[asan] Implement interception on AIX #131870

jakeegan · 2025-03-18T17:41:22Z

Adjust asan interceptor support for AIX. AIX uses dlsym to retrieve addresses of exported functions. However, some functions in libc.a, such as memcpy, are not exported, so we currently have a limitation in retrieving these addresses.

Issue: #138916

github-actions · 2025-03-18T17:48:15Z

✅ With the latest revision this PR passed the C/C++ code formatter.

This reverts commit 549b134.

llvmbot · 2025-03-24T17:42:03Z

@llvm/pr-subscribers-compiler-rt-sanitizer

Author: Jake Egan (jakeegan)

Changes

Builds on the effort to support AIX in sanitizer_common/asan. Intercepting functions in shared libraries is not supported on AIX yet.

Previous PR: #131868

Patch is 20.14 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/131870.diff

9 Files Affected:

(modified) compiler-rt/lib/interception/CMakeLists.txt (+2)
(modified) compiler-rt/lib/interception/interception.h (+17-2)
(added) compiler-rt/lib/interception/interception_aix.cpp (+45)
(added) compiler-rt/lib/interception/interception_aix.h (+36)
(modified) compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc (+42-12)
(modified) compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_ioctl.inc (+2)
(modified) compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_memintrinsics.inc (+2)
(modified) compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h (+33-19)
(modified) compiler-rt/lib/sanitizer_common/sanitizer_redefine_builtins.h (+1-1)

diff --git a/compiler-rt/lib/interception/CMakeLists.txt b/compiler-rt/lib/interception/CMakeLists.txt
index fe7fa27fbc78b..57c8c8ba20141 100644
--- a/compiler-rt/lib/interception/CMakeLists.txt
+++ b/compiler-rt/lib/interception/CMakeLists.txt
@@ -1,6 +1,7 @@
 # Build for the runtime interception helper library.
 
 set(INTERCEPTION_SOURCES
+  interception_aix.cpp
   interception_linux.cpp
   interception_mac.cpp
   interception_win.cpp
@@ -9,6 +10,7 @@ set(INTERCEPTION_SOURCES
 
 set(INTERCEPTION_HEADERS
   interception.h
+  interception_aix.h
   interception_linux.h
   interception_mac.h
   interception_win.h
diff --git a/compiler-rt/lib/interception/interception.h b/compiler-rt/lib/interception/interception.h
index 3cb6b446638e0..3a20cfb04dcd6 100644
--- a/compiler-rt/lib/interception/interception.h
+++ b/compiler-rt/lib/interception/interception.h
@@ -19,7 +19,7 @@
 
 #if !SANITIZER_LINUX && !SANITIZER_FREEBSD && !SANITIZER_APPLE &&    \
     !SANITIZER_NETBSD && !SANITIZER_WINDOWS && !SANITIZER_FUCHSIA && \
-    !SANITIZER_SOLARIS
+    !SANITIZER_SOLARIS && !SANITIZER_AIX
 #  error "Interception doesn't work on this operating system."
 #endif
 
@@ -168,6 +168,16 @@ const interpose_substitution substitution_##func_name[]             \
     extern "C" ret_type func(__VA_ARGS__);
 # define DECLARE_WRAPPER_WINAPI(ret_type, func, ...)  \
     extern "C" __declspec(dllimport) ret_type __stdcall func(__VA_ARGS__);
+#elif SANITIZER_AIX
+#  define WRAP(x) __interceptor_##x
+#  define TRAMPOLINE(x) WRAP(x)
+// # define WRAPPER_NAME(x) "__interceptor_" #x
+#  define INTERCEPTOR_ATTRIBUTE __attribute__((visibility("default")))
+// AIX's linker will not select the weak symbol, so don't use weak for the
+// interceptors.
+#  define DECLARE_WRAPPER(ret_type, func, ...) \
+    extern "C" ret_type func(__VA_ARGS__)      \
+        __attribute__((alias("__interceptor_" #func), visibility("default")));
 #elif !SANITIZER_FUCHSIA  // LINUX, FREEBSD, NETBSD, SOLARIS
 # define INTERCEPTOR_ATTRIBUTE __attribute__((visibility("default")))
 # if ASM_INTERCEPTOR_TRAMPOLINE_SUPPORT
@@ -367,7 +377,12 @@ inline void DoesNotSupportStaticLinking() {}
 
 #define INCLUDED_FROM_INTERCEPTION_LIB
 
-#if SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD || \
+#if SANITIZER_AIX
+# include "interception_aix.h"
+# define INTERCEPT_FUNCTION(func) INTERCEPT_FUNCTION_AIX(func)
+# define INTERCEPT_FUNCTION_VER(func, symver) INTERCEPT_FUNCTION_AIX(func)
+
+#elif SANITIZER_LINUX || SANITIZER_FREEBSD || SANITIZER_NETBSD || \
     SANITIZER_SOLARIS
 
 # include "interception_linux.h"
diff --git a/compiler-rt/lib/interception/interception_aix.cpp b/compiler-rt/lib/interception/interception_aix.cpp
new file mode 100644
index 0000000000000..953bbad96eb47
--- /dev/null
+++ b/compiler-rt/lib/interception/interception_aix.cpp
@@ -0,0 +1,45 @@
+//===-- interception_aix.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.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of AddressSanitizer, an address sanity checker.
+//
+// AIX-specific interception methods.
+//===----------------------------------------------------------------------===//
+
+#include "interception.h"
+#include "sanitizer_common/sanitizer_common.h"
+
+#if SANITIZER_AIX
+
+#  include <dlfcn.h>  // for dlsym()
+
+namespace __interception {
+
+static void *GetFuncAddr(const char *name, uptr wrapper_addr) {
+  // AIX dlsym can only defect the functions that are exported, so
+  // on AIX, we can not intercept some basic functions like memcpy.
+  // FIXME: if we are going to ship dynamic asan library, we may need to search
+  // all the loaded modules with RTLD_DEFAULT if RTLD_NEXT failed.
+  void *addr = dlsym(RTLD_NEXT, name);
+
+  // 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)
+    addr = nullptr;
+  return addr;
+}
+
+bool InterceptFunction(const char *name, uptr *ptr_to_real, uptr func,
+                       uptr wrapper) {
+  void *addr = GetFuncAddr(name, wrapper);
+  *ptr_to_real = (uptr)addr;
+  return addr && (func == wrapper);
+}
+
+}  // namespace __interception
+#endif  // SANITIZER_AIX
diff --git a/compiler-rt/lib/interception/interception_aix.h b/compiler-rt/lib/interception/interception_aix.h
new file mode 100644
index 0000000000000..08d17caeb6a8d
--- /dev/null
+++ b/compiler-rt/lib/interception/interception_aix.h
@@ -0,0 +1,36 @@
+//===-- interception_aix.h --------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of AddressSanitizer, an address sanity checker.
+//
+// AIX-specific interception methods.
+//===----------------------------------------------------------------------===//
+
+#if SANITIZER_AIX
+
+#  if !defined(INCLUDED_FROM_INTERCEPTION_LIB)
+#    error \
+        "interception_aix.h should be included from interception library only"
+#  endif
+
+#  ifndef INTERCEPTION_AIX_H
+#    define INTERCEPTION_AIX_H
+
+namespace __interception {
+bool InterceptFunction(const char *name, uptr *ptr_to_real, uptr func,
+                       uptr wrapper);
+}  // namespace __interception
+
+#    define INTERCEPT_FUNCTION_AIX(func)                \
+      ::__interception::InterceptFunction(              \
+          #func, (::__interception::uptr *)&REAL(func), \
+          (::__interception::uptr) & (func),            \
+          (::__interception::uptr)&WRAP(func))
+
+#  endif  // INTERCEPTION_AIX_H
+#endif    // SANITIZER_AIX
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc
index 24a8a2d4dc55b..35876084f8107 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc
@@ -480,7 +480,7 @@ INTERCEPTOR(char*, textdomain, const char *domainname) {
 #define INIT_TEXTDOMAIN
 #endif
 
-#if SANITIZER_INTERCEPT_STRCMP || SANITIZER_INTERCEPT_MEMCMP
+#if SANITIZER_INTERCEPT_MEMCMP
 static inline int CharCmpX(unsigned char c1, unsigned char c2) {
   return (c1 == c2) ? 0 : (c1 < c2) ? -1 : 1;
 }
@@ -953,7 +953,7 @@ INTERCEPTOR(double, frexp, double x, int *exp) {
 #define INIT_FREXP
 #endif  // SANITIZER_INTERCEPT_FREXP
 
-#if SANITIZER_INTERCEPT_FREXPF_FREXPL
+#if SANITIZER_INTERCEPT_FREXPF
 INTERCEPTOR(float, frexpf, float x, int *exp) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, frexpf, x, exp);
@@ -963,6 +963,12 @@ INTERCEPTOR(float, frexpf, float x, int *exp) {
   return res;
 }
 
+#define INIT_FREXPF COMMON_INTERCEPT_FUNCTION(frexpf);
+#else
+#define INIT_FREXPF
+#endif
+
+#if SANITIZER_INTERCEPT_FREXPL
 INTERCEPTOR(long double, frexpl, long double x, int *exp) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, frexpl, x, exp);
@@ -972,12 +978,10 @@ INTERCEPTOR(long double, frexpl, long double x, int *exp) {
   return res;
 }
 
-#define INIT_FREXPF_FREXPL           \
-  COMMON_INTERCEPT_FUNCTION(frexpf); \
-  COMMON_INTERCEPT_FUNCTION_LDBL(frexpl)
+#define INIT_FREXPL COMMON_INTERCEPT_FUNCTION_LDBL(frexpl)
 #else
-#define INIT_FREXPF_FREXPL
-#endif  // SANITIZER_INTERCEPT_FREXPF_FREXPL
+#define INIT_FREXPL
+#endif
 
 #if SI_POSIX
 static void write_iovec(void *ctx, struct __sanitizer_iovec *iovec,
@@ -1346,7 +1350,8 @@ INTERCEPTOR(unsigned long, time, unsigned long *t) {
 #if SANITIZER_INTERCEPT_LOCALTIME_AND_FRIENDS
 static void unpoison_tm(void *ctx, __sanitizer_tm *tm) {
   COMMON_INTERCEPTOR_WRITE_RANGE(ctx, tm, sizeof(*tm));
-#if !SANITIZER_SOLARIS
+// AIX tm struct does not have tm_zone field.
+#if !SANITIZER_SOLARIS && !SANITIZER_AIX
   if (tm->tm_zone) {
     // Can not use COMMON_INTERCEPTOR_WRITE_RANGE here, because tm->tm_zone
     // can point to shared memory and tsan would report a data race.
@@ -1731,8 +1736,10 @@ INTERCEPTOR(int, __vsprintf_chk, char *str, int flag, SIZE_T size_to,
 VSPRINTF_INTERCEPTOR_IMPL(vsprintf, str, format, ap)
 #endif
 
+#if SANITIZER_INTERCEPT_VASPRINTF
 INTERCEPTOR(int, vasprintf, char **strp, const char *format, va_list ap)
 VASPRINTF_INTERCEPTOR_IMPL(vasprintf, strp, format, ap)
+#endif
 
 #if SANITIZER_INTERCEPT_ISOC99_PRINTF
 INTERCEPTOR(int, __isoc99_vprintf, const char *format, va_list ap)
@@ -1783,8 +1790,10 @@ INTERCEPTOR(int, __snprintf_chk, char *str, SIZE_T size, int flag,
 FORMAT_INTERCEPTOR_IMPL(__snprintf_chk, vsnprintf, str, size, format)
 #endif
 
+#if SANITIZER_INTERCEPT_ASPRINTF
 INTERCEPTOR(int, asprintf, char **strp, const char *format, ...)
 FORMAT_INTERCEPTOR_IMPL(asprintf, vasprintf, strp, format)
+#endif
 
 #if SANITIZER_INTERCEPT_ISOC99_PRINTF
 INTERCEPTOR(int, __isoc99_printf, const char *format, ...)
@@ -1807,17 +1816,29 @@ FORMAT_INTERCEPTOR_IMPL(__isoc99_snprintf, __isoc99_vsnprintf, str, size,
 #endif  // SANITIZER_INTERCEPT_PRINTF
 
 #if SANITIZER_INTERCEPT_PRINTF
+#if SANITIZER_AIX
+#define INIT_PRINTF                     \
+  COMMON_INTERCEPT_FUNCTION_LDBL(printf);    \
+  COMMON_INTERCEPT_FUNCTION_LDBL(sprintf);   \
+  COMMON_INTERCEPT_FUNCTION_LDBL(snprintf);  \
+  COMMON_INTERCEPT_FUNCTION_LDBL(fprintf);   \
+  COMMON_INTERCEPT_FUNCTION_LDBL(vprintf);   \
+  COMMON_INTERCEPT_FUNCTION_LDBL(vsprintf);  \
+  COMMON_INTERCEPT_FUNCTION_LDBL(vsnprintf); \
+  COMMON_INTERCEPT_FUNCTION_LDBL(vfprintf);
+#else
 #define INIT_PRINTF                     \
   COMMON_INTERCEPT_FUNCTION_LDBL(printf);    \
   COMMON_INTERCEPT_FUNCTION_LDBL(sprintf);   \
   COMMON_INTERCEPT_FUNCTION_LDBL(snprintf);  \
-  COMMON_INTERCEPT_FUNCTION_LDBL(asprintf);  \
   COMMON_INTERCEPT_FUNCTION_LDBL(fprintf);   \
   COMMON_INTERCEPT_FUNCTION_LDBL(vprintf);   \
   COMMON_INTERCEPT_FUNCTION_LDBL(vsprintf);  \
   COMMON_INTERCEPT_FUNCTION_LDBL(vsnprintf); \
+  COMMON_INTERCEPT_FUNCTION_LDBL(asprintf);  \
   COMMON_INTERCEPT_FUNCTION_LDBL(vasprintf); \
   COMMON_INTERCEPT_FUNCTION_LDBL(vfprintf);
+#endif
 #else
 #define INIT_PRINTF
 #endif
@@ -3901,7 +3922,11 @@ INTERCEPTOR(SIZE_T, wcrtomb, char *dest, wchar_t src, void *ps) {
   if (res != ((SIZE_T)-1)) {
     CHECK_LE(res, sizeof(local_dest));
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dest, res);
-    REAL(memcpy)(dest, local_dest, res);
+#if !SANITIZER_AIX
+      REAL(memcpy)(dest, local_dest, res);
+#else
+      internal_memcpy(dest, local_dest, res);
+#endif
   }
   return res;
 }
@@ -3923,7 +3948,11 @@ INTERCEPTOR(int, wctomb, char *dest, wchar_t src) {
   if (res != -1) {
     CHECK_LE(res, sizeof(local_dest));
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dest, res);
-    REAL(memcpy)(dest, local_dest, res);
+#if !SANITIZER_AIX
+      REAL(memcpy)(dest, local_dest, res);
+#else
+      internal_memcpy(dest, local_dest, res);
+#endif
   }
   return res;
 }
@@ -10329,7 +10358,8 @@ static void InitializeCommonInterceptors() {
   INIT_PRINTF_L;
   INIT_ISOC99_PRINTF;
   INIT_FREXP;
-  INIT_FREXPF_FREXPL;
+  INIT_FREXPF;
+  INIT_FREXPL;
   INIT_GETPWNAM_AND_FRIENDS;
   INIT_GETPWNAM_R_AND_FRIENDS;
   INIT_GETPWENT;
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_ioctl.inc b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_ioctl.inc
index 49ec4097c900b..5b23039954a43 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_ioctl.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_ioctl.inc
@@ -73,7 +73,9 @@ static void ioctl_table_fill() {
   _(TIOCNXCL, NONE, 0);
   _(TIOCOUTQ, WRITE, sizeof(int));
   _(TIOCPKT, READ, sizeof(int));
+#if !SANITIZER_AIX
   _(TIOCSCTTY, NONE, 0);
+#endif
   _(TIOCSETD, READ, sizeof(int));
   _(TIOCSPGRP, READ, pid_t_sz);
   _(TIOCSTI, READ, sizeof(char));
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_memintrinsics.inc b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_memintrinsics.inc
index 1565a494140f6..e542a6c7b438a 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_memintrinsics.inc
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors_memintrinsics.inc
@@ -36,6 +36,8 @@
 #define PLATFORM_HAS_DIFFERENT_MEMCPY_AND_MEMMOVE 0
 #elif SANITIZER_WINDOWS64
 #define PLATFORM_HAS_DIFFERENT_MEMCPY_AND_MEMMOVE 0
+#elif SANITIZER_AIX
+#define PLATFORM_HAS_DIFFERENT_MEMCPY_AND_MEMMOVE 0
 #else
 #define PLATFORM_HAS_DIFFERENT_MEMCPY_AND_MEMMOVE 1
 #endif  // SANITIZER_APPLE
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h b/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h
index febd233bb1e3c..43a37037c971d 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_platform_interceptors.h
@@ -139,6 +139,12 @@ SANITIZER_WEAK_IMPORT void *aligned_alloc(__sanitizer::usize __alignment,
 #define SI_SOLARIS 0
 #endif
 
+#if SANITIZER_AIX
+#define SI_NOT_AIX 0
+#else
+#define SI_NOT_AIX 1
+#endif
+
 #if SANITIZER_SOLARIS32
 #define SI_SOLARIS32 1
 #else
@@ -159,20 +165,20 @@ SANITIZER_WEAK_IMPORT void *aligned_alloc(__sanitizer::usize __alignment,
 
 #define SANITIZER_INTERCEPT_STRLEN SI_NOT_FUCHSIA
 #define SANITIZER_INTERCEPT_STRNLEN (SI_NOT_MAC && SI_NOT_FUCHSIA)
-#define SANITIZER_INTERCEPT_STRCMP SI_NOT_FUCHSIA
+#define SANITIZER_INTERCEPT_STRCMP (SI_NOT_FUCHSIA && SI_NOT_AIX)
 #define SANITIZER_INTERCEPT_STRSTR SI_NOT_FUCHSIA
-#define SANITIZER_INTERCEPT_STRCASESTR SI_POSIX
+#define SANITIZER_INTERCEPT_STRCASESTR (SI_POSIX && SI_NOT_AIX)
 #define SANITIZER_INTERCEPT_STRTOK SI_NOT_FUCHSIA
 #define SANITIZER_INTERCEPT_STRCHR SI_NOT_FUCHSIA
-#define SANITIZER_INTERCEPT_STRCHRNUL SI_POSIX_NOT_MAC
+#define SANITIZER_INTERCEPT_STRCHRNUL (SI_POSIX_NOT_MAC && SI_NOT_AIX)
 #define SANITIZER_INTERCEPT_STRRCHR SI_NOT_FUCHSIA
 #define SANITIZER_INTERCEPT_STRSPN SI_NOT_FUCHSIA
 #define SANITIZER_INTERCEPT_STRPBRK SI_NOT_FUCHSIA
 #define SANITIZER_INTERCEPT_TEXTDOMAIN SI_LINUX_NOT_ANDROID || SI_SOLARIS
 #define SANITIZER_INTERCEPT_STRCASECMP SI_POSIX
 #define SANITIZER_INTERCEPT_MEMSET 1
-#define SANITIZER_INTERCEPT_MEMMOVE 1
-#define SANITIZER_INTERCEPT_MEMCPY 1
+#define SANITIZER_INTERCEPT_MEMMOVE SI_NOT_AIX
+#define SANITIZER_INTERCEPT_MEMCPY SI_NOT_AIX
 #define SANITIZER_INTERCEPT_MEMCMP SI_NOT_FUCHSIA
 #define SANITIZER_INTERCEPT_BCMP \
   SANITIZER_INTERCEPT_MEMCMP &&  \
@@ -231,6 +237,8 @@ SANITIZER_WEAK_IMPORT void *aligned_alloc(__sanitizer::usize __alignment,
 #define SANITIZER_INTERCEPT_ISOC99_SCANF SI_GLIBC
 
 #ifndef SANITIZER_INTERCEPT_PRINTF
+#define SANITIZER_INTERCEPT_ASPRINTF SI_NOT_AIX
+#define SANITIZER_INTERCEPT_VASPRINTF SI_NOT_AIX
 #define SANITIZER_INTERCEPT_PRINTF SI_POSIX
 #define SANITIZER_INTERCEPT_PRINTF_L (SI_FREEBSD || SI_NETBSD)
 #define SANITIZER_INTERCEPT_ISOC99_PRINTF SI_GLIBC
@@ -239,8 +247,10 @@ SANITIZER_WEAK_IMPORT void *aligned_alloc(__sanitizer::usize __alignment,
 #define SANITIZER_INTERCEPT___PRINTF_CHK \
   (SANITIZER_INTERCEPT_PRINTF && SI_GLIBC)
 
-#define SANITIZER_INTERCEPT_FREXP SI_NOT_FUCHSIA
-#define SANITIZER_INTERCEPT_FREXPF_FREXPL SI_POSIX
+// AIX libc does not export FREXP and FREXP.
+#define SANITIZER_INTERCEPT_FREXP (SI_NOT_FUCHSIA && SI_NOT_AIX)
+#define SANITIZER_INTERCEPT_FREXPF (SI_POSIX && SI_NOT_AIX)
+#define SANITIZER_INTERCEPT_FREXPL SI_POSIX
 
 #define SANITIZER_INTERCEPT_GETPWNAM_AND_FRIENDS SI_POSIX
 #define SANITIZER_INTERCEPT_GETPWNAM_R_AND_FRIENDS \
@@ -289,7 +299,7 @@ SANITIZER_WEAK_IMPORT void *aligned_alloc(__sanitizer::usize __alignment,
 #define SANITIZER_INTERCEPT_ACCEPT4 \
   (SI_LINUX_NOT_ANDROID || SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_PACCEPT SI_NETBSD
-#define SANITIZER_INTERCEPT_MODF SI_POSIX
+#define SANITIZER_INTERCEPT_MODF (SI_POSIX && SI_NOT_AIX)
 #define SANITIZER_INTERCEPT_RECVMSG SI_POSIX
 #define SANITIZER_INTERCEPT_SENDMSG SI_POSIX
 #define SANITIZER_INTERCEPT_RECVMMSG SI_LINUX
@@ -324,8 +334,9 @@ SANITIZER_WEAK_IMPORT void *aligned_alloc(__sanitizer::usize __alignment,
 #define SANITIZER_INTERCEPT___WCSXFRM_L SI_LINUX
 #define SANITIZER_INTERCEPT_WCSNRTOMBS \
   (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
-#define SANITIZER_INTERCEPT_WCRTOMB \
-  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
+#define SANITIZER_INTERCEPT_WCRTOMB                                           \
+  (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS || \
+   !SI_NOT_AIX)
 #define SANITIZER_INTERCEPT_WCTOMB \
   (SI_FREEBSD || SI_NETBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_TCGETATTR SI_LINUX_NOT_ANDROID || SI_SOLARIS
@@ -365,7 +376,8 @@ SANITIZER_WEAK_IMPORT void *aligned_alloc(__sanitizer::usize __alignment,
 #define SANITIZER_INTERCEPT_GETMNTENT_R SI_LINUX_NOT_ANDROID
 #define SANITIZER_INTERCEPT_STATFS \
   (SI_FREEBSD || SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
-#define SANITIZER_INTERCEPT_STATFS64 SI_GLIBC && SANITIZER_HAS_STATFS64
+#define SANITIZER_INTERCEPT_STATFS64 \
+  ((SI_GLIBC || !SI_NOT_AIX) && SANITIZER_HAS_STATFS64)
 #define SANITIZER_INTERCEPT_STATVFS \
   (SI_FREEBSD || SI_NETBSD || SI_LINUX_NOT_ANDROID)
 #define SANITIZER_INTERCEPT_STATVFS64 SI_GLIBC
@@ -414,10 +426,10 @@ SANITIZER_WEAK_IMPORT void *aligned_alloc(__sanitizer::usize __alignment,
 #define SANITIZER_INTERCEPT_TTYNAME_R SI_POSIX
 #define SANITIZER_INTERCEPT_TEMPNAM SI_POSIX
 #define SANITIZER_INTERCEPT_SINCOS SI_LINUX || SI_SOLARIS
-#define SANITIZER_INTERCEPT_REMQUO SI_POSIX
-#define SANITIZER_INTERCEPT_REMQUOL (SI_POSIX && !SI_NETBSD)
-#define SANITIZER_INTERCEPT_LGAMMA SI_POSIX
-#define SANITIZER_INTERCEPT_LGAMMAL (SI_POSIX && !SI_NETBSD)
+#define SANITIZER_INTERCEPT_REMQUO (SI_POSIX && SI_NOT_AIX)
+#define SANITIZER_INTERCEPT_REMQUOL (SI_POSIX && !SI_NETBSD && SI_NOT_AIX)
+#define SANITIZER_INTERCEPT_LGAMMA (SI_POSIX && SI_NOT_AIX)
+#define SANITIZER_INTERCEPT_LGAMMAL (SI_POSIX && !SI_NETBSD && SI_NOT_AIX)
 #define SANITIZER_INTERCEPT_LGAMMA_R (SI_FREEBSD || SI_LINUX || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_LGAMMAL_R SI_LINUX_NOT_ANDROID || SI_SOLARIS
 #define SANITIZER_INTERCEPT_DRAND48_R SI_GLIBC
@@ -502,9 +514,11 @@ SANITIZER_WEAK_IMPORT void *aligned_alloc(__sanitizer::usize __alignment,
 #define SI_STAT_LINUX (SI_LINUX && __GLIBC_PREREQ(2, 33))
 #define SANITIZER_INTERCEPT_STAT                                        \
   (SI_FREEBSD || SI_MAC || SI_ANDROID || SI_NETBSD || SI_SOLARIS ||     \
-   SI_STAT_LINUX)
-#define SANITIZER_INTERCEPT_STAT64 SI_STAT_LINUX && SANITIZER_HAS_STAT64
-#define SANITIZER_INTERCEPT_LSTAT (SI_NETBSD || SI_FREEBSD || SI_STAT_LINUX)
+   SI_STAT_LINUX || !SI_NOT_AIX)
+#define SANITIZER_INTERCEPT_STAT64 \
+  ((SI_STAT_LINUX || !SI_NOT_AIX) && SANITIZER_HAS_STAT64)
+#define SANITIZER_INTERCEPT_LSTAT \
+  (SI_NETBSD || SI_FREEBSD || SI_STAT_LINUX || !SI_NOT_AIX)
 #define SANITIZER_INTERCEPT___XSTAT \
   ((!SANITIZER_INTERCEPT_STAT && SI_POSIX) || SI_STAT_LINUX)
 #define SANITIZER_INTERCEPT___XSTAT64 SI_GLIBC
@@ -572,7 +586,7 @@ SANITIZER_WEAK_IMPORT void *aligned_alloc(__sanitizer::usize __alignment,
 #define SANITIZER_INTERCEPT_PROTOENT_R SI_GLIBC
 #define SANITIZER_INTERCEPT_NETENT (SI_LINUX || SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_SETVBUF \
-  (SI_NETBSD || SI_FREEBSD || SI_LINUX || SI_MAC)
+  (SI_NETBSD || SI_FREEBSD || SI_LINUX || SI_MAC || !SI_NOT_AIX)
 #define SANITIZER_INTERCEPT_GETMNTINFO (SI_NETBSD || SI_FREEBSD || SI_MAC)
 #define SANITIZER_INTERCEPT_MI_VECTOR_HASH SI_NETBSD
 #define SANITIZER_INTERCEPT_GETVFSSTAT SI_NETBSD
diff --git a/compiler-rt/lib/sanitizer_common/sanitizer_redefine_builtins.h b/compiler-rt/lib/sanitizer_common/sanitizer_redefine_builtins.h
index 41e0613d6fc13..bda0f04687693 100644
--- a/compiler-rt/lib/sanitizer_common/sanitizer_redefine_builtins.h
+++ b/compiler-rt/lib/sanitizer_common/sanitizer_redefine_builtins.h
@@ -15,7 +15,7 @@
 #    define SANITIZER_REDEFINE_BUILTINS_H
 
 // The asm hack only works...
[truncated]

compiler-rt/lib/asan/asan_interceptors.cpp

MaskRay · 2025-06-13T04:00:55Z

Can you add some description to the commit message? What's the interception strategy on AIX?

jakeegan · 2025-06-24T17:18:20Z

Updated PR and description. If there's no more comments, I'll merge soon.

hubert-reinterpretcast

The patch is generally lacking in code comments to explain why things are different on AIX.

Format the whole `asan_interceptors.h` file to prepare for some changes in PR #131870.

Format the whole `asan_interceptors.h` file to prepare for some changes in PR llvm/llvm-project#131870.

Format the whole `asan_interceptors.h` file to prepare for some changes in PR llvm#131870.

jakeegan · 2025-12-10T18:46:58Z

Rebased the PR after the file asan_interceptors.h was reformatted. Are there any further comments/reviews?

hubert-reinterpretcast · 2025-12-11T04:24:22Z

compiler-rt/lib/asan/asan_allocator.cpp

+      // AIX currently can't retrieve memcpy's address, we have to use
+      // internal_memcpy here.


The comment as written is misleading: it implies that being unable to retrieve an address for the real memcpy as a member of a shared library (which, afaik, is not something that we plan to change) necessarily means that we have to use internal_memcpy (i.e., there are no other options).

Perhaps change this comment to reflect any future plans to use ___memcpy on AIX? Better yet: Update this PR to make use of ___memcpy (which is what the memcpy wrapper that AIX libc provides for static linking forwards to).

hubert-reinterpretcast · 2025-12-11T04:28:31Z

compiler-rt/lib/asan/asan_interceptors.cpp

+#  endif
+}
+
+[[maybe_unused]] static inline uptr MaybeRealStrnlen(const char* s,


What is the reason for the addition of [[maybe_unused]] here as part of this PR?

hubert-reinterpretcast · 2025-12-11T04:40:56Z

compiler-rt/lib/asan/asan_interceptors.cpp

      result = REAL(pthread_create)(thread, attr, asan_thread_start, t);
+// AIX pthread_t is unsigned int.
+#    if SANITIZER_AIX
+      return result ? 0 : *(unsigned*)(thread);


Nit: Match style with similar line below.

Suggested change

return result ? 0 : *(unsigned*)(thread);

return result ? 0 : *(unsigned *)(thread);

hubert-reinterpretcast · 2025-12-11T04:46:07Z

compiler-rt/lib/asan/asan_interceptors.cpp

+#  endif

-#if ASAN_INTERCEPT__LONGJMP
+#  if ASAN_INTERCEPT__LONGJMP


This looks like a formatting change that should have happened on a separate PR?

hubert-reinterpretcast · 2025-12-11T05:05:00Z

compiler-rt/lib/asan/asan_interceptors.cpp

-#if SANITIZER_INTERCEPT_STRNLEN
+static inline void internal_or_real_memcpy(void* new_mem, const char* s,
+                                           uptr length) {
+#  if SANITIZER_INTERCEPT_MEMCPY


I don't see the logical reason to use SANITIZER_INTERCEPT_MEMCPY here as

it is unexplained why enabling interception for memcpy means we cannot use internal_memcpy in this context, and

this PR is supposed to be an AIX-specific change, so why have a change that reads as being platform-agnostic?

SANITIZER_INTERCEPT_MEMCPY was added in d7c6cad, which basically makes it an alias of SI_NOT_AIX. I believe this should just use SI_NOT_AIX directly (with a TODO to remove this function when REAL(memcpy) is set up to work on AIX).

hubert-reinterpretcast · 2025-12-11T05:14:34Z

compiler-rt/lib/asan/asan_interceptors.cpp

+#  if ASAN_INTERCEPT_EXIT
+INTERCEPTOR(void, exit, int status) {
+  AsanInitFromRtl();
+  StopInitOrderChecking();
+  REAL(exit)(status);
+}
+#  endif


Maybe split adding new interceptors out to a separate PR and focus this PR on compatibility changes to existing sanitizer code.

hubert-reinterpretcast · 2025-12-11T05:24:44Z

compiler-rt/lib/asan/asan_interceptors.h

 // really defined to replace libc functions.
 #if !SANITIZER_FUCHSIA

+// AIX currently can't retrieve the address of longjmp


Add TODO comment to investigate using longjmpx (and also intercepting it).

Also, for all "AIX currently can't retrieve the address of [ ... ]" comments, the correct phrasing is "Sanitizer on AIX is currently unable to retrieve the address of the real [ ... ] (or an alternative thereto)".

hubert-reinterpretcast · 2025-12-11T05:25:06Z

compiler-rt/lib/asan/asan_interceptors.h

 // intercepted on a given platform.
 #  if !SANITIZER_WINDOWS
-#    define ASAN_INTERCEPT__LONGJMP 1
+   // AIX currently can't retrieve the address of _longjmp


Add TODO comment re: _longjmpx.

hubert-reinterpretcast · 2025-12-11T05:26:13Z

compiler-rt/lib/asan/asan_interceptors.h

 #  endif

-#  if !SANITIZER_WINDOWS
+// AIX currently can't retrieve the address of siglongjmp


Add TODO comment re: sigsetjmpx.

hubert-reinterpretcast · 2025-12-11T05:46:19Z

compiler-rt/lib/asan/asan_interceptors.h

 #  endif

-#  if !SANITIZER_WINDOWS
+// AIX currently can't retrieve the address of __cxa_atexit


Suggested change

// AIX currently can't retrieve the address of __cxa_atexit

// Clang on AIX neither uses `__cxa_atexit` nor links against a library with

// such.

// TODO: Consider intercepting `atexit` and `unatexit` on AIX.

hubert-reinterpretcast · 2025-12-11T06:05:25Z

compiler-rt/lib/asan/asan_interceptors.h

 #  endif

+#  if SANITIZER_AIX
+#    define ASAN_INTERCEPT_EXIT 1


What is special about AIX that we need it for AIX and not other platforms? What happens on AIX if we don't have this?

[sanitizer_common] Implement interceptors for AIX

1c2952c

Define SANITIZER_AIX temporarily to build

549b134

jakeegan changed the title ~~[sanitizer_common] Implement interceptors for AIX~~ [sanitizer_common] Implement address sanitizer on AIX: interceptors (5/n) Mar 24, 2025

jakeegan added 2 commits March 24, 2025 08:44

Revert "Define SANITIZER_AIX temporarily to build"

d3301a3

This reverts commit 549b134.

convert to preprocessor directives

6ffa232

jakeegan mentioned this pull request Mar 24, 2025

[sanitizer_common] Implement address sanitizer on AIX: add/update functions #131868

Open

jakeegan marked this pull request as ready for review March 24, 2025 17:41

llvmbot added compiler-rt compiler-rt:sanitizer labels Mar 24, 2025

jakeegan requested review from hubert-reinterpretcast, vitalybuka and w2yehia March 24, 2025 17:41

jakeegan requested a review from mandlebug March 24, 2025 17:42

jakeegan added 2 commits April 1, 2025 10:12

Fix formatting

4890013

Add asan interceptor part

dcde670

llvmbot added the compiler-rt:asan Address sanitizer label Apr 24, 2025

jakeegan changed the title ~~[sanitizer_common] Implement address sanitizer on AIX: interceptors (5/n)~~ [sanitizer_common][asan] Implement address sanitizer on AIX: interceptors (5/n) Apr 24, 2025

jakeegan mentioned this pull request Apr 24, 2025

[asan] Implement address sanitizer on AIX: shadow memory #136874

Open

jakeegan and others added 7 commits April 27, 2025 13:42

Merge branch 'main' into asan_common3

c254027

Remove typo

878a567

Fix undeclared identifier

7cc64a1

Correct backslash mistake

4159aa9

Fix return

b0098db

Fix formatting

b159f20

Move misplaced endif

67317ad

jakeegan changed the title ~~[sanitizer_common][asan] Implement address sanitizer on AIX: interceptors (5/n)~~ [sanitizer_common][asan] Implement address sanitizer on AIX: interceptors (5/6) Apr 29, 2025

vitalybuka reviewed May 1, 2025

View reviewed changes

compiler-rt/lib/asan/asan_interceptors.cpp Outdated Show resolved Hide resolved

vitalybuka reviewed May 1, 2025

View reviewed changes

compiler-rt/lib/asan/asan_interceptors.cpp Outdated Show resolved Hide resolved

jakeegan added 4 commits May 26, 2025 23:41

Address comments

fb70384

Address comments

b024cac

Forgot __strdup

b57eee8

Fix formatting

c3ff763

jakeegan requested a review from vitalybuka June 12, 2025 00:19

vitalybuka approved these changes Jun 13, 2025

View reviewed changes

MaskRay reviewed Jun 13, 2025

View reviewed changes

compiler-rt/lib/asan/asan_interceptors.cpp Outdated Show resolved Hide resolved

Swap to unsigned

ab72dfa

vitalybuka requested a review from MaskRay June 24, 2025 17:39

hubert-reinterpretcast reviewed Jun 25, 2025

View reviewed changes

jakeegan added 4 commits July 7, 2025 16:20

Update comments

72b85c9

Forgot siglongjmp

93a5093

Fix siglongjmp comment

ad0b183

Fix formatting

1b40995

jakeegan requested a review from hubert-reinterpretcast July 7, 2025 20:37

jakeegan added 4 commits November 19, 2025 14:45

Merge branch 'main' into asan_common3

b7d52db

Merge branch 'main' into asan_common3

cab7b84

Handle memmove too

18a60b9

Fix formatting

0048e02

jakeegan mentioned this pull request Dec 2, 2025

[NFC][asan] Fix formatting of asan_interceptors.h #170361

Merged

jakeegan added a commit that referenced this pull request Dec 5, 2025

[NFC][asan] Fix formatting of asan_interceptors.h (#170361)

2833d9f

Format the whole `asan_interceptors.h` file to prepare for some changes in PR #131870.

llvm-sync bot pushed a commit to arm/arm-toolchain that referenced this pull request Dec 5, 2025

Automerge: [NFC][asan] Fix formatting of asan_interceptors.h (#170361)

340df98

Format the whole `asan_interceptors.h` file to prepare for some changes in PR llvm/llvm-project#131870.

honeygoyal pushed a commit to honeygoyal/llvm-project that referenced this pull request Dec 9, 2025

[NFC][asan] Fix formatting of asan_interceptors.h (llvm#170361)

d43c186

Format the whole `asan_interceptors.h` file to prepare for some changes in PR llvm#131870.

jakeegan added 2 commits December 10, 2025 09:39

Merge branch 'main' into asan_common3

54f6a42

Fix formatting

cb545f5

hubert-reinterpretcast reviewed Dec 11, 2025

View reviewed changes

		// AIX currently can't retrieve memcpy's address, we have to use
		// internal_memcpy here.

	return result ? 0 : (unsigned)(thread);
	return result ? 0 : (unsigned )(thread);

-// AIX currently can't retrieve the address of __cxa_atexit
+// Clang on AIX neither uses `__cxa_atexit` nor links against a library with
+// such.
+// TODO: Consider intercepting `atexit` and `unatexit` on AIX.

[asan] Implement interception on AIX #131870

Are you sure you want to change the base?

[asan] Implement interception on AIX #131870

Uh oh!

Conversation

jakeegan commented Mar 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Mar 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

llvmbot commented Mar 24, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

MaskRay commented Jun 13, 2025

Uh oh!

jakeegan commented Jun 24, 2025

Uh oh!

hubert-reinterpretcast left a comment

Choose a reason for hiding this comment

Uh oh!

jakeegan commented Dec 10, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

hubert-reinterpretcast Dec 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

jakeegan commented Mar 18, 2025 •

edited

Loading

github-actions bot commented Mar 18, 2025 •

edited

Loading

hubert-reinterpretcast Dec 11, 2025 •

edited

Loading