[DFSAN] Add support for strnlen

[tkuchta]

Summary

This patch adds a support for the libc strnlen() function in DFSAN

[tkuchta]

apply clang format

[browneee]

LGTM!

Sorry for my slow review, I was busy with other stuff.

Do you want me to submit?

[tkuchta]

No worries! Thank you for the review. Yes, please submit the patch. I will prepare the next function for the review.

compiler-rt/lib/dfsan/dfsan_custom.cpp

SANITIZER_INTERFACE_ATTRIBUTE size_t __dfso_strlen(const char *s,
                                                    dfsan_origin s_origin,
                                                    dfsan_origin *ret_origin) {
   size_t ret = __dfsw_strlen(s, s_label, ret_label);
   if (!flags().strict_data_dependencies)
     *ret_origin = dfsan_read_origin_of_first_taint(s, ret + 1);
   return ret;
 }
 
+SANITIZER_INTERFACE_ATTRIBUTE size_t __dfsw_strnlen(const char *s,
+                                                    size_t maxlen,
+                                                    dfsan_label s_label,
+                                                    dfsan_label maxlen_label,
+                                                    dfsan_label *ret_label) {
+  size_t ret = strnlen(s, maxlen);
+  if (flags().strict_data_dependencies) {
+    *ret_label = 0;
+  } else {
+    size_t full_len = strlen(s);
+    size_t covered_len = maxlen > (full_len + 1) ? (full_len + 1) : maxlen;
+    *ret_label = dfsan_union(maxlen_label, dfsan_read_label(s, covered_len));
+  }
+  return ret;
+}
+
+SANITIZER_INTERFACE_ATTRIBUTE size_t __dfso_strnlen(
+    const char *s, size_t maxlen, dfsan_label s_label, dfsan_label maxlen_label,
+    dfsan_label *ret_label, dfsan_origin s_origin, dfsan_origin maxlen_origin,
+    dfsan_origin *ret_origin) {
+  size_t ret = __dfsw_strnlen(s, maxlen, s_label, maxlen_label, ret_label);
+  if (!flags().strict_data_dependencies) {
+    size_t full_len = strlen(s);
+    size_t covered_len = maxlen > (full_len + 1) ? (full_len + 1) : maxlen;
+    dfsan_origin o = dfsan_read_origin_of_first_taint(s, covered_len);
+    *ret_origin = o ? o : maxlen_origin;
+  }
+  return ret;
+}
+
 static void *dfsan_memmove(void *dest, const void *src, size_t n) {
   dfsan_label *sdest = shadow_for(dest);
   const dfsan_label *ssrc = shadow_for(src);
   internal_memmove((void *)sdest, (const void *)ssrc, n * sizeof(dfsan_label));
   return internal_memmove(dest, src, n);
 }
 
 static void *dfsan_memmove_with_origin(void *dest, const void *src, size_t n) {

compiler-rt/lib/dfsan/done_abilist.txt

 # necessarily data dependent.
 fun:bcmp=custom
 fun:memchr=custom
 fun:memcmp=custom
 fun:strcasecmp=custom
 fun:strchr=custom
 fun:strcmp=custom
 fun:strlen=custom
+fun:strnlen=custom
 fun:strncasecmp=custom
 fun:strncmp=custom
 fun:strpbrk=custom
 fun:strrchr=custom
 fun:strstr=custom
 fun:strsep=custom
 
 # Functions which take action based on global state, such as running a callback

compiler-rt/test/dfsan/custom.cpp

 #ifdef STRICT_DATA_DEPENDENCIES
   ASSERT_ZERO_LABEL(rv);
 #else
   ASSERT_LABEL(rv, i_label);
   ASSERT_EQ_ORIGIN(rv, str1[3]);
 #endif
 }
 
+void test_strnlen() {
+  char str1[] = "str1";
+  dfsan_set_label(i_label, &str1[3], 1);
+
+  int maxlen = 4;
+  dfsan_set_label(j_label, &maxlen, sizeof(maxlen));
+
+  int rv = strnlen(str1, maxlen);
+  assert(rv == 4);
+#ifdef STRICT_DATA_DEPENDENCIES
+  ASSERT_ZERO_LABEL(rv);
+#else
+  ASSERT_LABEL(rv, dfsan_union(i_label, j_label));
+  ASSERT_EQ_ORIGIN(rv, str1[3]);
+#endif
+
+  maxlen = 2;
+  dfsan_set_label(j_label, &maxlen, sizeof(maxlen));
+  rv = strnlen(str1, maxlen);
+  assert(rv == 2);
+#ifdef STRICT_DATA_DEPENDENCIES
+  ASSERT_ZERO_LABEL(rv);
+#else
+  ASSERT_LABEL(rv, j_label);
+  ASSERT_EQ_ORIGIN(rv, maxlen);
+#endif
+}
+
 void test_strdup() {
   char str1[] = "str1";
   dfsan_set_label(i_label, &str1[3], 1);
   DEFINE_AND_SAVE_ORIGINS(str1)
 
   char *strd = strdup(str1);
   ASSERT_ZERO_LABEL(strd);
   ASSERT_ZERO_LABEL(strd[0]);
int main(void) {
   test_stat();
   test_strcasecmp();
   test_strchr();
   test_strcmp();
   test_strcat();
   test_strcpy();
   test_strdup();
   test_strlen();
+  test_strnlen();
   test_strncasecmp();
   test_strncmp();
   test_strncpy();
   test_strpbrk();
   test_strsep();
   test_strrchr();
   test_strstr();
   test_strtod();
   test_strtol();
   test_strtoll();
   test_strtoul();
   test_strtoull();
   test_time();
   test_write();
 }