diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -277,6 +277,7 @@
 --------------------
 
 - Support ``-mindirect-branch-cs-prefix`` for call and jmp to indirect thunk.
+- Fix 32-bit ``__fastcall`` and ``__vectorcall`` ABI mismatch with MSVC.
 
 DWARF Support in Clang
 ----------------------
diff --git a/clang/lib/CodeGen/TargetInfo.cpp b/clang/lib/CodeGen/TargetInfo.cpp
--- a/clang/lib/CodeGen/TargetInfo.cpp
+++ b/clang/lib/CodeGen/TargetInfo.cpp
@@ -1771,21 +1771,22 @@
 }
 
 bool X86_32ABIInfo::shouldPrimitiveUseInReg(QualType Ty, CCState &State) const {
-  if (!updateFreeRegs(Ty, State))
+  bool IsPtrOrInt = (getContext().getTypeSize(Ty) <= 32) &&
+                    (Ty->isIntegralOrEnumerationType() || Ty->isPointerType() ||
+                     Ty->isReferenceType());
+
+  if (!IsPtrOrInt && (State.CC == llvm::CallingConv::X86_FastCall ||
+                      State.CC == llvm::CallingConv::X86_VectorCall))
     return false;
 
-  if (IsMCUABI)
+  if (!updateFreeRegs(Ty, State))
     return false;
 
-  if (State.CC == llvm::CallingConv::X86_FastCall ||
-      State.CC == llvm::CallingConv::X86_VectorCall ||
-      State.CC == llvm::CallingConv::X86_RegCall) {
-    if (getContext().getTypeSize(Ty) > 32)
-      return false;
+  if (!IsPtrOrInt && State.CC == llvm::CallingConv::X86_RegCall)
+    return false;
 
-    return (Ty->isIntegralOrEnumerationType() || Ty->isPointerType() ||
-        Ty->isReferenceType());
-  }
+  if (IsMCUABI)
+    return false;
 
   return true;
 }
diff --git a/clang/test/CodeGen/mangle-windows.c b/clang/test/CodeGen/mangle-windows.c
--- a/clang/test/CodeGen/mangle-windows.c
+++ b/clang/test/CodeGen/mangle-windows.c
@@ -47,7 +47,7 @@
 // X64: define dso_local void @f8(
 
 void __fastcall f9(long long a, char b, char c, short d) {}
-// CHECK: define dso_local x86_fastcallcc void @"\01@f9@20"(i64 noundef %a, i8 noundef signext %b, i8 noundef signext %c, i16 noundef signext %d)
+// CHECK: define dso_local x86_fastcallcc void @"\01@f9@20"(i64 noundef %a, i8 inreg noundef signext %b, i8 inreg noundef signext %c, i16 noundef signext %d)
 // X64: define dso_local void @f9(
 
 void f12(void) {}
@@ -81,3 +81,6 @@
 void __vectorcall v6(char a, char b) {}
 // CHECK: define dso_local x86_vectorcallcc void @"\01v6@@8"(
 // X64: define dso_local x86_vectorcallcc void @"\01v6@@16"(
+
+void __vectorcall v7(long long a, char b, char c, short d) {}
+// CHECK: define dso_local x86_vectorcallcc void @"\01v7@@20"(i64 noundef %a, i8 inreg noundef signext %b, i8 inreg noundef signext %c, i16 noundef signext %d)