vendor/lld/lld-trunk-r256633

1 files changed, 34 insertions, 0 deletions

diff --git a/lib/ReaderWriter/ELF/AMDGPU/AMDGPUExecutableWriter.cpp b/lib/ReaderWriter/ELF/AMDGPU/AMDGPUExecutableWriter.cpp
new file mode 100644
index 000000000000..89efeb23d6f8
--- /dev/null
+++ b/lib/ReaderWriter/ELF/AMDGPU/AMDGPUExecutableWriter.cpp
@@ -0,0 +1,34 @@
+//===- lib/ReaderWriter/ELF/AMDGPU/AMDGPUExecutableWriter.cpp -------------===//
+//
+//                             The LLVM Linker
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#include "AMDGPUExecutableWriter.h"
+
+using namespace lld;
+using namespace lld::elf;
+
+AMDGPUExecutableWriter::AMDGPUExecutableWriter(AMDGPULinkingContext &ctx,
+                                               AMDGPUTargetLayout &layout)
+    : ExecutableWriter(ctx, layout), _ctx(ctx) {}
+
+void AMDGPUExecutableWriter::createImplicitFiles(
+    std::vector<std::unique_ptr<File>> &Result) {
+  // ExecutableWriter::createImplicitFiles() adds C runtime symbols that we
+  // don't need, so we use the OutputELFWriter implementation instead.
+  OutputELFWriter<ELF64LE>::createImplicitFiles(Result);
+}
+
+void AMDGPUExecutableWriter::finalizeDefaultAtomValues() {
+
+  // ExecutableWriter::finalizeDefaultAtomValues() assumes the presence of
+  // C runtime symbols.  However, since we skip the call to
+  // ExecutableWriter::createImplicitFiles(), these symbols are never added
+  // and ExectuableWriter::finalizeDefaultAtomValues() will crash if we call
+  // it.
+  OutputELFWriter<ELF64LE>::finalizeDefaultAtomValues();
+}