Add interceptors for wcsxfrm, wcsxfrm_l

[oliverchang]

fix condition

[vitalybuka]

Are you going to move strxfrm into common as well? In the different patch?
If not, I'll move it myself, for consistency.

lib/sanitizer_common/sanitizer_common_interceptors.inc

6536

sizeof(wchar_t) * (REAL(wcslen)(src) + 1)

6539

Please remove {}

6550

sizeof(wchar_t) * (REAL(wcslen)(src) + 1)

test/msan/wcsxfrm.cc

12

you may try to add test case where it catches poisoned bytes in the src string.

[vitalybuka]

because of COMMON_INTERCEPTOR_READ_RANGE argument

[oliverchang]

address comments

[oliverchang]

lib/sanitizer_common/sanitizer_common_interceptors.inc

6536

Fixed. thanks!

6539

Done.

6550

Done.

test/msan/wcsxfrm.cc

12

Added one here.

[oliverchang]

Please commit this if this looks good, thanks!

[vitalybuka]

I'll commit.

[oliverchang]

Thanks!

And oops, missed your comment about strxfrm. I can move that in a subsequent patch if you haven't already started doing that.

[kubamracek]

Hi, this seems to be failing at http://green.lab.llvm.org/green/job/clang-stage1-configure-RA/43377/console. Can you take a look?

[oliverchang]

Hmm, looks like we might need "#include <xlocale.h>" when compiling on macos for both compiler-rt/test/sanitizer_common/TestCases/Posix/strxfrm.c and compiler-rt/test/sanitizer_common/TestCases/Posix/wcsxfrm.c ?

lib/sanitizer_common/sanitizer_common_interceptors.inc

 }
 #define INIT_WCSCAT                  \
   COMMON_INTERCEPT_FUNCTION(wcscat); \
   COMMON_INTERCEPT_FUNCTION(wcsncat);
 #else
 #define INIT_WCSCAT
 #endif
 
+#if SANITIZER_INTERCEPT_WCSXFRM
+INTERCEPTOR(SIZE_T, wcsxfrm, wchar_t *dest, const wchar_t *src, SIZE_T len) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, wcsxfrm, dest, src, len);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, src,
+                                sizeof(wchar_t) * (REAL(wcslen)(src) + 1));
+
+  SIZE_T res = REAL(wcsxfrm)(dest, src, len);
+  if (res < len)
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dest, sizeof(wchar_t) * (res + 1));
+
+  return res;
+}
+
+INTERCEPTOR(SIZE_T, wcsxfrm_l, wchar_t *dest, const wchar_t *src, SIZE_T len,
+            void *locale) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, wcsxfrm_l, dest, src, len, locale);
+  COMMON_INTERCEPTOR_READ_RANGE(ctx, src,
+                                sizeof(wchar_t) * (REAL(wcslen)(src) + 1));
+
+  SIZE_T res = REAL(wcsxfrm_l)(dest, src, len, locale);
+  if (res < len)
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, dest, sizeof(wchar_t) * (res + 1));
+
+  return res;
+}
+
+#define INIT_WCSXFRM                  \
+  COMMON_INTERCEPT_FUNCTION(wcsxfrm); \
+  COMMON_INTERCEPT_FUNCTION(wcsxfrm_l);
+#else
+#define INIT_WCSXFRM
+#endif
+
 #if SANITIZER_INTERCEPT_ACCT
 INTERCEPTOR(int, acct, const char *file) {
   void *ctx;
   COMMON_INTERCEPTOR_ENTER(ctx, acct, file);
   if (file)
     COMMON_INTERCEPTOR_READ_RANGE(ctx, file, REAL(strlen)(file) + 1);
   return REAL(acct)(file);
 }
static void InitializeCommonInterceptors() {
   INIT___LXSTAT;
   INIT___LXSTAT64;
   // FIXME: add other *stat interceptors.
   INIT_UTMP;
   INIT_UTMPX;
   INIT_GETLOADAVG;
   INIT_WCSLEN;
   INIT_WCSCAT;
+  INIT_WCSXFRM;
   INIT_ACCT;
   INIT_USER_FROM_UID;
   INIT_UID_FROM_USER;
   INIT_GROUP_FROM_GID;
   INIT_GID_FROM_GROUP;
   INIT_ACCESS;
   INIT_FACCESSAT;
   INIT_GETGROUPLIST;

lib/sanitizer_common/sanitizer_platform_interceptors.h

 #define SANITIZER_INTERCEPT_SETLOCALE SI_POSIX
 #define SANITIZER_INTERCEPT_GETCWD SI_POSIX
 #define SANITIZER_INTERCEPT_GET_CURRENT_DIR_NAME SI_LINUX_NOT_ANDROID
 #define SANITIZER_INTERCEPT_STRTOIMAX SI_POSIX
 #define SANITIZER_INTERCEPT_MBSTOWCS SI_POSIX
 #define SANITIZER_INTERCEPT_MBSNRTOWCS \
   (SI_MAC || SI_LINUX_NOT_ANDROID || SI_SOLARIS)
 #define SANITIZER_INTERCEPT_WCSTOMBS SI_POSIX
+#define SANITIZER_INTERCEPT_WCSXFRM SI_POSIX
 #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_TCGETATTR SI_LINUX_NOT_ANDROID || SI_SOLARIS
 #define SANITIZER_INTERCEPT_REALPATH SI_POSIX
 #define SANITIZER_INTERCEPT_CANONICALIZE_FILE_NAME \
   (SI_LINUX_NOT_ANDROID || SI_SOLARIS)

test/msan/wcsxfrm.cc

+// RUN: %clangxx_msan -O0 -g %s -o %t && not %run %t
+
+#include <assert.h>
+#include <locale.h>
+#include <sanitizer/msan_interface.h>
+#include <stdlib.h>
+#include <wchar.h>
+
+int main(void) {
+  wchar_t q[10];
+  size_t n = wcsxfrm(q, L"abcdef", sizeof(q) / sizeof(wchar_t));
+  assert(n < sizeof(q));
+  __msan_check_mem_is_initialized(q, (n + 1) * sizeof(wchar_t));
+
+  locale_t loc = newlocale(LC_ALL_MASK, "", (locale_t)0);
+
+  __msan_poison(&q, sizeof(q));
+  n = wcsxfrm_l(q, L"qwerty", sizeof(q) / sizeof(wchar_t), loc);
+  assert(n < sizeof(q));
+  __msan_check_mem_is_initialized(q, (n + 1) * sizeof(wchar_t));
+
+  q[0] = 'A';
+  q[1] = '\x00';
+  __msan_poison(&q, sizeof(q));
+  wcsxfrm(NULL, q, 0);
+
+  // CHECK: WARNING: MemorySanitizer: use-of-uninitialized-value
+  // CHECK:    in main {{.*}}wcsxfrm.cc:25
+  return 0;
+}

test/sanitizer_common/TestCases/Posix/wcsxfrm.c

+// RUN: %clang -O0 %s -o %t && %run %t
+
+#include <assert.h>
+#include <locale.h>
+#include <wchar.h>
+
+int main(int argc, char **argv) {
+  wchar_t q[10];
+  size_t n = wcsxfrm(q, L"abcdef", sizeof(q) / sizeof(wchar_t));
+  assert(n < sizeof(q));
+
+  wchar_t q2[10];
+  locale_t loc = newlocale(LC_ALL_MASK, "", (locale_t)0);
+  n = wcsxfrm_l(q2, L"qwerty", sizeof(q) / sizeof(wchar_t), loc);
+  assert(n < sizeof(q2));
+
+  freelocale(loc);
+  return 0;
+}