Index: include/llvm/Target/TargetRegisterInfo.h
===================================================================
--- include/llvm/Target/TargetRegisterInfo.h
+++ include/llvm/Target/TargetRegisterInfo.h
@@ -460,6 +460,10 @@
     llvm_unreachable("target does not provide no preserved mask");
   }
 
+  /// Return true if all bits in the register mask \p mask0 are set in \p mask1
+  /// as well.
+  bool regmaskSubsetEqual(const uint32_t *mask0, const uint32_t *mask1) const;
+
   /// Return all the call-preserved register masks defined for this target.
   virtual ArrayRef<const uint32_t *> getRegMasks() const = 0;
   virtual ArrayRef<const char *> getRegMaskNames() const = 0;
Index: lib/CodeGen/TargetRegisterInfo.cpp
===================================================================
--- lib/CodeGen/TargetRegisterInfo.cpp
+++ lib/CodeGen/TargetRegisterInfo.cpp
@@ -388,6 +388,16 @@
   return false;
 }
 
+bool TargetRegisterInfo::regmaskSubsetEqual(const uint32_t *mask0,
+                                            const uint32_t *mask1) const {
+  unsigned N = (getNumRegs()+31) / 32;
+  for (unsigned I = 0; I < N; ++I) {
+    if ((mask0[I] & mask1[I]) != mask0[I])
+      return false;
+  }
+  return true;
+}
+
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
 void
 TargetRegisterInfo::dumpReg(unsigned Reg, unsigned SubRegIndex,
Index: lib/Target/AArch64/AArch64ISelLowering.cpp
===================================================================
--- lib/Target/AArch64/AArch64ISelLowering.cpp
+++ lib/Target/AArch64/AArch64ISelLowering.cpp
@@ -2816,13 +2816,6 @@
   CallingConv::ID CallerCC = CallerF->getCallingConv();
   bool CCMatch = CallerCC == CalleeCC;
 
-  // Disable tailcall for CXX_FAST_TLS when callee and caller have different
-  // calling conventions, given that CXX_FAST_TLS has a bigger CSR set.
-  if (!CCMatch &&
-      (CallerCC == CallingConv::CXX_FAST_TLS ||
-       CalleeCC == CallingConv::CXX_FAST_TLS))
-    return false;
-
   // Byval parameters hand the function a pointer directly into the stack area
   // we want to reuse during a tail call. Working around this *is* possible (see
   // X86) but less efficient and uglier in LowerCall.
@@ -2882,6 +2875,13 @@
                                   CCAssignFnForCall(CalleeCC, isVarArg),
                                   CCAssignFnForCall(CallerCC, isVarArg)))
     return false;
+  // The callee has to preserve all registers the caller needs to preserve.
+  if (!CCMatch) {
+    const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
+    if (!TRI->regmaskSubsetEqual(TRI->getCallPreservedMask(MF, CallerCC),
+                                 TRI->getCallPreservedMask(MF, CalleeCC)))
+      return false;
+  }
 
   // Nothing more to check if the callee is taking no arguments
   if (Outs.empty())
Index: lib/Target/ARM/ARMISelLowering.cpp
===================================================================
--- lib/Target/ARM/ARMISelLowering.cpp
+++ lib/Target/ARM/ARMISelLowering.cpp
@@ -2105,14 +2105,6 @@
   MachineFunction &MF = DAG.getMachineFunction();
   const Function *CallerF = MF.getFunction();
   CallingConv::ID CallerCC = CallerF->getCallingConv();
-  bool CCMatch = CallerCC == CalleeCC;
-
-  // Disable tailcall for CXX_FAST_TLS when callee and caller have different
-  // calling conventions, given that CXX_FAST_TLS has a bigger CSR set.
-  if (!CCMatch &&
-      (CallerCC == CallingConv::CXX_FAST_TLS ||
-       CalleeCC == CallingConv::CXX_FAST_TLS))
-    return false;
 
   assert(Subtarget->supportsTailCall());
 
@@ -2156,6 +2148,13 @@
                                   CCAssignFnForNode(CalleeCC, true, isVarArg),
                                   CCAssignFnForNode(CallerCC, true, isVarArg)))
     return false;
+  // The callee has to preserve all registers the caller needs to preserve.
+  if (CalleeCC != CallerCC) {
+    const ARMBaseRegisterInfo *TRI = Subtarget->getRegisterInfo();
+    if (!TRI->regmaskSubsetEqual(TRI->getCallPreservedMask(MF, CallerCC),
+                                 TRI->getCallPreservedMask(MF, CalleeCC)))
+      return false;
+  }
 
   // If Caller's vararg or byval argument has been split between registers and
   // stack, do not perform tail call, since part of the argument is in caller's
Index: lib/Target/X86/X86ISelLowering.cpp
===================================================================
--- lib/Target/X86/X86ISelLowering.cpp
+++ lib/Target/X86/X86ISelLowering.cpp
@@ -3802,13 +3802,6 @@
   if (IsCalleeWin64 != IsCallerWin64)
     return false;
 
-  // Disable tailcall for CXX_FAST_TLS when callee and caller have different
-  // calling conventions, given that CXX_FAST_TLS has a bigger CSR set.
-  if (!CCMatch &&
-      (CallerCC == CallingConv::CXX_FAST_TLS ||
-       CalleeCC == CallingConv::CXX_FAST_TLS))
-    return false;
-
   if (DAG.getTarget().Options.GuaranteedTailCallOpt) {
     if (canGuaranteeTCO(CalleeCC) && CCMatch)
       return true;
@@ -3872,6 +3865,13 @@
   if (!CCState::resultsCompatible(CalleeCC, CallerCC, MF, C, Ins,
                                   RetCC_X86, RetCC_X86))
     return false;
+  // The callee has to preserve all registers the caller needs to preserve.
+  if (!CCMatch) {
+    const X86RegisterInfo *TRI = Subtarget.getRegisterInfo();
+    if (!TRI->regmaskSubsetEqual(TRI->getCallPreservedMask(MF, CallerCC),
+                                 TRI->getCallPreservedMask(MF, CalleeCC)))
+      return false;
+  }
 
   unsigned StackArgsSize = 0;
 
Index: test/CodeGen/AArch64/tailcall-ccmismatch.ll
===================================================================
--- /dev/null
+++ test/CodeGen/AArch64/tailcall-ccmismatch.ll
@@ -0,0 +1,18 @@
+; RUN: llc -o - %s | FileCheck %s
+target triple="aarch64--"
+
+declare void @somefunc()
+define preserve_mostcc void @test_ccmismatch_notail() {
+; CHECK-LABEL: test_ccmismatch_notail:
+; CHECK: bl somefunc
+  tail call void @somefunc()
+  ret void
+}
+
+declare preserve_mostcc void @some_preserve_most_func()
+define void @test_ccmismatch_tail() {
+; CHECK-LABEL: test_ccmismatch_tail:
+; CHECK: b some_preserve_most_func
+  tail call preserve_mostcc void @some_preserve_most_func()
+  ret void
+}
Index: test/CodeGen/ARM/cxx-tlscc.ll
===================================================================
--- test/CodeGen/ARM/cxx-tlscc.ll
+++ test/CodeGen/ARM/cxx-tlscc.ll
@@ -126,5 +126,23 @@
   ret void
 }
 
+declare void @somefunc()
+define cxx_fast_tlscc void @test_ccmismatch_notail() {
+; CHECK-LABEL: test_ccmismatch_notail:
+; CHECK-NOT: b _somefunc
+; CHECK: bl _somefunc
+  tail call void @somefunc()
+  ret void
+}
+
+declare cxx_fast_tlscc void @some_fast_tls_func()
+define void @test_ccmismatch_tail() {
+; CHECK-LABEL: test_ccmismatch_tail:
+; CHECK-NOT: bl _some_fast_tls_func
+; CHECK: b _some_fast_tls_func
+  tail call cxx_fast_tlscc void @some_fast_tls_func()
+  ret void
+}
+
 attributes #0 = { nounwind "no-frame-pointer-elim"="true" }
 attributes #1 = { nounwind }