Update tensorflow/compiler/mlir/hlo/CMakeLists.txt
Co-authored-by: Geoffrey Martin-Noble <gcmn@google.com>
diff --git a/tensorflow/compiler/mlir/hlo/CMakeLists.txt b/tensorflow/compiler/mlir/hlo/CMakeLists.txt
index f2d31a0..e35c3ee 100644
--- a/tensorflow/compiler/mlir/hlo/CMakeLists.txt
+++ b/tensorflow/compiler/mlir/hlo/CMakeLists.txt
@@ -59,9 +59,9 @@
# MLIR/LLVM Configuration
#-------------------------------------------------------------------------------
-# Skip to find MLIR, which allows to build MHLO as part of another project.
-# This is necessary if the project which embeds MHLO uses a bundled version of
-# MLIR and does not use an installed version of MLIR.
+# Find MLIR to install if we are building standalone. If building as part of
+# another project, let it handle the MLIR dependency. The dependent project
+# might use a bundled version of MLIR instead of installing, for instance.
if(NOT MHLO_BUILD_EMBEDDED)
find_package(MLIR REQUIRED CONFIG)
message(STATUS "Using MLIRConfig.cmake in: ${MLIR_DIR}")
