Index: lib/Transforms/Instrumentation/AddressSanitizer.cpp
===================================================================
--- lib/Transforms/Instrumentation/AddressSanitizer.cpp
+++ lib/Transforms/Instrumentation/AddressSanitizer.cpp
@@ -947,6 +947,14 @@
     PtrOperand = XCHG->getPointerOperand();
   }
 
+  // Do not instrument acesses from different address spaces; we cannot deal
+  // with them.
+  if (PtrOperand) {
+    Type *PtrTy = cast<PointerType>(PtrOperand->getType()->getScalarType());
+    if (PtrTy->getPointerAddressSpace() != 0)
+      return nullptr;
+  }
+
   // Treat memory accesses to promotable allocas as non-interesting since they
   // will not cause memory violations. This greatly speeds up the instrumented
   // executable at -O0.
Index: lib/Transforms/Instrumentation/ThreadSanitizer.cpp
===================================================================
--- lib/Transforms/Instrumentation/ThreadSanitizer.cpp
+++ lib/Transforms/Instrumentation/ThreadSanitizer.cpp
@@ -259,6 +259,15 @@
         return false;
     }
   }
+
+  // Do not instrument acesses from different address spaces; we cannot deal
+  // with them.
+  if (Addr) {
+    Type *PtrTy = cast<PointerType>(Addr->getType()->getScalarType());
+    if (PtrTy->getPointerAddressSpace() != 0)
+      return false;
+  }
+
   return true;
 }
 
Index: test/Instrumentation/AddressSanitizer/asan_address_space_attr.ll
===================================================================
--- /dev/null
+++ test/Instrumentation/AddressSanitizer/asan_address_space_attr.ll
@@ -0,0 +1,32 @@
+; RUN: opt < %s -asan -S | FileCheck %s
+target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.11.0"
+
+; Checks that we do not instrument loads and stores comming from custom address space.
+; These result in invalid (false positive) reports.
+; int foo(int argc, const char * argv[]) {
+;   void *__attribute__((address_space(256))) *gs_base = (((void * __attribute__((address_space(256))) *)0));
+;   void *somevalue = gs_base[-1];
+;   return somevalue;
+; }
+
+define i32 @foo(i32 %argc, i8** %argv) sanitize_address {
+entry:
+  %retval = alloca i32, align 4
+  %argc.addr = alloca i32, align 4
+  %argv.addr = alloca i8**, align 8
+  %gs_base = alloca i8* addrspace(256)*, align 8
+  %somevalue = alloca i8*, align 8
+  store i32 0, i32* %retval, align 4
+  store i32 %argc, i32* %argc.addr, align 4
+  store i8** %argv, i8*** %argv.addr, align 8
+  store i8* addrspace(256)* null, i8* addrspace(256)** %gs_base, align 8
+  %0 = load i8* addrspace(256)*, i8* addrspace(256)** %gs_base, align 8
+  %arrayidx = getelementptr inbounds i8*, i8* addrspace(256)* %0, i64 -1
+  %1 = load i8*, i8* addrspace(256)* %arrayidx, align 8
+  store i8* %1, i8** %somevalue, align 8
+  %2 = load i8*, i8** %somevalue, align 8
+  %3 = ptrtoint i8* %2 to i32
+  ret i32 %3
+}
+; CHECK-NOT: call void @__asan_report_load8
Index: test/Instrumentation/ThreadSanitizer/tsan_address_space_attr.ll
===================================================================
--- /dev/null
+++ test/Instrumentation/ThreadSanitizer/tsan_address_space_attr.ll
@@ -0,0 +1,32 @@
+; RUN: opt < %s -tsan -S | FileCheck %s
+target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.11.0"
+
+; Checks that we do not instrument loads and stores comming from custom address space.
+; These result in crashing the compiler.
+; int foo(int argc, const char * argv[]) {
+;   void *__attribute__((address_space(256))) *gs_base = (((void * __attribute__((address_space(256))) *)0));
+;   void *somevalue = gs_base[-1];
+;   return somevalue;
+; }
+
+define i32 @foo(i32 %argc, i8** %argv) sanitize_thread {
+entry:
+  %retval = alloca i32, align 4
+  %argc.addr = alloca i32, align 4
+  %argv.addr = alloca i8**, align 8
+  %gs_base = alloca i8* addrspace(256)*, align 8
+  %somevalue = alloca i8*, align 8
+  store i32 0, i32* %retval, align 4
+  store i32 %argc, i32* %argc.addr, align 4
+  store i8** %argv, i8*** %argv.addr, align 8
+  store i8* addrspace(256)* null, i8* addrspace(256)** %gs_base, align 8
+  %0 = load i8* addrspace(256)*, i8* addrspace(256)** %gs_base, align 8
+  %arrayidx = getelementptr inbounds i8*, i8* addrspace(256)* %0, i64 -1
+  %1 = load i8*, i8* addrspace(256)* %arrayidx, align 8
+  store i8* %1, i8** %somevalue, align 8
+  %2 = load i8*, i8** %somevalue, align 8
+  %3 = ptrtoint i8* %2 to i32
+  ret i32 %3
+}
+; CHECK-NOT: addrspacecast