[asan] Implement interception on AIX #131870

jakeegan · 2025-03-18T17:41:22Z

Adjust asan interceptor support for AIX.

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

vitalybuka · 2025-05-01T07:36:37Z

compiler-rt/lib/asan/asan_interceptors.cpp

@@ -596,7 +609,13 @@ INTERCEPTOR(char*, strdup, const char *s) {
  GET_STACK_TRACE_MALLOC;
  void *new_mem = asan_malloc(length + 1, &stack);
  if (new_mem) {
+#  if SANITIZER_AIX
+    // memcpy is a static function defined in libc.a on AIX. It can not be
+    // intercepted, so REAL(memcpy) is null on AIX. Use internal_memcpy instead.


Can you please extract inline internal_or_real_memcpy() in the top of the file?

#if SANITIZER_INTERCEPT_MEMCPY rather than SANITIZER_AIX?

vitalybuka · 2025-05-01T07:39:07Z

compiler-rt/lib/asan/asan_interceptors.cpp

@@ -56,6 +56,7 @@ namespace __asan {
 #  define ASAN_READ_STRING(ctx, s, n) \
    ASAN_READ_STRING_OF_LEN((ctx), (s), internal_strlen(s), (n))

+#  if SANITIZER_INTERCEPT_STRCAT || SANITIZER_INTERCEPT_STRCPY
 static inline uptr MaybeRealStrnlen(const char *s, uptr maxlen) {


Let's use [[maybe_unused]] instead of #if

compiler-rt/lib/asan/asan_interceptors.h

compiler-rt/lib/asan/asan_interceptors_memintrinsics.cpp

vitalybuka · 2025-05-01T07:55:55Z

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


[[maybe_unused]] instead of #if?

compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc

compiler-rt/lib/interception/interception.h

vitalybuka · 2025-05-01T08:03:53Z

compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc

@@ -1807,17 +1816,29 @@ FORMAT_INTERCEPTOR_IMPL(__isoc99_snprintf, __isoc99_vsnprintf, str, size,
 #endif  // SANITIZER_INTERCEPT_PRINTF

 #if SANITIZER_INTERCEPT_PRINTF


Maybe something like:

#if SANITIZER_INTERCEPT_PRINTF # define INIT_PRINTF_COMMON \ 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); # if !SANITIZER_AIX # define INIT_PRINTF_EXTRA \ COMMON_INTERCEPT_FUNCTION_LDBL(asprintf); \ COMMON_INTERCEPT_FUNCTION_LDBL(vasprintf); \ #else #define INIT_PRINTF_EXTRA #endif # define INIT_PRINTF INIT_PRINTF_COMMON INIT_PRINTF_EXTRA # endif #else #define INIT_PRINTF #endif

compiler-rt/lib/sanitizer_common/sanitizer_common_interceptors.inc

jakeegan · 2025-05-06T14:01:31Z

Split part of the PR to:
#138606
#138608

Still have to address some comments.

[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 vitalybuka, hubert-reinterpretcast 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