diff --git a/mlir.spec b/mlir.spec index c4e25c6..f35a080 100644 --- a/mlir.spec +++ b/mlir.spec @@ -7,7 +7,7 @@ Name: mlir Version: %{mlir_version}%{?rc_ver:~rc%{rc_ver}} -Release: 1%{?dist} +Release: 2%{?dist} Summary: Multi-Level Intermediate Representation Overview License: Apache-2.0 WITH LLVM-exception @@ -65,6 +65,7 @@ find ../* -maxdepth 0 ! -name '%{name}' -exec rm -rf {} + -DLLVM_BUILD_LLVM_DYLIB=ON \ -DCMAKE_PREFIX_PATH=%{_libdir}/cmake/llvm/ \ -DLLVM_EXTERNAL_LIT=%{_bindir}/lit \ + -DLLVM_BUILD_TOOLS:BOOL=ON \ -DLLVM_BUILD_UTILS:BOOL=ON \ -DMLIR_INCLUDE_DOCS:BOOL=ON \ -DMLIR_INCLUDE_TESTS:BOOL=ON \ @@ -89,9 +90,15 @@ export LD_LIBRARY_PATH=%{_builddir}/%{mlir_srcdir}/%{name}/%{_build}/%{_lib} # Remove tablegen tests, as they rely on includes from llvm/. rm -rf test/mlir-tblgen -# TODO: Test currently fails on i686. %ifarch %{ix86} +# TODO: Test currently fails on i686. rm test/IR/file-metadata-resources.mlir + +# TODO: There's two issues here (see https://github.com/llvm/llvm-project/issues/58357): +# 1. The async dialect hardcodes a 64-bit assumption. +# 2. The cpu runner tests call mlir-opt without awareness of the host index size. +# For this reason, skip mlir-cpu-runner tests on 32-bit. +rm -rf test/mlir-cpu-runner %endif # Test execution normally relies on RPATH, so set LD_LIBRARY_PATH instead. @@ -109,7 +116,15 @@ export LD_LIBRARY_PATH=%{buildroot}/%{_libdir} %{_libdir}/libMLIR*.a %files devel +%{_bindir}/mlir-cpu-runner +%{_bindir}/mlir-linalg-ods-yaml-gen +%{_bindir}/mlir-lsp-server +%{_bindir}/mlir-opt +%{_bindir}/mlir-pdll-lsp-server +%{_bindir}/mlir-reduce %{_bindir}/mlir-tblgen +%{_bindir}/mlir-translate +%{_bindir}/tblgen-lsp-server %{_libdir}/libMLIR*.so %{_libdir}/libmlir_async_runtime.so %{_libdir}/libmlir_c_runner_utils.so @@ -119,6 +134,9 @@ export LD_LIBRARY_PATH=%{buildroot}/%{_libdir} %{_libdir}/cmake/mlir %changelog +* Thu Dec 22 2022 Nikita Popov - 15.0.6-2 +- rhbz#2127916: Add mlir tools to mlir-devel + * Mon Dec 05 2022 Nikita Popov - 15.0.6-1 - Update to LLVM 15.0.6