devtools/rust/rust_1.90.0.bb +35
-5@@ -7,7 +7,7 @@ LIC_FILES_CHKSUM = "file://COPYRIGHT;md5=11a3899825f4376896e438c8c753f8dc" inherit rust inherit cargo_common -DEPENDS += "rust-llvm" +DEPENDS += "llvm" # native rust uses cargo/rustc from binary snapshots to bootstrap # but everything else should use our native builds DEPENDS:append:class-target = " cargo-native rust-native"@@ -28,8 +28,8 @@ PV .= "${@bb.utils.contains('RUST_CHANNEL', 'stable', '', '-${RUST_CHANNEL}', d) export FORCE_CRATE_HASH = "${BB_TASKHASH}" -RUST_ALTERNATE_EXE_PATH ?= "${STAGING_LIBDIR}/llvm-rust/bin/llvm-config" -RUST_ALTERNATE_EXE_PATH_NATIVE = "${STAGING_LIBDIR_NATIVE}/llvm-rust/bin/llvm-config" +RUST_ALTERNATE_EXE_PATH ?= "${STAGING_BINDIR}/llvm-config" +RUST_ALTERNATE_EXE_PATH_NATIVE = "${STAGING_BINDIR_NATIVE}/llvm-config" # We don't want to use bitbakes vendoring because the rust sources do their # own vendoring.@@ -188,6 +188,37 @@ python do_configure() { bb.build.exec_func("setup_cargo_environment", d) } +# llvm-config expects static libraries to be in the 'lib' directory rather than 'lib64' when +# multilibs enabled. Since we are copying the natively built llvm-config into the target sysroot +# and executing it there, it will default to searching in 'lib', as it is unaware of the 'lib64' +# directory. To ensure llvm-config can locate the necessary libraries, create a symlink from 'lib' +do_compile:append:class-target() { + # Ensure llvm-config can find static libraries in multilib setup + lib64_dir="${STAGING_DIR_TARGET}/usr/lib64" + lib_dir="${STAGING_DIR_TARGET}/usr/lib" + + if [ -d "$lib64_dir" ]; then + # If lib does not exist, symlink it to lib64 + if [ ! -e "$lib_dir" ]; then + ln -s lib64 "$lib_dir" + fi + + # Only do per-file symlinking if lib is a real directory (not symlink) + if [ -d "$lib_dir" ] && [ ! -L "$lib_dir" ]; then + for lib64_file in "${lib64_dir}"/libLLVM*.a; do + if [ -e "$lib64_file" ]; then + lib_name=$(basename "${lib64_file}") + target_link="${lib_dir}/${lib_name}" + + if [ ! -e "${target_link}" ]; then + ln -s "../lib64/${lib_name}" "${target_link}" + fi + fi + done + fi + fi +} + rust_runx () { echo "COMPILE ${PN}" "$@"@@ -199,7 +230,7 @@ rust_runx () { unset CXXFLAGS unset CPPFLAGS - export RUSTFLAGS="${RUST_DEBUG_REMAP}" + export RUSTFLAGS="${RUST_DEBUG_REMAP} -Clink-arg=-lz -Clink-arg=-lzstd" # Copy the natively built llvm-config into the target so we can run it. Horrible, # but works! --2.49.0
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#224888): https://lists.openembedded.org/g/openembedded-core/message/224888
Mute This Topic: https://lists.openembedded.org/mt/115768136/1997914
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
Re: [OE-core] [PATCH V5 2/3] rust: Use llvm instead of rust-llvm
Quoted 5 lines · Khem Raj
CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know the content is safe.I am seeing a build regression when using clang as default compiler
Quoted 7 lines
https://errors.yoctoproject.org/Errors/Details/885430/
PREFERRED_TOOLCHAIN_TARGET?= "clang"PREFERRED_TOOLCHAIN_SDK?= "clang"
TOOLCHAIN_NATIVE:pn-rust ?= "clang"
TOOLCHAIN_NATIVE:pn-rust-native ?= "clang"
TOOLCHAIN_NATIVE:pn-nativesdk-rust ?= "clang"
Hi Khem,
WIth the above mentioned conf in local.conf, I tried building rust for qemuarm and qemuarm64 quite few times and the build completed successfully without any issues for both native and target. Do I need to add any other layers on top of poky ? Do you have any branch where I can clone and reproduce the issue ? Am I missing any step in reproducing the issue ?
Regards,
Deepesh
Quoted 185 lines
Quoted 181 lines · Deepesh via lists.openembedded.org
From: Deepesh Varatharajan <[email protected]>
Updated the Rust build to depend on llvm instead.
*Summary of discussion with the rust upstream about using latest LLVM instead of Rust maintained LLVM fork.
https://internals.rust-lang.org/t/can-we-use-proper-clang-instead-of-llvm-fork-what-rust-uses/23489*Upstream LLVM is generally compatible:
- Rust does support building with upstream (vanilla) LLVM, especially the latest major release and the one or two preceding ones.
https://rustc-dev-guide.rust-lang.org/backend/updating-llvm.html#updating-llvm*Impact on Yocto Rust upgrades:
- Rust upgrades shall always check for updates on rust forked llvm and backport the relevant patches to llvm.*Regarding the rust forked llvm local patches:
- There are no local patches on rust forked llvm other than the backported fixes from llvm master.*We are copying the natively built
llvm-configbinary into the target sysroot and running it. However, thisllvm-confighas compile time dependencies on various other arch's LLVM libraries because native-llvm is built for all oe-core supported targets.Attempting to work around this by symlinking the missing libraries from the native sysroot into the target sysroot leads to mixed architectures in the final
.rlib. Specifically, the object files extracted from those symlinked libraries withinlibrustc_llvm-<hash>.rlibare built for the host, while others are correctly built for the target This results in linker failures due to file format not recognized.To resolve this, we now build llvm-target also for all oe-core supported architectures in addition to the native-llvm build. This ensures that
llvm-configand all associated libraries are built for the correct target, eliminating cross-architecture contamination and linker issues.*We are enabling -DLLVM_INSTALL_UTILS=ON to ensure essential LLVM utilities like FileCheck are available, as they are required by the Rust build.
Without this, the build fails with an error as below:
| thread 'main' panicked atsrc/bootstrap/src/core/sanity.rs:315:21:
| FileCheck executable "poky/build/tmp/work/x86_64-linux/rust-native/1.90.0/recipe-sysroot -native/usr/bin/FileCheck" does not exist*We now add these flags "-Clink-arg=-lz -Clink-arg=-lzstd" because of this following diff otherwise we will get errors during link time.
Setup in rust-llvm
(quoted diff) +0
-3-DLLVM_ENABLE_ZLIB=OFF \ -DLLVM_ENABLE_ZSTD=OFF \ -DLLVM_ENABLE_FFI=OFF \Setup in llvm
-DLLVM_ENABLE_FFI=ON \*When multilibs enabled:
llvm-config expects static libraries to be located in the lib directory rather than lib64. However, since we are copying the natively built llvm-config to target sysroot and running it and llvm-config doesn't know anything about lib64 existence. To accommodate this without breaking multilib behavior, we are creating a symlink from 'lib' to 'lib64' directory.
Previously, when we depended on rust-llvm, this worked because we specified:
-DCMAKE_INSTALL_PREFIX:PATH=${libdir}/llvm-rustWith this setup, llvm-config was installed inside ${libdir}/llvm-rust, which included its own bin and lib directories. Thus, llvm-config located in bin would correctly find the libraries in the adjacent lib directory.
Even when multilib was enabled or not, llvm-config would still look for libraries under lib in this structure, so everything functioned as expected.
Signed-off-by Deepesh Varatharajan <[email protected]>--- meta/recipes-devtools/clang/llvm_git.bb | 2 +- meta/recipes-devtools/rust/rust_1.90.0.bb | 39 ++++++++++++++++++++--- 2 files changed, 36 insertions(+), 5 deletions(-)meta/recipes-devtools/clang/llvm_git.bb +1
-1@@ -27,7 +27,6 @@ OECMAKE_SOURCEPATH = "${S}/llvm" # https://github.com/llvm/llvm-project/blob/main/llvm/CMakeLists.txt LLVM_TARGETS_GPU ?= "${@bb.utils.contains_any('DISTRO_FEATURES', 'opencl opengl vulkan', 'AMDGPU;NVPTX;SPIRV', '', d)}" LLVM_TARGETS_TO_BUILD ?= "AArch64;ARM;BPF;Mips;PowerPC;RISCV;X86;LoongArch;${LLVM_TARGETS_GPU}" -LLVM_TARGETS_TO_BUILD:class-target ?= "${@get_clang_host_arch(bb, d)};BPF;${LLVM_TARGETS_GPU}" LLVM_EXPERIMENTAL_TARGETS_TO_BUILD ?= ""@@ -37,6 +36,7 @@ HF[vardepvalue] = "${HF}" EXTRA_OECMAKE += "-DCMAKE_BUILD_TYPE=MinSizeRel \ -DLLVM_ENABLE_BINDINGS=OFF \ + -DLLVM_INSTALL_UTILS=ON \ -DLLVM_ENABLE_FFI=ON \ -DLLVM_ENABLE_RTTI=ON \ -DLLVM_TARGETS_TO_BUILD='${LLVM_TARGETS_TO_BUILD}' \meta/recipes-devtools/rust/rust_1.90.0.bb +35
-5@@ -7,7 +7,7 @@ LIC_FILES_CHKSUM = "file://COPYRIGHT;md5=11a3899825f4376896e438c8c753f8dc" inherit rust inherit cargo_common -DEPENDS += "rust-llvm" +DEPENDS += "llvm" # native rust uses cargo/rustc from binary snapshots to bootstrap # but everything else should use our native builds DEPENDS:append:class-target = " cargo-native rust-native"@@ -28,8 +28,8 @@ PV .= "${@bb.utils.contains('RUST_CHANNEL', 'stable', '', '-${RUST_CHANNEL}', d) export FORCE_CRATE_HASH = "${BB_TASKHASH}" -RUST_ALTERNATE_EXE_PATH ?= "${STAGING_LIBDIR}/llvm-rust/bin/llvm-config" -RUST_ALTERNATE_EXE_PATH_NATIVE = "${STAGING_LIBDIR_NATIVE}/llvm-rust/bin/llvm-config" +RUST_ALTERNATE_EXE_PATH ?= "${STAGING_BINDIR}/llvm-config" +RUST_ALTERNATE_EXE_PATH_NATIVE = "${STAGING_BINDIR_NATIVE}/llvm-config" # We don't want to use bitbakes vendoring because the rust sources do their # own vendoring.@@ -188,6 +188,37 @@ python do_configure() { bb.build.exec_func("setup_cargo_environment", d) } +# llvm-config expects static libraries to be in the 'lib' directory rather than 'lib64' when +# multilibs enabled. Since we are copying the natively built llvm-config into the target sysroot +# and executing it there, it will default to searching in 'lib', as it is unaware of the 'lib64' +# directory. To ensure llvm-config can locate the necessary libraries, create a symlink from 'lib' +do_compile:append:class-target() { + # Ensure llvm-config can find static libraries in multilib setup + lib64_dir="${STAGING_DIR_TARGET}/usr/lib64" + lib_dir="${STAGING_DIR_TARGET}/usr/lib" + + if [ -d "$lib64_dir" ]; then + # If lib does not exist, symlink it to lib64 + if [ ! -e "$lib_dir" ]; then + ln -s lib64 "$lib_dir" + fi + + # Only do per-file symlinking if lib is a real directory (not symlink) + if [ -d "$lib_dir" ] && [ ! -L "$lib_dir" ]; then + for lib64_file in "${lib64_dir}"/libLLVM*.a; do + if [ -e "$lib64_file" ]; then + lib_name=$(basename "${lib64_file}") + target_link="${lib_dir}/${lib_name}" + + if [ ! -e "${target_link}" ]; then + ln -s "../lib64/${lib_name}" "${target_link}" + fi + fi + done + fi + fi +} + rust_runx () { echo "COMPILE ${PN}" "$@"@@ -199,7 +230,7 @@ rust_runx () { unset CXXFLAGS unset CPPFLAGS - export RUSTFLAGS="${RUST_DEBUG_REMAP}" + export RUSTFLAGS="${RUST_DEBUG_REMAP} -Clink-arg=-lz -Clink-arg=-lzstd" # Copy the natively built llvm-config into the target so we can run it. Horrible, # but works! --2.49.0
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#224888): https://lists.openembedded.org/g/openembedded-core/message/224888
Mute This Topic: https://lists.openembedded.org/mt/115768136/1997914
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
Re: [OE-core] [PATCH V5 2/3] rust: Use llvm instead of rust-llvm
Quoted 17 lines · Deepesh Varatharajan
Quoted 5 lines · Khem Raj
CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know the content is safe.I am seeing a build regression when using clang as default compiler
Quoted 7 lines
https://errors.yoctoproject.org/Errors/Details/885430/
PREFERRED_TOOLCHAIN_TARGET?= "clang"PREFERRED_TOOLCHAIN_SDK?= "clang"
TOOLCHAIN_NATIVE:pn-rust ?= "clang"
TOOLCHAIN_NATIVE:pn-rust-native ?= "clang"
TOOLCHAIN_NATIVE:pn-nativesdk-rust ?= "clang"Hi Khem,
Quoted 8 lines
WIth the above mentioned conf in local.conf, I tried building rust for qemuarm and qemuarm64 quite
few times and the build completed successfully without any issues for both native and target. Do I need
to add any other layers on top of poky ? Do you have any branch where I can clone and reproduce the
issue ? Am I missing any step in reproducing the issue ?
Yeah, I missed a setting which I think could be the reason
TC_CXX_RUNTIME = "llvm"
Quoted 187 lines
Regards,
DeepeshQuoted 185 lines
Quoted 181 lines · Deepesh via lists.openembedded.org
From: Deepesh Varatharajan <[email protected]>
Updated the Rust build to depend on llvm instead.
*Summary of discussion with the rust upstream about using latest LLVM instead of Rust maintained LLVM fork.
https://internals.rust-lang.org/t/can-we-use-proper-clang-instead-of-llvm-fork-what-rust-uses/23489*Upstream LLVM is generally compatible:
- Rust does support building with upstream (vanilla) LLVM, especially the latest major release and the one or two preceding ones.
https://rustc-dev-guide.rust-lang.org/backend/updating-llvm.html#updating-llvm*Impact on Yocto Rust upgrades:
- Rust upgrades shall always check for updates on rust forked llvm and backport the relevant patches to llvm.*Regarding the rust forked llvm local patches:
- There are no local patches on rust forked llvm other than the backported fixes from llvm master.*We are copying the natively built
llvm-configbinary into the target sysroot and running it. However, thisllvm-confighas compile time dependencies on various other arch's LLVM libraries because native-llvm is built for all oe-core supported targets.Attempting to work around this by symlinking the missing libraries from the native sysroot into the target sysroot leads to mixed architectures in the final
.rlib. Specifically, the object files extracted from those symlinked libraries withinlibrustc_llvm-<hash>.rlibare built for the host, while others are correctly built for the target This results in linker failures due to file format not recognized.To resolve this, we now build llvm-target also for all oe-core supported architectures in addition to the native-llvm build. This ensures that
llvm-configand all associated libraries are built for the correct target, eliminating cross-architecture contamination and linker issues.*We are enabling -DLLVM_INSTALL_UTILS=ON to ensure essential LLVM utilities like FileCheck are available, as they are required by the Rust build.
Without this, the build fails with an error as below:
| thread 'main' panicked atsrc/bootstrap/src/core/sanity.rs:315:21:
| FileCheck executable "poky/build/tmp/work/x86_64-linux/rust-native/1.90.0/recipe-sysroot -native/usr/bin/FileCheck" does not exist*We now add these flags "-Clink-arg=-lz -Clink-arg=-lzstd" because of this following diff otherwise we will get errors during link time.
Setup in rust-llvm
(quoted diff) +0
-3-DLLVM_ENABLE_ZLIB=OFF \ -DLLVM_ENABLE_ZSTD=OFF \ -DLLVM_ENABLE_FFI=OFF \Setup in llvm
-DLLVM_ENABLE_FFI=ON \*When multilibs enabled:
llvm-config expects static libraries to be located in the lib directory rather than lib64. However, since we are copying the natively built llvm-config to target sysroot and running it and llvm-config doesn't know anything about lib64 existence. To accommodate this without breaking multilib behavior, we are creating a symlink from 'lib' to 'lib64' directory.
Previously, when we depended on rust-llvm, this worked because we specified:
-DCMAKE_INSTALL_PREFIX:PATH=${libdir}/llvm-rustWith this setup, llvm-config was installed inside ${libdir}/llvm-rust, which included its own bin and lib directories. Thus, llvm-config located in bin would correctly find the libraries in the adjacent lib directory.
Even when multilib was enabled or not, llvm-config would still look for libraries under lib in this structure, so everything functioned as expected.
Signed-off-by Deepesh Varatharajan <[email protected]>--- meta/recipes-devtools/clang/llvm_git.bb | 2 +- meta/recipes-devtools/rust/rust_1.90.0.bb | 39 ++++++++++++++++++++--- 2 files changed, 36 insertions(+), 5 deletions(-)meta/recipes-devtools/clang/llvm_git.bb +1
-1@@ -27,7 +27,6 @@ OECMAKE_SOURCEPATH = "${S}/llvm" # https://github.com/llvm/llvm-project/blob/main/llvm/CMakeLists.txt LLVM_TARGETS_GPU ?= "${@bb.utils.contains_any('DISTRO_FEATURES', 'opencl opengl vulkan', 'AMDGPU;NVPTX;SPIRV', '', d)}" LLVM_TARGETS_TO_BUILD ?= "AArch64;ARM;BPF;Mips;PowerPC;RISCV;X86;LoongArch;${LLVM_TARGETS_GPU}" -LLVM_TARGETS_TO_BUILD:class-target ?= "${@get_clang_host_arch(bb, d)};BPF;${LLVM_TARGETS_GPU}" LLVM_EXPERIMENTAL_TARGETS_TO_BUILD ?= ""@@ -37,6 +36,7 @@ HF[vardepvalue] = "${HF}" EXTRA_OECMAKE += "-DCMAKE_BUILD_TYPE=MinSizeRel \ -DLLVM_ENABLE_BINDINGS=OFF \ + -DLLVM_INSTALL_UTILS=ON \ -DLLVM_ENABLE_FFI=ON \ -DLLVM_ENABLE_RTTI=ON \ -DLLVM_TARGETS_TO_BUILD='${LLVM_TARGETS_TO_BUILD}' \meta/recipes-devtools/rust/rust_1.90.0.bb +35
-5@@ -7,7 +7,7 @@ LIC_FILES_CHKSUM = "file://COPYRIGHT;md5=11a3899825f4376896e438c8c753f8dc" inherit rust inherit cargo_common -DEPENDS += "rust-llvm" +DEPENDS += "llvm" # native rust uses cargo/rustc from binary snapshots to bootstrap # but everything else should use our native builds DEPENDS:append:class-target = " cargo-native rust-native"@@ -28,8 +28,8 @@ PV .= "${@bb.utils.contains('RUST_CHANNEL', 'stable', '', '-${RUST_CHANNEL}', d) export FORCE_CRATE_HASH = "${BB_TASKHASH}" -RUST_ALTERNATE_EXE_PATH ?= "${STAGING_LIBDIR}/llvm-rust/bin/llvm-config" -RUST_ALTERNATE_EXE_PATH_NATIVE = "${STAGING_LIBDIR_NATIVE}/llvm-rust/bin/llvm-config" +RUST_ALTERNATE_EXE_PATH ?= "${STAGING_BINDIR}/llvm-config" +RUST_ALTERNATE_EXE_PATH_NATIVE = "${STAGING_BINDIR_NATIVE}/llvm-config" # We don't want to use bitbakes vendoring because the rust sources do their # own vendoring.@@ -188,6 +188,37 @@ python do_configure() { bb.build.exec_func("setup_cargo_environment", d) } +# llvm-config expects static libraries to be in the 'lib' directory rather than 'lib64' when +# multilibs enabled. Since we are copying the natively built llvm-config into the target sysroot +# and executing it there, it will default to searching in 'lib', as it is unaware of the 'lib64' +# directory. To ensure llvm-config can locate the necessary libraries, create a symlink from 'lib' +do_compile:append:class-target() { + # Ensure llvm-config can find static libraries in multilib setup + lib64_dir="${STAGING_DIR_TARGET}/usr/lib64" + lib_dir="${STAGING_DIR_TARGET}/usr/lib" + + if [ -d "$lib64_dir" ]; then + # If lib does not exist, symlink it to lib64 + if [ ! -e "$lib_dir" ]; then + ln -s lib64 "$lib_dir" + fi + + # Only do per-file symlinking if lib is a real directory (not symlink) + if [ -d "$lib_dir" ] && [ ! -L "$lib_dir" ]; then + for lib64_file in "${lib64_dir}"/libLLVM*.a; do + if [ -e "$lib64_file" ]; then + lib_name=$(basename "${lib64_file}") + target_link="${lib_dir}/${lib_name}" + + if [ ! -e "${target_link}" ]; then + ln -s "../lib64/${lib_name}" "${target_link}" + fi + fi + done + fi + fi +} + rust_runx () { echo "COMPILE ${PN}" "$@"@@ -199,7 +230,7 @@ rust_runx () { unset CXXFLAGS unset CPPFLAGS - export RUSTFLAGS="${RUST_DEBUG_REMAP}" + export RUSTFLAGS="${RUST_DEBUG_REMAP} -Clink-arg=-lz -Clink-arg=-lzstd" # Copy the natively built llvm-config into the target so we can run it. Horrible, # but works! --2.49.0
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#224888): https://lists.openembedded.org/g/openembedded-core/message/224888
Mute This Topic: https://lists.openembedded.org/mt/115768136/1997914
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
Re: [OE-core] [PATCH V5 2/3] rust: Use llvm instead of rust-llvm
Quoted 29 lines · Khem Raj
CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know the content is safe.Quoted 16 lines · Deepesh Varatharajan
Quoted 5 lines · Khem Raj
CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know the content is safe.I am seeing a build regression when using clang as default compiler
Quoted 7 lines
https://errors.yoctoproject.org/Errors/Details/885430/
PREFERRED_TOOLCHAIN_TARGET?= "clang"PREFERRED_TOOLCHAIN_SDK?= "clang"
TOOLCHAIN_NATIVE:pn-rust ?= "clang"
TOOLCHAIN_NATIVE:pn-rust-native ?= "clang"
TOOLCHAIN_NATIVE:pn-nativesdk-rust ?= "clang"Hi Khem,
Quoted 8 lines
WIth the above mentioned conf in local.conf, I tried building rust for qemuarm and qemuarm64 quite
few times and the build completed successfully without any issues for both native and target. Do I need
to add any other layers on top of poky ? Do you have any branch where I can clone and reproduce the
issue ? Am I missing any step in reproducing the issue ?
Quoted 3 lines
Yeah, I missed a setting which I think could be the reason
TC_CXX_RUNTIME= "llvm"
Hi Khem,
After adding the above in locl.conf, reproduced the issue. I've looked into the issue and here’s what I found:
Currently, both rust-native and rust-target rely on the natively built llvm-config binary. However, upon inspecting the runtime dependencies of llvm-config, I noticed the following:
$ ldd ./llvm-config linux-vdso.so.1 => (0x00007ffdec9eb000) libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6
As shown, llvm-config is linked against libstdc++. Meanwhile, we're enforcing Rust to be built using libc++, and this mismatch is causing symbol resolution issues at runtime.
To address the ABI mismatch, I tried building llvm-native with libc++ by introducing a dependency on libcxx-native. However, this leads to a circular dependency:
>llvm-native depends on libcxx-native
>libcxx-native depends on clang-native
>clang-native depends back on llvm-native
This cycle blocks the build.
One possible solution could be to create a separate llvm-config recipe that builds with a dependency on libcxx-native, and have Rust depend on that version specifically.
Do we have any alternative ways to break this circular dependency?
Regards,
Deepesh
Quoted 188 lines
Quoted 188 lines
Regards,
DeepeshQuoted 184 lines
Quoted 181 lines · Deepesh via lists.openembedded.org
From: Deepesh Varatharajan <[email protected]>
Updated the Rust build to depend on llvm instead.
*Summary of discussion with the rust upstream about using latest LLVM instead of Rust maintained LLVM fork.
https://internals.rust-lang.org/t/can-we-use-proper-clang-instead-of-llvm-fork-what-rust-uses/23489*Upstream LLVM is generally compatible:
- Rust does support building with upstream (vanilla) LLVM, especially the latest major release and the one or two preceding ones.
https://rustc-dev-guide.rust-lang.org/backend/updating-llvm.html#updating-llvm*Impact on Yocto Rust upgrades:
- Rust upgrades shall always check for updates on rust forked llvm and backport the relevant patches to llvm.*Regarding the rust forked llvm local patches:
- There are no local patches on rust forked llvm other than the backported fixes from llvm master.*We are copying the natively built
llvm-configbinary into the target sysroot and running it. However, thisllvm-confighas compile time dependencies on various other arch's LLVM libraries because native-llvm is built for all oe-core supported targets.Attempting to work around this by symlinking the missing libraries from the native sysroot into the target sysroot leads to mixed architectures in the final
.rlib. Specifically, the object files extracted from those symlinked libraries withinlibrustc_llvm-<hash>.rlibare built for the host, while others are correctly built for the target This results in linker failures due to file format not recognized.To resolve this, we now build llvm-target also for all oe-core supported architectures in addition to the native-llvm build. This ensures that
llvm-configand all associated libraries are built for the correct target, eliminating cross-architecture contamination and linker issues.*We are enabling -DLLVM_INSTALL_UTILS=ON to ensure essential LLVM utilities like FileCheck are available, as they are required by the Rust build.
Without this, the build fails with an error as below:
| thread 'main' panicked atsrc/bootstrap/src/core/sanity.rs:315:21:
| FileCheck executable "poky/build/tmp/work/x86_64-linux/rust-native/1.90.0/recipe-sysroot -native/usr/bin/FileCheck" does not exist*We now add these flags "-Clink-arg=-lz -Clink-arg=-lzstd" because of this following diff otherwise we will get errors during link time.
Setup in rust-llvm
(quoted diff) +0
-3-DLLVM_ENABLE_ZLIB=OFF \ -DLLVM_ENABLE_ZSTD=OFF \ -DLLVM_ENABLE_FFI=OFF \Setup in llvm
-DLLVM_ENABLE_FFI=ON \*When multilibs enabled:
llvm-config expects static libraries to be located in the lib directory rather than lib64. However, since we are copying the natively built llvm-config to target sysroot and running it and llvm-config doesn't know anything about lib64 existence. To accommodate this without breaking multilib behavior, we are creating a symlink from 'lib' to 'lib64' directory.
Previously, when we depended on rust-llvm, this worked because we specified:
-DCMAKE_INSTALL_PREFIX:PATH=${libdir}/llvm-rustWith this setup, llvm-config was installed inside ${libdir}/llvm-rust, which included its own bin and lib directories. Thus, llvm-config located in bin would correctly find the libraries in the adjacent lib directory.
Even when multilib was enabled or not, llvm-config would still look for libraries under lib in this structure, so everything functioned as expected.
Signed-off-by Deepesh Varatharajan <[email protected]>--- meta/recipes-devtools/clang/llvm_git.bb | 2 +- meta/recipes-devtools/rust/rust_1.90.0.bb | 39 ++++++++++++++++++++--- 2 files changed, 36 insertions(+), 5 deletions(-)meta/recipes-devtools/clang/llvm_git.bb +1
-1@@ -27,7 +27,6 @@ OECMAKE_SOURCEPATH = "${S}/llvm" # https://github.com/llvm/llvm-project/blob/main/llvm/CMakeLists.txt LLVM_TARGETS_GPU ?= "${@bb.utils.contains_any('DISTRO_FEATURES', 'opencl opengl vulkan', 'AMDGPU;NVPTX;SPIRV', '', d)}" LLVM_TARGETS_TO_BUILD ?= "AArch64;ARM;BPF;Mips;PowerPC;RISCV;X86;LoongArch;${LLVM_TARGETS_GPU}" -LLVM_TARGETS_TO_BUILD:class-target ?= "${@get_clang_host_arch(bb, d)};BPF;${LLVM_TARGETS_GPU}" LLVM_EXPERIMENTAL_TARGETS_TO_BUILD ?= ""@@ -37,6 +36,7 @@ HF[vardepvalue] = "${HF}" EXTRA_OECMAKE += "-DCMAKE_BUILD_TYPE=MinSizeRel \ -DLLVM_ENABLE_BINDINGS=OFF \ + -DLLVM_INSTALL_UTILS=ON \ -DLLVM_ENABLE_FFI=ON \ -DLLVM_ENABLE_RTTI=ON \ -DLLVM_TARGETS_TO_BUILD='${LLVM_TARGETS_TO_BUILD}' \meta/recipes-devtools/rust/rust_1.90.0.bb +35
-5@@ -7,7 +7,7 @@ LIC_FILES_CHKSUM = "file://COPYRIGHT;md5=11a3899825f4376896e438c8c753f8dc" inherit rust inherit cargo_common -DEPENDS += "rust-llvm" +DEPENDS += "llvm" # native rust uses cargo/rustc from binary snapshots to bootstrap # but everything else should use our native builds DEPENDS:append:class-target = " cargo-native rust-native"@@ -28,8 +28,8 @@ PV .= "${@bb.utils.contains('RUST_CHANNEL', 'stable', '', '-${RUST_CHANNEL}', d) export FORCE_CRATE_HASH = "${BB_TASKHASH}" -RUST_ALTERNATE_EXE_PATH ?= "${STAGING_LIBDIR}/llvm-rust/bin/llvm-config" -RUST_ALTERNATE_EXE_PATH_NATIVE = "${STAGING_LIBDIR_NATIVE}/llvm-rust/bin/llvm-config" +RUST_ALTERNATE_EXE_PATH ?= "${STAGING_BINDIR}/llvm-config" +RUST_ALTERNATE_EXE_PATH_NATIVE = "${STAGING_BINDIR_NATIVE}/llvm-config" # We don't want to use bitbakes vendoring because the rust sources do their # own vendoring.@@ -188,6 +188,37 @@ python do_configure() { bb.build.exec_func("setup_cargo_environment", d) } +# llvm-config expects static libraries to be in the 'lib' directory rather than 'lib64' when +# multilibs enabled. Since we are copying the natively built llvm-config into the target sysroot +# and executing it there, it will default to searching in 'lib', as it is unaware of the 'lib64' +# directory. To ensure llvm-config can locate the necessary libraries, create a symlink from 'lib' +do_compile:append:class-target() { + # Ensure llvm-config can find static libraries in multilib setup + lib64_dir="${STAGING_DIR_TARGET}/usr/lib64" + lib_dir="${STAGING_DIR_TARGET}/usr/lib" + + if [ -d "$lib64_dir" ]; then + # If lib does not exist, symlink it to lib64 + if [ ! -e "$lib_dir" ]; then + ln -s lib64 "$lib_dir" + fi + + # Only do per-file symlinking if lib is a real directory (not symlink) + if [ -d "$lib_dir" ] && [ ! -L "$lib_dir" ]; then + for lib64_file in "${lib64_dir}"/libLLVM*.a; do + if [ -e "$lib64_file" ]; then + lib_name=$(basename "${lib64_file}") + target_link="${lib_dir}/${lib_name}" + + if [ ! -e "${target_link}" ]; then + ln -s "../lib64/${lib_name}" "${target_link}" + fi + fi + done + fi + fi +} + rust_runx () { echo "COMPILE ${PN}" "$@"@@ -199,7 +230,7 @@ rust_runx () { unset CXXFLAGS unset CPPFLAGS - export RUSTFLAGS="${RUST_DEBUG_REMAP}" + export RUSTFLAGS="${RUST_DEBUG_REMAP} -Clink-arg=-lz -Clink-arg=-lzstd" # Copy the natively built llvm-config into the target so we can run it. Horrible, # but works! --2.49.0
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#224888): https://lists.openembedded.org/g/openembedded-core/message/224888
Mute This Topic: https://lists.openembedded.org/mt/115768136/1997914
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
Re: [OE-core] [PATCH V5 2/3] rust: Use llvm instead of rust-llvm
In reply to [email protected]Hi,
Quoted 2 lines · Deepesh via lists.openembedded.org
Updated the Rust build to depend on llvm instead.
Because I was curious, I looked at the build stats before and after this series being merged. Here are the diffs for cputime and walltime when I built llvm-native/rust-llvm-native and rust-native:
PKG TASK ABSDIFF RELDIFF CPUTIME1 -> CPUTIME2 - rust-llvm-native do_compile -12241.8s -100.0% 12241.8s -> 0.0s - rust-llvm-native do_unpack -31.4s -100.0% 31.4s -> 0.0s - rust-llvm-native do_configure -9.4s -100.0% 9.4s -> 0.0s - rust-llvm-native do_populate_sysroot -6.2s -100.0% 6.2s -> 0.0s + llvm-project-source-21.1.3 do_patch 5.8s +inf% 0.0s -> 5.8s rust-native do_populate_sysroot 9.5s +323.0% 2.9s -> 12.5s + llvm-native do_configure 12.9s +inf% 0.0s -> 12.9s + llvm-native do_populate_sysroot 18.6s +inf% 0.0s -> 18.6s + llvm-project-source-21.1.3 do_unpack 20.0s +inf% 0.0s -> 20.0s + llvm-tblgen-native do_configure 29.0s +inf% 0.0s -> 29.0s + llvm-tblgen-native do_compile 593.1s +inf% 0.0s -> 593.1s rust-native do_install 1724.2s +26.1% 6611.3s -> 8335.5s + llvm-native do_compile 12051.6s +inf% 0.0s -> 12051.6s
PKG TASK ABSDIFF RELDIFF WALLTIME1 -> WALLTIME2 - rust-llvm-native do_compile -401.1s -100.0% 401.1s -> 0.0s - rust-llvm-native do_unpack -23.0s -100.0% 23.0s -> 0.0s rust-native do_unpack -21.3s -42.5% 50.3s -> 28.9s - rust-llvm-native do_configure -9.0s -100.0% 9.0s -> 0.0s + llvm-project-source-21.1.3 do_unpack 7.0s +inf% 0.0s -> 7.0s + llvm-native do_configure 12.3s +inf% 0.0s -> 12.3s + llvm-tblgen-native do_configure 28.5s +inf% 0.0s -> 28.5s + llvm-tblgen-native do_compile 30.1s +inf% 0.0s -> 30.1s rust-native do_install 33.4s +6.0% 553.4s -> 586.8s + llvm-native do_compile 378.0s +inf% 0.0s -> 378.0s
For llvm vs rust-llvm, I’ve a local tweak to not build some of the GPU backends (as they take a long time) so it’s actually comparing like-for-like, and it’s good to see that the compile time is close. It’s slightly faster in llvm even though it’s building more backends, so I’m glad my build tweaking was useful.
However, rust-native takes 26% more cpu time (just 6% more wall time, due to parallelisation). I’m wondering if this is fallout from using -Os instead of -O2, or something else about how we’re building rust-llvm vs llvm. Insights welcome, but I’ll likely fire a test build without -Os to see how that impacts the numbers later.
Cheers,
Ross
Re: [OE-core] [PATCH V5 2/3] rust: Use llvm instead of rust-llvm
Quoted 32 lines · Ross Burton via lists.openembedded.org
Hi,
Quoted 2 lines · Deepesh via lists.openembedded.org
Updated the Rust build to depend on llvm instead.
Because I was curious, I looked at the build stats before and after this series being merged. Here are the diffs for cputime and walltime when I built llvm-native/rust-llvm-native and rust-native:
PKG TASK ABSDIFF RELDIFF CPUTIME1 -> CPUTIME2
- rust-llvm-native do_compile -12241.8s -100.0% 12241.8s -> 0.0s - rust-llvm-native do_unpack -31.4s -100.0% 31.4s -> 0.0s - rust-llvm-native do_configure -9.4s -100.0% 9.4s -> 0.0s - rust-llvm-native do_populate_sysroot -6.2s -100.0% 6.2s -> 0.0s + llvm-project-source-21.1.3 do_patch 5.8s +inf% 0.0s -> 5.8s rust-native do_populate_sysroot 9.5s +323.0% 2.9s -> 12.5s + llvm-native do_configure 12.9s +inf% 0.0s -> 12.9s + llvm-native do_populate_sysroot 18.6s +inf% 0.0s -> 18.6s + llvm-project-source-21.1.3 do_unpack 20.0s +inf% 0.0s -> 20.0s + llvm-tblgen-native do_configure 29.0s +inf% 0.0s -> 29.0s+ llvm-tblgen-native
do_compile593.1s +inf% 0.0s -> 593.1s
rust-nativedo_install1724.2s +26.1% 6611.3s -> 8335.5s
+ llvm-nativedo_compile12051.6s +inf% 0.0s -> 12051.6sPKG TASK ABSDIFF RELDIFF WALLTIME1 -> WALLTIME2
- rust-llvm-native do_compile -401.1s -100.0% 401.1s -> 0.0s - rust-llvm-native do_unpack -23.0s -100.0% 23.0s -> 0.0s rust-native do_unpack -21.3s -42.5% 50.3s -> 28.9s - rust-llvm-native do_configure -9.0s -100.0% 9.0s -> 0.0s + llvm-project-source-21.1.3 do_unpack 7.0s +inf% 0.0s -> 7.0s + llvm-native do_configure 12.3s +inf% 0.0s -> 12.3s + llvm-tblgen-native do_configure 28.5s +inf% 0.0s -> 28.5s + llvm-tblgen-native do_compile 30.1s +inf% 0.0s -> 30.1s
Quoted 6 lines
rust-native
do_install33.4s +6.0% 553.4s -> 586.8s
+ llvm-nativedo_compile378.0s +inf% 0.0s -> 378.0sFor llvm vs rust-llvm, I’ve a local tweak to not build some of the GPU backends (as they take a long time) so it’s actually comparing like-for-like, and it’s good to see that the compile time is close. It’s slightly faster in llvm even though it’s building more backends, so I’m glad my build tweaking was useful.
However, rust-native takes 26% more cpu time (just 6% more wall time, due to parallelisation). I’m wondering if this is fallout from using -Os instead of -O2, or something else about how we’re building rust-llvm vs llvm. Insights welcome, but I’ll likely fire a test build without -Os to see how that impacts the numbers later.
Followup, comparing current master to master but with LLVM built as RelWithDebInfo instead of MinSizeRel:
PKG TASK ABSDIFF RELDIFF CPUTIME1 -> CPUTIME2 rust-native do_install -1690.4s -20.3% 8335.5s -> 6645.1s llvm-native do_compile 6081.7s +50.5% 12051.6s -> 18133.3s
Cumulative cputime:
4117.4s +19.5% 5:52:36.1 (21156.1s) -> 7:01:13.5 (25273.5s)
PKG TASK ABSDIFF RELDIFF WALLTIME1 -> WALLTIME2 rust-native do_install 24.0s +4.1% 586.8s -> 610.8s llvm-native do_compile 238.9s +63.2% 378.0s -> 616.8s
Cumulative walltime:
251.3s +22.8% 18:24.7 (1104.7s) -> 22:36.0 (1356.0s)
So rust builds a bit faster, but LLVM is a lot slower to build, with a net result that the build is actually slower overall. I’m happy with the choice to use MinSizeRel in LLVM.
Ross
Re: [OE-core] [PATCH V5 2/3] rust: Use llvm instead of rust-llvm
Quoted 51 lines · Ross Burton via lists.openembedded.org
Quoted 32 lines · Ross Burton via lists.openembedded.org
Hi,
Quoted 2 lines · Deepesh via lists.openembedded.org
Updated the Rust build to depend on llvm instead.
Because I was curious, I looked at the build stats before and after this series being merged. Here are the diffs for cputime and walltime when I built llvm-native/rust-llvm-native and rust-native:
PKG TASK ABSDIFF RELDIFF CPUTIME1 -> CPUTIME2
- rust-llvm-native do_compile -12241.8s -100.0% 12241.8s -> 0.0s - rust-llvm-native do_unpack -31.4s -100.0% 31.4s -> 0.0s - rust-llvm-native do_configure -9.4s -100.0% 9.4s -> 0.0s - rust-llvm-native do_populate_sysroot -6.2s -100.0% 6.2s -> 0.0s + llvm-project-source-21.1.3 do_patch 5.8s +inf% 0.0s -> 5.8s rust-native do_populate_sysroot 9.5s +323.0% 2.9s -> 12.5s + llvm-native do_configure 12.9s +inf% 0.0s -> 12.9s + llvm-native do_populate_sysroot 18.6s +inf% 0.0s -> 18.6s + llvm-project-source-21.1.3 do_unpack 20.0s +inf% 0.0s -> 20.0s + llvm-tblgen-native do_configure 29.0s +inf% 0.0s -> 29.0s+ llvm-tblgen-native
do_compile593.1s +inf% 0.0s -> 593.1s
rust-nativedo_install1724.2s +26.1% 6611.3s -> 8335.5s
+ llvm-nativedo_compile12051.6s +inf% 0.0s -> 12051.6sPKG TASK ABSDIFF RELDIFF WALLTIME1 -> WALLTIME2
- rust-llvm-native do_compile -401.1s -100.0% 401.1s -> 0.0s - rust-llvm-native do_unpack -23.0s -100.0% 23.0s -> 0.0s rust-native do_unpack -21.3s -42.5% 50.3s -> 28.9s - rust-llvm-native do_configure -9.0s -100.0% 9.0s -> 0.0s + llvm-project-source-21.1.3 do_unpack 7.0s +inf% 0.0s -> 7.0s + llvm-native do_configure 12.3s +inf% 0.0s -> 12.3s + llvm-tblgen-native do_configure 28.5s +inf% 0.0s -> 28.5s + llvm-tblgen-native do_compile 30.1s +inf% 0.0s -> 30.1sQuoted 6 lines
rust-native
do_install33.4s +6.0% 553.4s -> 586.8s
+ llvm-nativedo_compile378.0s +inf% 0.0s -> 378.0sFor llvm vs rust-llvm, I’ve a local tweak to not build some of the GPU backends (as they take a long time) so it’s actually comparing like-for-like, and it’s good to see that the compile time is close. It’s slightly faster in llvm even though it’s building more backends, so I’m glad my build tweaking was useful.
However, rust-native takes 26% more cpu time (just 6% more wall time, due to parallelisation). I’m wondering if this is fallout from using -Os instead of -O2, or something else about how we’re building rust-llvm vs llvm. Insights welcome, but I’ll likely fire a test build without -Os to see how that impacts the numbers later.
Followup, comparing current master to master but with LLVM built as RelWithDebInfo instead of MinSizeRel:
PKG TASK ABSDIFF RELDIFF CPUTIME1 -> CPUTIME2 rust-native do_install -1690.4s -20.3% 8335.5s -> 6645.1s llvm-native do_compile 6081.7s +50.5% 12051.6s -> 18133.3sCumulative cputime:
4117.4s +19.5% 5:52:36.1 (21156.1s) -> 7:01:13.5 (25273.5s)
Quoted 8 lines
PKG TASK ABSDIFF RELDIFF WALLTIME1 -> WALLTIME2 rust-native do_install 24.0s +4.1% 586.8s -> 610.8s llvm-native do_compile 238.9s +63.2% 378.0s -> 616.8sCumulative walltime:
251.3s +22.8% 18:24.7 (1104.7s) -> 22:36.0 (1356.0s)So rust builds a bit faster, but LLVM is a lot slower to build, with a net result that the build is actually slower overall. I’m happy with the choice to use MinSizeRel in LLVM.
A large set of users of OE would not build compiler as often as some system build engineers who build distros for them will do, so I wonder how fast is resulting compiler in compiling applications, because in such
settings it will be used in often and save a lot of time for users such that spending extra time in compiling it in OE can be justified. Of couse I am not recommending that
OE adopts that option as default.
Quoted 10 lines
Ross
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#225121): https://lists.openembedded.org/g/openembedded-core/message/225121
Mute This Topic: https://lists.openembedded.org/mt/115768136/1997914
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
PATCH v5 3/3 rust: Drop rust-llvm
In reply to [email protected]From: Deepesh Varatharajan <[email protected]>
Dropping rust-llvm because now rust depends on llvm.
0001-AsmMatcherEmitter-sort-ClassInfo-lists-by-name-as-we.patch 0002-llvm-allow-env-override-of-exe-path.patch
(We have the above two patches in llvm)
--- meta/conf/distro/include/maintainers.inc | 1 - .../distro/include/rust_security_flags.inc | 1 - meta/conf/distro/include/tcmode-default.inc | 2 - ...r-sort-ClassInfo-lists-by-name-as-we.patch | 31 ------ ...-llvm-allow-env-override-of-exe-path.patch | 33 ------- .../recipes-devtools/rust/rust-llvm_1.90.0.bb | 98 ------------------- 6 files changed, 166 deletions(-) delete mode 100644 meta/recipes-devtools/rust/rust-llvm/0001-AsmMatcherEmitter-sort-ClassInfo-lists-by-name-as-we.patch delete mode 100644 meta/recipes-devtools/rust/rust-llvm/0002-llvm-allow-env-override-of-exe-path.patch delete mode 100644 meta/recipes-devtools/rust/rust-llvm_1.90.0.bb
meta/conf/distro/include/maintainers.inc +0 -1
@@ -774,7 +774,6 @@ RECIPE_MAINTAINER:pn-ruby = "Ross Burton <[email protected]>" RECIPE_MAINTAINER:pn-run-postinsts = "Ross Burton <[email protected]>" RECIPE_MAINTAINER:pn-rust = "Randy MacLeod <[email protected]>" RECIPE_MAINTAINER:pn-rust-cross-canadian-${TRANSLATED_TARGET_ARCH} = "Randy MacLeod <[email protected]>" -RECIPE_MAINTAINER:pn-rust-llvm = "Randy MacLeod <[email protected]>" RECIPE_MAINTAINER:pn-rxvt-unicode = "Unassigned <[email protected]>" RECIPE_MAINTAINER:pn-sassc = "Simone Weiß <[email protected]>" RECIPE_MAINTAINER:pn-sato-screenshot = "Ross Burton <[email protected]>"
meta/conf/distro/include/rust_security_flags.inc +0 -1
@@ -2,6 +2,5 @@ SECURITY_CFLAGS:pn-rust-native = "${SECURITY_NO_PIE_CFLAGS}" SECURITY_CFLAGS:pn-rust-cross-${TARGET_ARCH} = "${SECURITY_NO_PIE_CFLAGS}" SECURITY_CFLAGS:pn-rust = "${SECURITY_NO_PIE_CFLAGS}" -SECURITY_CFLAGS:pn-rust-llvm = "${SECURITY_NO_PIE_CFLAGS}" SECURITY_LDFLAGS:pn-rust-cross-arm = " -lssp_nonshared -lssp"
meta/conf/distro/include/tcmode-default.inc +0 -2
@@ -43,7 +43,5 @@ PREFERRED_VERSION_cargo-native ?= "${RUSTVERSION}" PREFERRED_VERSION_libstd-rs ?= "${RUSTVERSION}" PREFERRED_VERSION_rust ?= "${RUSTVERSION}" PREFERRED_VERSION_rust-cross-${TARGET_ARCH} ?= "${RUSTVERSION}" -PREFERRED_VERSION_rust-llvm ?= "${RUSTVERSION}" -PREFERRED_VERSION_rust-llvm-native ?= "${RUSTVERSION}" PREFERRED_VERSION_rust-native ?= "${RUSTVERSION}"
meta/recipes-devtools/rust/rust-llvm/0001-AsmMatcherEmitter-sort-ClassInfo-lists-by-name-as-we.patch deleted +0 -31
@@ -1,31 +0,0 @@ -From 86940d87026432683fb6741cd8a34d3b9b18e40d Mon Sep 17 00:00:00 2001 -From: Alexander Kanavin <[email protected]> -Date: Fri, 27 Nov 2020 10:11:08 +0000 -Subject: [PATCH] AsmMatcherEmitter: sort ClassInfo lists by name as well - -Otherwise, there are instances which are identical in -every other field and therefore sort non-reproducibly -(which breaks binary and source reproducibiliy). - -Upstream-Status: Submitted [https://reviews.llvm.org/D97477] -Signed-off-by: Alexander Kanavin <[email protected]> ---- - llvm/utils/TableGen/AsmMatcherEmitter.cpp | 5 ++++- - 1 file changed, 4 insertions(+), 1 deletion(-) - -diff --git a/llvm/utils/TableGen/AsmMatcherEmitter.cpp b/llvm/utils/TableGen/AsmMatcherEmitter.cpp -index ccf0959389b..1f801e83b7d 100644 ---- a/llvm/utils/TableGen/AsmMatcherEmitter.cpp -+++ b/llvm/utils/TableGen/AsmMatcherEmitter.cpp -@@ -359,7 +359,10 @@ public: - // name of a class shouldn't be significant. However, some of the backends - // accidentally rely on this behaviour, so it will have to stay like this - // until they are fixed. -- return ValueName < RHS.ValueName; -+ if (ValueName != RHS.ValueName) -+ return ValueName < RHS.ValueName; -+ // All else being equal, we should sort by name, for source and binary reproducibility -+ return Name < RHS.Name; - } - }; -
meta/recipes-devtools/rust/rust-llvm/0002-llvm-allow-env-override-of-exe-path.patch deleted +0 -33
@@ -1,33 +0,0 @@ -From 7111770e8290082530d920e120995bf81431b0aa Mon Sep 17 00:00:00 2001 -From: Martin Kelly <[email protected]> -Date: Fri, 19 May 2017 00:22:57 -0700 -Subject: [PATCH 12/18] llvm: allow env override of exe path - -When using a native llvm-config from inside a sysroot, we need llvm-config to -return the libraries, include directories, etc. from inside the sysroot rather -than from the native sysroot. Thus provide an env override for calling -llvm-config from a target sysroot. - -Upstream-Status: Inappropriate [oe-core specific] -Signed-off-by: Martin Kelly <[email protected]> -Signed-off-by: Khem Raj <[email protected]> ---- - llvm/tools/llvm-config/llvm-config.cpp | 7 +++++++ - 1 file changed, 7 insertions(+) - ---- a/llvm/tools/llvm-config/llvm-config.cpp -+++ b/llvm/tools/llvm-config/llvm-config.cpp -@@ -226,6 +226,13 @@ Typical components:\n\ - - /// Compute the path to the main executable. - std::string GetExecutablePath(const char *Argv0) { -+ // Hack for Yocto: we need to override the root path when we are using -+ // llvm-config from within a target sysroot. -+ const char *Sysroot = std::getenv("YOCTO_ALTERNATE_EXE_PATH"); -+ if (Sysroot != nullptr) { -+ return Sysroot; -+ } -+ - // This just needs to be some symbol in the binary; C++ doesn't - // allow taking the address of ::main however. - void *P = (void *)(intptr_t)GetExecutablePath;
meta/recipes-devtools/rust/rust-llvm_1.90.0.bb deleted +0 -98
98 lines not on this page · raw message
Signature · 1 line
2.49.0
Quoted 27 lines · Deepesh via lists.openembedded.org
From: Deepesh Varatharajan <[email protected]>
Rightnow rust depends on llvm instead of rust-llvm
Setup in rust-llvm
CFLAGS:remove = "-g"
CXXFLAGS:remove = "-g"Setup in llvm
DEBUG_LEVELFLAG= "-g1"As a result, the stage1 compiler crate binaries include debug symbols, increasing their size. These binaries are used to run tests inside QEMU. To accommodate this, increase the QEMU RAM allocation to 1024 MB.
Signed-off-by Deepesh Varatharajan <[email protected]>---
meta/lib/oeqa/selftest/cases/rust.py | 5 +++--1 file changed, 3 insertions(+), 2 deletions(-)meta/lib/oeqa/selftest/cases/rust.py +0
-0@@ -102,8 +102,9 @@ class RustSelfTestSystemEmulated(OESelftestTestCase, OEPTestResultTestCase):testargs =
exclude_fail_tests+ " --no-doc --no-fail-fast --bless"
Quoted 9 lines
# wrap the execution with a qemu instance.
- # Tests are run with 512 tasks in parallel to execute all tests very quickly
- with runqemu("core-image-minimal", runqemuparams = "nographic", qemuparams = "-m 512") as qemu:
+ # Set QEMU RAM to 1024MB to support running unit tests for the compiler crate, including larger
+ # test cases and parallel execution in the test environment.
+ with runqemu("core-image-minimal", runqemuparams = "nographic", qemuparams = "-m 1024") as qemu:
# Copy remote-test-server to image through scp
host_sys = get_bb_var("RUST_BUILD_SYS", "rust")
ssh = SSHControl(ip=qemu.ip, logfile=qemu.sshlog, user="root")This failed in testing:
https://autobuilder.yoctoproject.org/valkyrie/#/builders/42/builds/2522
I suspect some qemu platforms don't allow changing the memory size so easily?
Cheers,
Richard
Quoted 42 lines · Richard Purdie
CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know the content is safe.Quoted 26 lines · Deepesh via lists.openembedded.org
From: Deepesh Varatharajan <[email protected]>
Rightnow rust depends on llvm instead of rust-llvm
Setup in rust-llvm
CFLAGS:remove = "-g"
CXXFLAGS:remove = "-g"Setup in llvm
DEBUG_LEVELFLAG= "-g1"As a result, the stage1 compiler crate binaries include debug symbols, increasing their size. These binaries are used to run tests inside QEMU. To accommodate this, increase the QEMU RAM allocation to 1024 MB.
Signed-off-by Deepesh Varatharajan <[email protected]>--- meta/lib/oeqa/selftest/cases/rust.py | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)meta/lib/oeqa/selftest/cases/rust.py +0
-0@@ -102,8 +102,9 @@ class RustSelfTestSystemEmulated(OESelftestTestCase, OEPTestResultTestCase): testargs = exclude_fail_tests + " --no-doc --no-fail-fast --bless"meta/lib/oeqa/selftest/cases/rust.py class RustSelfTestSystemEmulated(OESelftestTestCase, OEPTestResultTestCase): · line 110
Quoted 10 lines
meta/lib/oeqa/selftest/cases/rust.py +3
-2# wrap the execution with a qemu instance. - # Tests are run with 512 tasks in parallel to execute all tests very quickly - with runqemu("core-image-minimal", runqemuparams = "nographic", qemuparams = "-m 512") as qemu: + # Set QEMU RAM to 1024MB to support running unit tests for the compiler crate, including larger + # test cases and parallel execution in the test environment. + with runqemu("core-image-minimal", runqemuparams = "nographic", qemuparams = "-m 1024") as qemu: # Copy remote-test-server to image through scp host_sys = get_bb_var("RUST_BUILD_SYS", "rust") ssh = SSHControl(ip=qemu.ip, logfile=qemu.sshlog, user="root")This failed in testing:
Quoted 7 lines
https://autobuilder.yoctoproject.org/valkyrie/#/builders/42/builds/2522
I suspect some qemu platforms don't allow changing the memory size so easily?
Cheers,
Richard
Hi Richard,
I had a look at the failure and noticed that this test run includes a linux-firmware upgrade (20250917 → 20251011). While it’s not confirmed yet, it’s possible that the updated firmware may be affecting device initialization. Since the log shows a warning indicating that the target did not reach the login banner within 1000 seconds, confirming that the system hangs shortly after starting likely around the time virtio-gpu is initialized.
Additionally, I believe this issue is not related to the increased memory allocation because the GCC oe-selftest, which runs with 4096 MB of QEMU memory, completed successfully. I also checked previous autobuilder runs without this linux-firmware update which had these rust-llvm drop set of patches and the qemuarm-tc build completed without issues.
https://autobuilder.yoctoproject.org/valkyrie/?#/builders/42/builds/2519
Regards,
Deepesh