Index: lib/xray/xray_interface_internal.h
===================================================================
--- lib/xray/xray_interface_internal.h
+++ lib/xray/xray_interface_internal.h
@@ -28,13 +28,15 @@
   uint64_t Function;
   unsigned char Kind;
   unsigned char AlwaysInstrument;
-  unsigned char Padding[14]; // Need 32 bytes
+  unsigned char Version;
+  unsigned char Padding[13]; // Need 32 bytes
 #elif SANITIZER_WORDSIZE == 32
   uint32_t Address;
   uint32_t Function;
   unsigned char Kind;
   unsigned char AlwaysInstrument;
-  unsigned char Padding[6]; // Need 16 bytes
+  unsigned char Version;
+  unsigned char Padding[5]; // Need 16 bytes
 #else
 #error "Unsupported word size."
 #endif
Index: lib/xray/xray_trampoline_x86_64.S
===================================================================
--- lib/xray/xray_trampoline_x86_64.S
+++ lib/xray/xray_trampoline_x86_64.S
@@ -202,10 +202,7 @@
 	.type __xray_CustomEvent,@function
 __xray_CustomEvent:
   .cfi_startproc
-	subq $16, %rsp
-	.cfi_def_cfa_offset 24
-	movq %rbp, 8(%rsp)
-	movq %rax, 0(%rsp)
+	SAVE_REGISTERS
 
 	// We take two arguments to this trampoline, which should be in rdi	and rsi
 	// already. We also make sure that we stash %rax because we use that register
@@ -219,9 +216,7 @@
   callq *%rax
 
 .LcustomEventCleanup:
-	movq 0(%rsp), %rax
-	movq 8(%rsp), %rbp
-	addq $16, %rsp
+	RESTORE_REGISTERS
 	.cfi_def_cfa_offset 8
 	retq
 
Index: lib/xray/xray_x86_64.cc
===================================================================
--- lib/xray/xray_x86_64.cc
+++ lib/xray/xray_x86_64.cc
@@ -76,6 +76,7 @@
 static constexpr uint16_t MovR10Seq = 0xba41;
 static constexpr uint16_t Jmp9Seq = 0x09eb;
 static constexpr uint16_t Jmp20Seq = 0x14eb;
+static constexpr uint16_t Jmp15Seq = 0x0feb;
 static constexpr uint8_t JmpOpCode = 0xe9;
 static constexpr uint8_t RetOpCode = 0xc3;
 static constexpr uint16_t NopwSeq = 0x9066;
@@ -207,8 +208,10 @@
                       const XRaySledEntry &Sled) XRAY_NEVER_INSTRUMENT {
   // Here we do the dance of replacing the following sled:
   //
+  // In Version 0:
+  //
   // xray_sled_n:
-  //   jmp +19          // 2 bytes
+  //   jmp +20          // 2 bytes
   //   ...
   //
   // With the following:
@@ -216,24 +219,35 @@
   //   nopw             // 2 bytes*
   //   ...
   //
-  // We need to do this in the following order:
   //
-  // 1. Overwrite the 5-byte nop with the call (relative), where (relative) is
-  //    the relative offset to the __xray_CustomEvent trampoline.
-  // 2. Do a two-byte atomic write over the 'jmp +24' to turn it into a 'nopw'.
-  //    This allows us to "enable" this code once the changes have committed.
+  // The "unpatch" should just turn the 'nopw' back to a 'jmp +20'.
+  //
+  // ---
   //
-  // The "unpatch" should just turn the 'nopw' back to a 'jmp +24'.
+  // In Version 1:
+  //
+  //   The jump offset is now 15 bytes (0x0f), so when restoring the nopw back
+  //   to a jmp, use 15 bytes instead.
   //
   if (Enable) {
     std::atomic_store_explicit(
         reinterpret_cast<std::atomic<uint16_t> *>(Sled.Address), NopwSeq,
         std::memory_order_release);
   } else {
-    std::atomic_store_explicit(
-        reinterpret_cast<std::atomic<uint16_t> *>(Sled.Address), Jmp20Seq,
-        std::memory_order_release);
-  }
+    switch (Sled.Version) {
+    case 1:
+      std::atomic_store_explicit(
+          reinterpret_cast<std::atomic<uint16_t> *>(Sled.Address), Jmp15Seq,
+          std::memory_order_release);
+      break;
+    case 0:
+    default:
+      std::atomic_store_explicit(
+          reinterpret_cast<std::atomic<uint16_t> *>(Sled.Address), Jmp20Seq,
+          std::memory_order_release);
+      break;
+    }
+    }
   return false;
 }
 
Index: test/xray/TestCases/Linux/custom-event-handler-alignment.cc
===================================================================
--- /dev/null
+++ test/xray/TestCases/Linux/custom-event-handler-alignment.cc
@@ -0,0 +1,40 @@
+// Make sure we're aligning the stack properly when lowering the custom event
+// calls.
+//
+// RUN: %clangxx_xray -std=c++11 %s -o %t
+// RUN: XRAY_OPTIONS="patch_premain=false verbosity=1 xray_naive_log=false" \
+// RUN:     %run %t 2>&1
+// REQUIRES: x86_64-linux
+// REQUIRES: built-in-llvm-tree
+#include <xmmintrin.h>
+#include <stdio.h>
+#include "xray/xray_interface.h"
+
+[[clang::xray_never_instrument]] __attribute__((weak)) __m128 f(__m128 *i) {
+  return *i;
+}
+
+[[clang::xray_always_instrument]] void foo() {
+  __xray_customevent(0, 0);
+  __m128 v = {};
+  f(&v);
+}
+
+[[clang::xray_always_instrument]] void bar() {
+  __xray_customevent(0, 0);
+}
+
+void printer(void* ptr, size_t size) {
+  printf("handler called\n");
+}
+
+int main(int argc, char* argv[]) {
+  __xray_set_customevent_handler(printer);
+  __xray_patch();
+  foo();  // CHECK: handler called
+  bar();  // CHECK: handler called
+  __xray_unpatch();
+  __xray_remove_customevent_handler();
+  foo();
+  bar();
+}
Index: test/xray/TestCases/Linux/custom-event-logging.cc
===================================================================
--- test/xray/TestCases/Linux/custom-event-logging.cc
+++ test/xray/TestCases/Linux/custom-event-logging.cc
@@ -2,6 +2,8 @@
 //
 // RUN: %clangxx_xray -std=c++11 %s -o %t
 // RUN: XRAY_OPTIONS="patch_premain=false verbosity=1 xray_naive_log=false xray_logfile_base=custom-event-logging.xray-" %run %t 2>&1 | FileCheck %s
+// RUN: %clangxx_xray -std=c++11 -fpic -fpie %s -o %t
+// RUN: XRAY_OPTIONS="patch_premain=false verbosity=1 xray_naive_log=false xray_logfile_base=custom-event-logging.xray-" %run %t 2>&1 | FileCheck %s
 // FIXME: Support this in non-x86_64 as well
 // REQUIRES: x86_64-linux
 // REQUIRES: built-in-llvm-tree