release/23.x: [Clang][Interpreter] Don't pass -fPIC on MSVC (#210507)

[llvmbot]

Backport 617e4f5

Requested by: @dyung

[llvmbot]

@vgvassilev What do you think about merging this PR to the release branch?

[llvmorg-github-actions]

@llvm/pr-subscribers-clang

Author: llvmbot

Changes

Backport 617e4f5

Requested by: @dyung

---

Full diff: https://github.com/llvm/llvm-project/pull/210642.diff

1 Files Affected:

• (modified) clang/lib/Interpreter/Interpreter.cpp (+8-2)

diff --git a/clang/lib/Interpreter/Interpreter.cpp b/clang/lib/Interpreter/Interpreter.cpp
index 933a68b50db41..0536fdcd548a2 100644
--- a/clang/lib/Interpreter/Interpreter.cpp
+++ b/clang/lib/Interpreter/Interpreter.cpp
@@ -53,6 +53,7 @@
 #include "llvm/Support/VirtualFileSystem.h"
 #include "llvm/Support/raw_ostream.h"
 #include "llvm/TargetParser/Host.h"
+#include "llvm/TargetParser/Triple.h"
 #include "llvm/Transforms/Utils/Cloning.h" // for CloneModule
 
 #define DEBUG_TYPE "clang-repl"
@@ -228,8 +229,13 @@ IncrementalCompilerBuilder::create(std::string TT,
   // host symbol may be out of range when the JIT memory is mapped more than
   // 2GB away (as on FreeBSD), breaking tests such as
   // Interpreter/simple-exception.cpp. Insert before user arguments so it can
-  // still be overridden.
-  ClangArgv.insert(ClangArgv.begin() + 1, "-fPIC");
+  // still be overridden. On Windows (excluding Cygwin/MinGW) an explicit
+  // -fPIC is an unsupported driver option that would drop non-x86_64 targets
+  // to PIC level 0; PIC is already the forced default there where relevant,
+  // so don't inject it.
+  llvm::Triple TargetTriple(TT);
+  if (!TargetTriple.isOSWindows() || TargetTriple.isOSCygMing())
+    ClangArgv.insert(ClangArgv.begin() + 1, "-fPIC");
 
   // Prepending -c to force the driver to do something if no action was
   // specified. By prepending we allow users to override the default

[dyung]

This is needed to fix a test failure in the Windows-11-arm build which blocks it from completing successfully.

[vgvassilev]

Looks like a pre merge check is failing.

[vgvassilev]

Looks like a pre merge check is failing.

[dyung]

Looks like a pre merge check is failing.

The pre-merge check looks more like an infrastructure failure to me, but I've asked @tstellar to take a look since he had previously fixed this test.

On the release branch, I've just been committing fixes even with that failure, but still need a review!

[vgvassilev]

Lgtm, if the premerge checks fail due to infra.