libsanitizer: merge from upstream (c425db2eb558c263)

The following patch is result of libsanitizer/merge.sh
from c425db2eb558c263 (yesterday evening).

Bootstrapped/regtested on x86_64-linux and i686-linux (together with
the follow-up 3 patches I'm about to post).

BTW, seems upstream has added riscv64 support for I think lsan/tsan,
so if anyone is willing to try it there, it would be a matter of
copying e.g. the s390*-*-linux* libsanitizer/configure.tgt entry
to riscv64-*-linux* with the obvious s/s390x/riscv64/ change in it.

5 files changed, 234 insertions, 97 deletions

diff --git a/libsanitizer/interception/interception.h b/libsanitizer/interception/interception.h
index d97974e..069f73d 100644
--- a/libsanitizer/interception/interception.h
+++ b/libsanitizer/interception/interception.h
@@ -14,9 +14,10 @@
 #ifndef INTERCEPTION_H
 #define INTERCEPTION_H
 
+#include "sanitizer_common/sanitizer_asm.h"
 #include "sanitizer_common/sanitizer_internal_defs.h"
 
-#if !SANITIZER_LINUX && !SANITIZER_FREEBSD && !SANITIZER_APPLE &&      \
+#if !SANITIZER_LINUX && !SANITIZER_FREEBSD && !SANITIZER_APPLE &&    \
     !SANITIZER_NETBSD && !SANITIZER_WINDOWS && !SANITIZER_FUCHSIA && \
     !SANITIZER_SOLARIS
 #  error "Interception doesn't work on this operating system."
@@ -67,24 +68,50 @@ typedef __sanitizer::OFF64_T OFF64_T;
 //           for more details). To intercept such functions you need to use the
 //           INTERCEPTOR_WITH_SUFFIX(...) macro.
 
-// How it works:
-// To replace system functions on Linux we just need to declare functions
-// with same names in our library and then obtain the real function pointers
+// How it works on Linux
+// ---------------------
+//
+// To replace system functions on Linux we just need to declare functions with
+// the same names in our library and then obtain the real function pointers
 // using dlsym().
-// There is one complication. A user may also intercept some of the functions
-// we intercept. To resolve this we declare our interceptors with __interceptor_
-// prefix, and then make actual interceptors weak aliases to __interceptor_
-// functions.
 //
-// This is not so on Mac OS, where the two-level namespace makes
-// our replacement functions invisible to other libraries. This may be overcomed
-// using the DYLD_FORCE_FLAT_NAMESPACE, but some errors loading the shared
-// libraries in Chromium were noticed when doing so.
+// There is one complication: a user may also intercept some of the functions we
+// intercept. To allow for up to 3 interceptors (including ours) of a given
+// function "func", the interceptor implementation is in ___interceptor_func,
+// which is aliased by a weak function __interceptor_func, which in turn is
+// aliased (via a trampoline) by weak wrapper function "func".
+//
+// Most user interceptors should define a foreign interceptor as follows:
+//
+//  - provide a non-weak function "func" that performs interception;
+//  - if __interceptor_func exists, call it to perform the real functionality;
+//  - if it does not exist, figure out the real function and call it instead.
+//
+// In rare cases, a foreign interceptor (of another dynamic analysis runtime)
+// may be defined as follows (on supported architectures):
+//
+//  - provide a non-weak function __interceptor_func that performs interception;
+//  - if ___interceptor_func exists, call it to perform the real functionality;
+//  - if it does not exist, figure out the real function and call it instead;
+//  - provide a weak function "func" that is an alias to __interceptor_func.
+//
+// With this protocol, sanitizer interceptors, foreign user interceptors, and
+// foreign interceptors of other dynamic analysis runtimes, or any combination
+// thereof, may co-exist simultaneously.
+//
+// How it works on Mac OS
+// ----------------------
+//
+// This is not so on Mac OS, where the two-level namespace makes our replacement
+// functions invisible to other libraries. This may be overcomed using the
+// DYLD_FORCE_FLAT_NAMESPACE, but some errors loading the shared libraries in
+// Chromium were noticed when doing so.
+//
 // Instead we create a dylib containing a __DATA,__interpose section that
 // associates library functions with their wrappers. When this dylib is
-// preloaded before an executable using DYLD_INSERT_LIBRARIES, it routes all
-// the calls to interposed functions done through stubs to the wrapper
-// functions.
+// preloaded before an executable using DYLD_INSERT_LIBRARIES, it routes all the
+// calls to interposed functions done through stubs to the wrapper functions.
+//
 // As it's decided at compile time which functions are to be intercepted on Mac,
 // INTERCEPT_FUNCTION() is effectively a no-op on this system.
 
@@ -100,53 +127,102 @@ struct interpose_substitution {
 // For a function foo() create a global pair of pointers { wrap_foo, foo } in
 // the __DATA,__interpose section.
 // As a result all the calls to foo() will be routed to wrap_foo() at runtime.
-#define INTERPOSER(func_name) __attribute__((used)) \
+#define INTERPOSER(func_name) __attribute__((used))     \
 const interpose_substitution substitution_##func_name[] \
     __attribute__((section("__DATA, __interpose"))) = { \
-    { reinterpret_cast<const uptr>(WRAP(func_name)), \
-      reinterpret_cast<const uptr>(func_name) } \
+    { reinterpret_cast<const uptr>(WRAP(func_name)),    \
+      reinterpret_cast<const uptr>(func_name) }         \
 }
 
 // For a function foo() and a wrapper function bar() create a global pair
 // of pointers { bar, foo } in the __DATA,__interpose section.
 // As a result all the calls to foo() will be routed to bar() at runtime.
 #define INTERPOSER_2(func_name, wrapper_name) __attribute__((used)) \
-const interpose_substitution substitution_##func_name[] \
-    __attribute__((section("__DATA, __interpose"))) = { \
-    { reinterpret_cast<const uptr>(wrapper_name), \
-      reinterpret_cast<const uptr>(func_name) } \
+const interpose_substitution substitution_##func_name[]             \
+    __attribute__((section("__DATA, __interpose"))) = {             \
+    { reinterpret_cast<const uptr>(wrapper_name),                   \
+      reinterpret_cast<const uptr>(func_name) }                     \
 }
 
 # define WRAP(x) wrap_##x
-# define WRAPPER_NAME(x) "wrap_"#x
+# define TRAMPOLINE(x) WRAP(x)
 # define INTERCEPTOR_ATTRIBUTE
 # define DECLARE_WRAPPER(ret_type, func, ...)
 
 #elif SANITIZER_WINDOWS
 # define WRAP(x) __asan_wrap_##x
-# define WRAPPER_NAME(x) "__asan_wrap_"#x
+# define TRAMPOLINE(x) WRAP(x)
 # define INTERCEPTOR_ATTRIBUTE __declspec(dllexport)
-# define DECLARE_WRAPPER(ret_type, func, ...) \
+# define DECLARE_WRAPPER(ret_type, func, ...)         \
     extern "C" ret_type func(__VA_ARGS__);
-# define DECLARE_WRAPPER_WINAPI(ret_type, func, ...) \
+# define DECLARE_WRAPPER_WINAPI(ret_type, func, ...)  \
     extern "C" __declspec(dllimport) ret_type __stdcall func(__VA_ARGS__);
-#elif SANITIZER_FREEBSD || SANITIZER_NETBSD
-# define WRAP(x) __interceptor_ ## x
-# define WRAPPER_NAME(x) "__interceptor_" #x
+#elif !SANITIZER_FUCHSIA  // LINUX, FREEBSD, NETBSD, SOLARIS
 # define INTERCEPTOR_ATTRIBUTE __attribute__((visibility("default")))
+# if ASM_INTERCEPTOR_TRAMPOLINE_SUPPORT
+// Weak aliases of weak aliases do not work, therefore we need to set up a
+// trampoline function. The function "func" is a weak alias to the trampoline
+// (so that we may check if "func" was overridden), which calls the weak
+// function __interceptor_func, which in turn aliases the actual interceptor
+// implementation ___interceptor_func:
+//
+//    [wrapper "func": weak] --(alias)--> [TRAMPOLINE(func)]
+//                                                |
+//                     +--------(tail call)-------+
+//                     |
+//                     v
+//      [__interceptor_func: weak] --(alias)--> [WRAP(func)]
+//
+// We use inline assembly to define most of this, because not all compilers
+// support functions with the "naked" attribute with every architecture.
+#  define WRAP(x) ___interceptor_ ## x
+#  define TRAMPOLINE(x) __interceptor_trampoline_ ## x
+#  if SANITIZER_FREEBSD || SANITIZER_NETBSD
 // FreeBSD's dynamic linker (incompliantly) gives non-weak symbols higher
 // priority than weak ones so weak aliases won't work for indirect calls
 // in position-independent (-fPIC / -fPIE) mode.
-# define DECLARE_WRAPPER(ret_type, func, ...) \
-     extern "C" ret_type func(__VA_ARGS__) \
-     __attribute__((alias("__interceptor_" #func), visibility("default")));
-#elif !SANITIZER_FUCHSIA
-# define WRAP(x) __interceptor_ ## x
-# define WRAPPER_NAME(x) "__interceptor_" #x
-# define INTERCEPTOR_ATTRIBUTE __attribute__((visibility("default")))
-# define DECLARE_WRAPPER(ret_type, func, ...) \
-    extern "C" ret_type func(__VA_ARGS__) \
-    __attribute__((weak, alias("__interceptor_" #func), visibility("default")));
+#   define __ASM_WEAK_WRAPPER(func) ".globl " #func "\n"
+#  else
+#   define __ASM_WEAK_WRAPPER(func) ".weak " #func "\n"
+#  endif  // SANITIZER_FREEBSD || SANITIZER_NETBSD
+// Keep trampoline implementation in sync with sanitizer_common/sanitizer_asm.h
+#  define DECLARE_WRAPPER(ret_type, func, ...)                                 \
+     extern "C" ret_type func(__VA_ARGS__);                                    \
+     extern "C" ret_type TRAMPOLINE(func)(__VA_ARGS__);                        \
+     extern "C" ret_type __interceptor_##func(__VA_ARGS__)                     \
+       INTERCEPTOR_ATTRIBUTE __attribute__((weak)) ALIAS(WRAP(func));          \
+     asm(                                                                      \
+       ".text\n"                                                               \
+       __ASM_WEAK_WRAPPER(func)                                                \
+       ".set " #func ", " SANITIZER_STRINGIFY(TRAMPOLINE(func)) "\n"           \
+       ".globl " SANITIZER_STRINGIFY(TRAMPOLINE(func)) "\n"                    \
+       ".type  " SANITIZER_STRINGIFY(TRAMPOLINE(func)) ", %function\n"         \
+       SANITIZER_STRINGIFY(TRAMPOLINE(func)) ":\n"                             \
+       SANITIZER_STRINGIFY(CFI_STARTPROC) "\n"                                 \
+       SANITIZER_STRINGIFY(ASM_TAIL_CALL) " __interceptor_"                    \
+         SANITIZER_STRINGIFY(ASM_PREEMPTIBLE_SYM(func)) "\n"                   \
+       SANITIZER_STRINGIFY(CFI_ENDPROC) "\n"                                   \
+       ".size  " SANITIZER_STRINGIFY(TRAMPOLINE(func)) ", "                    \
+            ".-" SANITIZER_STRINGIFY(TRAMPOLINE(func)) "\n"                    \
+     );
+# else  // ASM_INTERCEPTOR_TRAMPOLINE_SUPPORT
+// Some architectures cannot implement efficient interceptor trampolines with
+// just a plain jump due to complexities of resolving a preemptible symbol. In
+// those cases, revert to just this scheme:
+//
+//    [wrapper "func": weak] --(alias)--> [WRAP(func)]
+//
+#  define WRAP(x) __interceptor_ ## x
+#  define TRAMPOLINE(x) WRAP(x)
+#  if SANITIZER_FREEBSD || SANITIZER_NETBSD
+#   define __ATTRIBUTE_WEAK_WRAPPER
+#  else
+#   define __ATTRIBUTE_WEAK_WRAPPER __attribute__((weak))
+#  endif  // SANITIZER_FREEBSD || SANITIZER_NETBSD
+#  define DECLARE_WRAPPER(ret_type, func, ...)                                 \
+     extern "C" ret_type func(__VA_ARGS__)                                     \
+       INTERCEPTOR_ATTRIBUTE __ATTRIBUTE_WEAK_WRAPPER ALIAS(WRAP(func));
+# endif  // ASM_INTERCEPTOR_TRAMPOLINE_SUPPORT
 #endif
 
 #if SANITIZER_FUCHSIA
@@ -162,10 +238,10 @@ const interpose_substitution substitution_##func_name[] \
 # define REAL(x) __interception::PTR_TO_REAL(x)
 # define FUNC_TYPE(x) x##_type
 
-# define DECLARE_REAL(ret_type, func, ...) \
+# define DECLARE_REAL(ret_type, func, ...)            \
     typedef ret_type (*FUNC_TYPE(func))(__VA_ARGS__); \
-    namespace __interception { \
-      extern FUNC_TYPE(func) PTR_TO_REAL(func); \
+    namespace __interception {                        \
+    extern FUNC_TYPE(func) PTR_TO_REAL(func);         \
     }
 # define ASSIGN_REAL(dst, src) REAL(dst) = REAL(src)
 #else  // SANITIZER_APPLE
@@ -176,14 +252,16 @@ const interpose_substitution substitution_##func_name[] \
 #endif  // SANITIZER_APPLE
 
 #if !SANITIZER_FUCHSIA
-#  define DECLARE_REAL_AND_INTERCEPTOR(ret_type, func, ...) \
+# define DECLARE_REAL_AND_INTERCEPTOR(ret_type, func, ...)  \
     DECLARE_REAL(ret_type, func, __VA_ARGS__)               \
+    extern "C" ret_type TRAMPOLINE(func)(__VA_ARGS__);      \
     extern "C" ret_type WRAP(func)(__VA_ARGS__);
 // Declare an interceptor and its wrapper defined in a different translation
 // unit (ex. asm).
-# define DECLARE_EXTERN_INTERCEPTOR_AND_WRAPPER(ret_type, func, ...)    \
-  extern "C" ret_type WRAP(func)(__VA_ARGS__); \
-  extern "C" ret_type func(__VA_ARGS__);
+# define DECLARE_EXTERN_INTERCEPTOR_AND_WRAPPER(ret_type, func, ...)  \
+    extern "C" ret_type TRAMPOLINE(func)(__VA_ARGS__);                \
+    extern "C" ret_type WRAP(func)(__VA_ARGS__);                      \
+    extern "C" ret_type func(__VA_ARGS__);
 #else
 # define DECLARE_REAL_AND_INTERCEPTOR(ret_type, func, ...)
 # define DECLARE_EXTERN_INTERCEPTOR_AND_WRAPPER(ret_type, func, ...)
@@ -215,12 +293,10 @@ const interpose_substitution substitution_##func_name[] \
 
 #elif !SANITIZER_APPLE
 
-#define INTERCEPTOR(ret_type, func, ...) \
-  DEFINE_REAL(ret_type, func, __VA_ARGS__) \
-  DECLARE_WRAPPER(ret_type, func, __VA_ARGS__) \
-  extern "C" \
-  INTERCEPTOR_ATTRIBUTE \
-  ret_type WRAP(func)(__VA_ARGS__)
+#define INTERCEPTOR(ret_type, func, ...)        \
+  DEFINE_REAL(ret_type, func, __VA_ARGS__)      \
+  DECLARE_WRAPPER(ret_type, func, __VA_ARGS__)  \
+  extern "C" INTERCEPTOR_ATTRIBUTE ret_type WRAP(func)(__VA_ARGS__)
 
 // We don't need INTERCEPTOR_WITH_SUFFIX on non-Darwin for now.
 #define INTERCEPTOR_WITH_SUFFIX(ret_type, func, ...) \
@@ -228,10 +304,10 @@ const interpose_substitution substitution_##func_name[] \
 
 #else  // SANITIZER_APPLE
 
-#define INTERCEPTOR_ZZZ(suffix, ret_type, func, ...) \
-  extern "C" ret_type func(__VA_ARGS__) suffix; \
-  extern "C" ret_type WRAP(func)(__VA_ARGS__); \
-  INTERPOSER(func); \
+#define INTERCEPTOR_ZZZ(suffix, ret_type, func, ...)  \
+  extern "C" ret_type func(__VA_ARGS__) suffix;       \
+  extern "C" ret_type WRAP(func)(__VA_ARGS__);        \
+  INTERPOSER(func);                                   \
   extern "C" INTERCEPTOR_ATTRIBUTE ret_type WRAP(func)(__VA_ARGS__)
 
 #define INTERCEPTOR(ret_type, func, ...) \
@@ -246,14 +322,12 @@ const interpose_substitution substitution_##func_name[] \
 #endif
 
 #if SANITIZER_WINDOWS
-# define INTERCEPTOR_WINAPI(ret_type, func, ...) \
+# define INTERCEPTOR_WINAPI(ret_type, func, ...)                \
     typedef ret_type (__stdcall *FUNC_TYPE(func))(__VA_ARGS__); \
-    namespace __interception { \
-      FUNC_TYPE(func) PTR_TO_REAL(func); \
-    } \
-    extern "C" \
-    INTERCEPTOR_ATTRIBUTE \
-    ret_type __stdcall WRAP(func)(__VA_ARGS__)
+    namespace __interception {                                  \
+      FUNC_TYPE(func) PTR_TO_REAL(func);                        \
+    }                                                           \
+    extern "C" INTERCEPTOR_ATTRIBUTE ret_type __stdcall WRAP(func)(__VA_ARGS__)
 #endif
 
 // ISO C++ forbids casting between pointer-to-function and pointer-to-object,
diff --git a/libsanitizer/interception/interception_linux.cpp b/libsanitizer/interception/interception_linux.cpp
index 5111a87..ef8136e 100644
--- a/libsanitizer/interception/interception_linux.cpp
+++ b/libsanitizer/interception/interception_linux.cpp
@@ -33,7 +33,7 @@ static int StrCmp(const char *s1, const char *s2) {
 }
 #endif
 
-static void *GetFuncAddr(const char *name, uptr wrapper_addr) {
+static void *GetFuncAddr(const char *name, uptr trampoline) {
 #if SANITIZER_NETBSD
   // FIXME: Find a better way to handle renames
   if (StrCmp(name, "sigaction"))
@@ -50,17 +50,17 @@ static void *GetFuncAddr(const char *name, uptr wrapper_addr) {
 
     // In case `name' is not loaded, dlsym ends up finding the actual wrapper.
     // We don't want to intercept the wrapper and have it point to itself.
-    if ((uptr)addr == wrapper_addr)
+    if ((uptr)addr == trampoline)
       addr = nullptr;
   }
   return addr;
 }
 
 bool InterceptFunction(const char *name, uptr *ptr_to_real, uptr func,
-                       uptr wrapper) {
-  void *addr = GetFuncAddr(name, wrapper);
+                       uptr trampoline) {
+  void *addr = GetFuncAddr(name, trampoline);
   *ptr_to_real = (uptr)addr;
-  return addr && (func == wrapper);
+  return addr && (func == trampoline);
 }
 
 // dlvsym is a GNU extension supported by some other platforms.
@@ -70,12 +70,12 @@ static void *GetFuncAddr(const char *name, const char *ver) {
 }
 
 bool InterceptFunction(const char *name, const char *ver, uptr *ptr_to_real,
-                       uptr func, uptr wrapper) {
+                       uptr func, uptr trampoline) {
   void *addr = GetFuncAddr(name, ver);
   *ptr_to_real = (uptr)addr;
-  return addr && (func == wrapper);
+  return addr && (func == trampoline);
 }
-#endif  // SANITIZER_GLIBC || SANITIZER_FREEBSD || SANITIZER_NETBSD
+#  endif  // SANITIZER_GLIBC || SANITIZER_FREEBSD || SANITIZER_NETBSD
 
 }  // namespace __interception
 
diff --git a/libsanitizer/interception/interception_linux.h b/libsanitizer/interception/interception_linux.h
index a08f8cb..433a3d9 100644
--- a/libsanitizer/interception/interception_linux.h
+++ b/libsanitizer/interception/interception_linux.h
@@ -15,7 +15,7 @@
     SANITIZER_SOLARIS
 
 #if !defined(INCLUDED_FROM_INTERCEPTION_LIB)
-# error "interception_linux.h should be included from interception library only"
+# error interception_linux.h should be included from interception library only
 #endif
 
 #ifndef INTERCEPTION_LINUX_H
@@ -23,26 +23,26 @@
 
 namespace __interception {
 bool InterceptFunction(const char *name, uptr *ptr_to_real, uptr func,
-                       uptr wrapper);
+                       uptr trampoline);
 bool InterceptFunction(const char *name, const char *ver, uptr *ptr_to_real,
-                       uptr func, uptr wrapper);
+                       uptr func, uptr trampoline);
 }  // namespace __interception
 
 #define INTERCEPT_FUNCTION_LINUX_OR_FREEBSD(func) \
   ::__interception::InterceptFunction(            \
       #func,                                      \
-      (::__interception::uptr *) & REAL(func),    \
-      (::__interception::uptr) & (func),          \
-      (::__interception::uptr) & WRAP(func))
+      (::__interception::uptr *)&REAL(func),      \
+      (::__interception::uptr)&(func),            \
+      (::__interception::uptr)&TRAMPOLINE(func))
 
 // dlvsym is a GNU extension supported by some other platforms.
 #if SANITIZER_GLIBC || SANITIZER_FREEBSD || SANITIZER_NETBSD
 #define INTERCEPT_FUNCTION_VER_LINUX_OR_FREEBSD(func, symver) \
   ::__interception::InterceptFunction(                        \
       #func, symver,                                          \
-      (::__interception::uptr *) & REAL(func),                \
-      (::__interception::uptr) & (func),                      \
-      (::__interception::uptr) & WRAP(func))
+      (::__interception::uptr *)&REAL(func),                  \
+      (::__interception::uptr)&(func),                        \
+      (::__interception::uptr)&TRAMPOLINE(func))
 #else
 #define INTERCEPT_FUNCTION_VER_LINUX_OR_FREEBSD(func, symver) \
   INTERCEPT_FUNCTION_LINUX_OR_FREEBSD(func)
diff --git a/libsanitizer/interception/interception_win.cpp b/libsanitizer/interception/interception_win.cpp
index faaa8ee..1b681ad 100644
--- a/libsanitizer/interception/interception_win.cpp
+++ b/libsanitizer/interception/interception_win.cpp
@@ -1,4 +1,4 @@
-//===-- interception_linux.cpp ----------------------------------*- C++ -*-===//
+//===-- interception_win.cpp ------------------------------------*- C++ -*-===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
@@ -141,8 +141,29 @@ static const int kBranchLength =
     FIRST_32_SECOND_64(kJumpInstructionLength, kIndirectJumpInstructionLength);
 static const int kDirectBranchLength = kBranchLength + kAddressLength;
 
+#  if defined(_MSC_VER)
+#    define INTERCEPTION_FORMAT(f, a)
+#  else
+#    define INTERCEPTION_FORMAT(f, a) __attribute__((format(printf, f, a)))
+#  endif
+
+static void (*ErrorReportCallback)(const char *format, ...)
+    INTERCEPTION_FORMAT(1, 2);
+
+void SetErrorReportCallback(void (*callback)(const char *format, ...)) {
+  ErrorReportCallback = callback;
+}
+
+#  define ReportError(...)                \
+    do {                                  \
+      if (ErrorReportCallback)            \
+        ErrorReportCallback(__VA_ARGS__); \
+    } while (0)
+
 static void InterceptionFailed() {
-  // Do we have a good way to abort with an error message here?
+  ReportError("interception_win: failed due to an unrecoverable error.\n");
+  // This acts like an abort when no debugger is attached. According to an old
+  // comment, calling abort() leads to an infinite recursion in CheckFailed.
   __debugbreak();
 }
 
@@ -249,8 +270,13 @@ static void WritePadding(uptr from, uptr size) {
 }
 
 static void WriteJumpInstruction(uptr from, uptr target) {
-  if (!DistanceIsWithin2Gig(from + kJumpInstructionLength, target))
+  if (!DistanceIsWithin2Gig(from + kJumpInstructionLength, target)) {
+    ReportError(
+        "interception_win: cannot write jmp further than 2GB away, from %p to "
+        "%p.\n",
+        (void *)from, (void *)target);
     InterceptionFailed();
+  }
   ptrdiff_t offset = target - from - kJumpInstructionLength;
   *(u8*)from = 0xE9;
   *(u32*)(from + 1) = offset;
@@ -274,6 +300,10 @@ static void WriteIndirectJumpInstruction(uptr from, uptr indirect_target) {
   int offset = indirect_target - from - kIndirectJumpInstructionLength;
   if (!DistanceIsWithin2Gig(from + kIndirectJumpInstructionLength,
                             indirect_target)) {
+    ReportError(
+        "interception_win: cannot write indirect jmp with target further than "
+        "2GB away, from %p to %p.\n",
+        (void *)from, (void *)indirect_target);
     InterceptionFailed();
   }
   *(u16*)from = 0x25FF;
@@ -427,6 +457,11 @@ static const u8 kPrologueWithShortJump2[] = {
 
 // Returns 0 on error.
 static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
+#if SANITIZER_ARM64
+  // An ARM64 instruction is 4 bytes long.
+  return 4;
+#endif
+
 #if SANITIZER_WINDOWS64
   if (memcmp((u8*)address, kPrologueWithShortJump1,
              sizeof(kPrologueWithShortJump1)) == 0 ||
@@ -492,6 +527,7 @@ static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
     case 0xFF8B:  // 8B FF : mov edi, edi
     case 0xEC8B:  // 8B EC : mov ebp, esp
     case 0xc889:  // 89 C8 : mov eax, ecx
+    case 0xE589:  // 89 E5 : mov ebp, esp
     case 0xC18B:  // 8B C1 : mov eax, ecx
     case 0xC033:  // 33 C0 : xor eax, eax
     case 0xC933:  // 33 C9 : xor ecx, ecx
@@ -588,7 +624,7 @@ static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
                       //   mov rax, QWORD PTR [rip + XXXXXXXX]
     case 0x25ff48:    // 48 ff 25 XX XX XX XX :
                       //   rex.W jmp QWORD PTR [rip + XXXXXXXX]
-
+    case 0x158D4C:    // 4c 8d 15 XX XX XX XX : lea r10, [rip + XX]
       // Instructions having offset relative to 'rip' need offset adjustment.
       if (rel_offset)
         *rel_offset = 3;
@@ -641,6 +677,8 @@ static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
     case 0x24448B:  // 8B 44 24 XX : mov eax, dword ptr [esp + XX]
     case 0x244C8B:  // 8B 4C 24 XX : mov ecx, dword ptr [esp + XX]
     case 0x24548B:  // 8B 54 24 XX : mov edx, dword ptr [esp + XX]
+    case 0x245C8B:  // 8B 5C 24 XX : mov ebx, dword ptr [esp + XX]
+    case 0x246C8B:  // 8B 6C 24 XX : mov ebp, dword ptr [esp + XX]
     case 0x24748B:  // 8B 74 24 XX : mov esi, dword ptr [esp + XX]
     case 0x247C8B:  // 8B 7C 24 XX : mov edi, dword ptr [esp + XX]
       return 4;
@@ -652,12 +690,20 @@ static size_t GetInstructionSize(uptr address, size_t* rel_offset = nullptr) {
   }
 #endif
 
-  // Unknown instruction!
-  // FIXME: Unknown instruction failures might happen when we add a new
-  // interceptor or a new compiler version. In either case, they should result
-  // in visible and readable error messages. However, merely calling abort()
-  // leads to an infinite recursion in CheckFailed.
-  InterceptionFailed();
+  // Unknown instruction! This might happen when we add a new interceptor, use
+  // a new compiler version, or if Windows changed how some functions are
+  // compiled. In either case, we print the address and 8 bytes of instructions
+  // to notify the user about the error and to help identify the unknown
+  // instruction. Don't treat this as a fatal error, though we can break the
+  // debugger if one has been attached.
+  u8 *bytes = (u8 *)address;
+  ReportError(
+      "interception_win: unhandled instruction at %p: %02x %02x %02x %02x %02x "
+      "%02x %02x %02x\n",
+      (void *)address, bytes[0], bytes[1], bytes[2], bytes[3], bytes[4],
+      bytes[5], bytes[6], bytes[7]);
+  if (::IsDebuggerPresent())
+    __debugbreak();
   return 0;
 }
 
@@ -678,16 +724,24 @@ static bool CopyInstructions(uptr to, uptr from, size_t size) {
   while (cursor != size) {
     size_t rel_offset = 0;
     size_t instruction_size = GetInstructionSize(from + cursor, &rel_offset);
-    _memcpy((void*)(to + cursor), (void*)(from + cursor),
+    if (!instruction_size)
+      return false;
+    _memcpy((void *)(to + cursor), (void *)(from + cursor),
             (size_t)instruction_size);
     if (rel_offset) {
-      uptr delta = to - from;
-      uptr relocated_offset = *(u32*)(to + cursor + rel_offset) - delta;
-#if SANITIZER_WINDOWS64
-      if (relocated_offset + 0x80000000U >= 0xFFFFFFFFU)
+#  if SANITIZER_WINDOWS64
+      // we want to make sure that the new relative offset still fits in 32-bits
+      // this will be untrue if relocated_offset \notin [-2**31, 2**31)
+      s64 delta = to - from;
+      s64 relocated_offset = *(s32 *)(to + cursor + rel_offset) - delta;
+      if (-0x8000'0000ll > relocated_offset || relocated_offset > 0x7FFF'FFFFll)
         return false;
-#endif
-      *(u32*)(to + cursor + rel_offset) = relocated_offset;
+#  else
+      // on 32-bit, the relative offset will always be correct
+      s32 delta = to - from;
+      s32 relocated_offset = *(s32 *)(to + cursor + rel_offset) - delta;
+#  endif
+      *(s32 *)(to + cursor + rel_offset) = relocated_offset;
     }
     cursor += instruction_size;
   }
@@ -895,6 +949,10 @@ static void **InterestingDLLsAvailable() {
       "msvcr120.dll",      // VS2013
       "vcruntime140.dll",  // VS2015
       "ucrtbase.dll",      // Universal CRT
+#if (defined(__MINGW32__) && defined(__i386__))
+      "libc++.dll",        // libc++
+      "libunwind.dll",     // libunwind
+#endif
       // NTDLL should go last as it exports some functions that we should
       // override in the CRT [presumably only used internally].
       "ntdll.dll", NULL};
diff --git a/libsanitizer/interception/interception_win.h b/libsanitizer/interception/interception_win.h
index 45900130..f6eca82 100644
--- a/libsanitizer/interception/interception_win.h
+++ b/libsanitizer/interception/interception_win.h
@@ -41,6 +41,11 @@ bool OverrideImportedFunction(const char *module_to_patch,
                               const char *function_name, uptr new_function,
                               uptr *orig_old_func);
 
+// Sets a callback to be used for reporting errors by interception_win. The
+// callback will be called with printf-like arguments. Intended to be used with
+// __sanitizer::Report. Pass nullptr to disable error reporting (default).
+void SetErrorReportCallback(void (*callback)(const char *format, ...));
+
 #if !SANITIZER_WINDOWS64
 // Exposed for unittests
 bool OverrideFunctionWithDetour(