[libc][GPU] Fix dependencies for externally installed stub files

[jhuber6]

Summary:
The GPU build has a lot of magic around how we package the output.
Generally, the GPU needs to exist as a secondary fatbinary image for
offloading languages. This is because offloading languages pretend like
offloading to an accelerator is a single file. This then needs to be put
into a single file to make it mesh with the existing build
infrastructure. To work with this, the libc makes an installed version
of the library that simply embeds the GPU code into an empty stub file.

This wasn't being updated correctly, which lead to the installed libc
static library not being updated correctly when the underlying file was
changed. The previous behaviour only updated when the entrypoint itself
was modified, but not any of its headers. By adding a dependcy on the
actual object file we should now capture the regular CMake semantics.

[llvmbot]

@llvm/pr-subscribers-libc

Changes

Summary:
The GPU build has a lot of magic around how we package the output.
Generally, the GPU needs to exist as a secondary fatbinary image for
offloading languages. This is because offloading languages pretend like
offloading to an accelerator is a single file. This then needs to be put
into a single file to make it mesh with the existing build
infrastructure. To work with this, the libc makes an installed version
of the library that simply embeds the GPU code into an empty stub file.

This wasn't being updated correctly, which lead to the installed libc
static library not being updated correctly when the underlying file was
changed. The previous behaviour only updated when the entrypoint itself
was modified, but not any of its headers. By adding a dependcy on the
actual object file we should now capture the regular CMake semantics.

---

Full diff: https://github.com/llvm/llvm-project/pull/66653.diff

1 Files Affected:

• (modified) libc/cmake/modules/LLVMLibCObjectRules.cmake (+3-3)

diff --git a/libc/cmake/modules/LLVMLibCObjectRules.cmake b/libc/cmake/modules/LLVMLibCObjectRules.cmake
index 891c298855e6131..994437f55d27407 100644
--- a/libc/cmake/modules/LLVMLibCObjectRules.cmake
+++ b/libc/cmake/modules/LLVMLibCObjectRules.cmake
@@ -210,7 +210,7 @@ function(_build_gpu_objects fq_target_name internal_target_name)
         list(APPEND packager_images
              --image=file=${input_file},arch=${gpu_arch},triple=${gpu_target_triple})
        endif()
-      list(APPEND gpu_target_names ${gpu_target_name})
+      list(APPEND gpu_target_objects ${input_file})
     endforeach()
 
     # After building the target for the desired GPUs we must package the output
@@ -222,7 +222,7 @@ function(_build_gpu_objects fq_target_name internal_target_name)
     add_custom_command(OUTPUT ${packaged_output_name}
                        COMMAND ${LIBC_CLANG_OFFLOAD_PACKAGER}
                                ${packager_images} -o ${packaged_output_name}
-                       DEPENDS ${gpu_target_names} ${add_gpu_obj_src} ${ADD_GPU_OBJ_HDRS}
+                       DEPENDS ${gpu_target_objects} ${add_gpu_obj_src} ${ADD_GPU_OBJ_HDRS}
                        COMMENT "Packaging LLVM offloading binary")
     add_custom_target(${packaged_target_name} DEPENDS ${packaged_output_name})
     list(APPEND packaged_gpu_names ${packaged_target_name})
@@ -242,7 +242,7 @@ function(_build_gpu_objects fq_target_name internal_target_name)
     OUTPUT "${CMAKE_CURRENT_BINARY_DIR}/stubs/${stub_filename}"
     COMMAND ${CMAKE_COMMAND} -E make_directory ${CMAKE_CURRENT_BINARY_DIR}/stubs/
     COMMAND ${CMAKE_COMMAND} -E touch ${CMAKE_CURRENT_BINARY_DIR}/stubs/${stub_filename}
-    DEPENDS ${gpu_target_names} ${ADD_GPU_OBJ_SRCS} ${ADD_GPU_OBJ_HDRS}
+    DEPENDS ${gpu_target_objects} ${ADD_GPU_OBJ_SRCS} ${ADD_GPU_OBJ_HDRS}
   )
   set(stub_target_name ${fq_target_name}.__stub__)
   add_custom_target(${stub_target_name} DEPENDS ${CMAKE_CURRENT_BINARY_DIR}/stubs/${stub_filename})