Add interceptors for the strtoi(3)/strtou(3) from NetBSD

[krytarowski]

Summary

strtoi/strtou converts string value to an intmax_t/uintmax_t integer.

Add a dedicated test.

Enable this API for NetBSD.

It's a reworked version of the original work by Yang Zheng.

[vitalybuka]

lib/sanitizer_common/sanitizer_common_interceptors.inc

7303

There is INTERCEPTORS_STRTO_BASE in msan
Probably we should try to move them into common, and than add new interceptors.
If this is hard to move, then just add strtoi there

[krytarowski]

lib/sanitizer_common/sanitizer_common_interceptors.inc

7303

Do you mean, to add strtoi for MSan only?

Yes, the interceptors shall be made generic.. however we are not able to do it in this iteration.

[krytarowski]

lib/sanitizer_common/sanitizer_common_interceptors.inc

7303

BTW. I will drop you a line in a mail.

[vitalybuka]

lib/sanitizer_common/sanitizer_common_interceptors.inc

7326

we need StrtolFixAndCheck and avoid reading entire string for strict_string_checks=0

[vitalybuka]

lib/sanitizer_common/sanitizer_common_interceptors.inc

7303

not sure what is long term is better
maybe adding this just into msan and later converting to _common would be easier

[krytarowski]

• rework the interceptor

[krytarowski]

• simplify

[vitalybuka]

test/sanitizer_common/TestCases/NetBSD/strtoi.cc

71

could you please make a single test function with arguments
now they are almost identical

[krytarowski]

test/sanitizer_common/TestCases/NetBSD/strtoi.cc

71

I will do it.

lib/sanitizer_common/sanitizer_common_interceptors.inc

INTERCEPTOR(void, mi_vector_hash, const void *key, SIZE_T len, u32 seed,
   if (hashes)
     COMMON_INTERCEPTOR_WRITE_RANGE(ctx, hashes, sizeof(hashes[0]) * 3);
 }
 #define INIT_MI_VECTOR_HASH COMMON_INTERCEPT_FUNCTION(mi_vector_hash)
 #else
 #define INIT_MI_VECTOR_HASH
 #endif
 
+#if SANITIZER_INTERCEPT_STRTOI
+INTERCEPTOR(INTMAX_T, strtoi, const char *nptr, char **endptr, int base,
+            INTMAX_T low, INTMAX_T high, int *rstatus) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, strtoi, nptr, endptr, base, low, high, rstatus);
+
+  if (nptr)
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, nptr, REAL(strlen)(nptr) + 1);
+  INTMAX_T ret = REAL(strtoi)(nptr, endptr, base, low, high, rstatus);
+  if (endptr) {
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, endptr, sizeof(char *));
+    if (*endptr)
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *endptr, REAL(strlen)(*endptr) + 1);
+  }
+  if (rstatus)
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, rstatus, sizeof(int));
+  return ret;
+}
+
+INTERCEPTOR(UINTMAX_T, strtou, const char *nptr, char **endptr, int base,
+            UINTMAX_T low, UINTMAX_T high, int *rstatus) {
+  void *ctx;
+  COMMON_INTERCEPTOR_ENTER(ctx, strtou, nptr, endptr, base, low, high, rstatus);
+
+  if (nptr)
+    COMMON_INTERCEPTOR_READ_RANGE(ctx, nptr, REAL(strlen)(nptr) + 1);
+  UINTMAX_T ret = REAL(strtou)(nptr, endptr, base, low, high, rstatus);
+  if (endptr) {
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, endptr, sizeof(char *));
+    if (*endptr)
+      COMMON_INTERCEPTOR_WRITE_RANGE(ctx, *endptr, REAL(strlen)(*endptr) + 1);
+  }
+  if (rstatus)
+    COMMON_INTERCEPTOR_WRITE_RANGE(ctx, rstatus, sizeof(int));
+  return ret;
+}
+#define INIT_STRTOI                                                            \
+  COMMON_INTERCEPT_FUNCTION(strtoi);                                           \
+  COMMON_INTERCEPT_FUNCTION(strtou)
+#else
+#define INIT_STRTOI
+#endif
+
 static void InitializeCommonInterceptors() {
   static u64 metadata_mem[sizeof(MetadataHashMap) / sizeof(u64) + 1];
   interceptor_metadata_map = new((void *)&metadata_mem) MetadataHashMap();
 
   INIT_MMAP;
   INIT_MMAP64;
   INIT_TEXTDOMAIN;
   INIT_STRLEN;
static void InitializeCommonInterceptors() {
   INIT_DEVNAME_R;
   INIT_FGETLN;
   INIT_STRMODE;
   INIT_TTYENT;
   INIT_PROTOENT;
   INIT_NETENT;
   INIT_GETMNTINFO;
   INIT_MI_VECTOR_HASH;
+  INIT_STRTOI;
 
   INIT___PRINTF_CHK;
 }

lib/sanitizer_common/sanitizer_platform_interceptors.h

 #define SANITIZER_INTERCEPT_DEVNAME_R (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_FGETLN (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_STRMODE (SI_NETBSD || SI_FREEBSD)
 #define SANITIZER_INTERCEPT_TTYENT SI_NETBSD
 #define SANITIZER_INTERCEPT_PROTOENT SI_NETBSD
 #define SANITIZER_INTERCEPT_NETENT SI_NETBSD
 #define SANITIZER_INTERCEPT_GETMNTINFO SI_NETBSD
 #define SANITIZER_INTERCEPT_MI_VECTOR_HASH SI_NETBSD
+#define SANITIZER_INTERCEPT_STRTOI SI_NETBSD
 
 #endif  // #ifndef SANITIZER_PLATFORM_INTERCEPTORS_H

test/sanitizer_common/TestCases/NetBSD/strtoi.cc

+// RUN: %clangxx -O0 -g %s -o %t && %run %t 2>&1 | FileCheck %s
+
+#include <inttypes.h>
+#include <stdio.h>
+
+void test_strtoi1() {
+  const char *cent = "100";
+  char *p;
+  int status;
+  intmax_t i = strtoi(cent, &p, 0, 1, 1000, &status);
+  printf("strtoi: conversion of '%s' to a number %s, using %jd, p=%#" PRIx8
+         "\n",
+         cent, status ? "failed" : "successful", i, *p);
+}
+
+void test_strtoi2() {
+  const char *cent = "100";
+  char *p;
+  int status;
+  intmax_t i = strtoi(cent, &p, 0, 1, 10, &status);
+  printf("strtoi: conversion of '%s' to a number %s, using %jd, p=%#" PRIx8
+         "\n",
+         cent, status ? "failed" : "successful", i, *p);
+}
+
+void test_strtoi3() {
+  const char *cent = "100xyz";
+  char *p;
+  int status;
+  intmax_t i = strtoi(cent, &p, 0, 1, 10, &status);
+  printf("strtoi: conversion of '%s' to a number %s, using %jd, p=%#" PRIx8
+         "\n",
+         cent, status ? "failed" : "successful", i, *p);
+}
+
+void test_strtou1() {
+  const char *cent = "100";
+  char *p;
+  int status;
+  uintmax_t i = strtou(cent, &p, 0, 1, 1000, &status);
+  printf("strtou: conversion of '%s' to a number %s, using %ju, p=%#" PRIx8
+         "\n",
+         cent, status ? "failed" : "successful", i, *p);
+}
+
+void test_strtou2() {
+  const char *cent = "100";
+  char *p;
+  int status;
+  uintmax_t i = strtou(cent, &p, 0, 1, 10, &status);
+  printf("strtou: conversion of '%s' to a number %s, using %ju, p=%#" PRIx8
+         "\n",
+         cent, status ? "failed" : "successful", i, *p);
+}
+
+void test_strtou3() {
+  const char *cent = "100xyz";
+  char *p;
+  int status;
+  uintmax_t i = strtou(cent, &p, 0, 1, 10, &status);
+  printf("strtou: conversion of '%s' to a number %s, using %ju, p=%#" PRIx8
+         "\n",
+         cent, status ? "failed" : "successful", i, *p);
+}
+
+int main(void) {
+  printf("strtoi\n");
+
+  test_strtoi1();
+  test_strtoi2();
+  test_strtoi3();
+  test_strtou1();
+  test_strtou2();
+  test_strtou3();
+
+  // CHECK: strtoi
+  // CHECK: strtoi: conversion of '100' to a number successful, using 100, p=0
+  // CHECK: strtoi: conversion of '100' to a number failed, using 10, p=0
+  // CHECK: strtoi: conversion of '100xyz' to a number failed, using 10, p=0x78
+  // CHECK: strtou: conversion of '100' to a number successful, using 100, p=0
+  // CHECK: strtou: conversion of '100' to a number failed, using 10, p=0
+  // CHECK: strtou: conversion of '100xyz' to a number failed, using 10, p=0x78
+
+  return 0;
+}