[libc] Add fuzz test for strcmp.

[cgyurgyik]

Summary

Adds a fuzz test for string comparison.

This takes in two strings with associated lengths.
Verifies each string contains at least one character, and that the last character is the null terminator.
Then, finds the first instance where one of the following does not hold:

1. i < min(size1, size2)
2. s1[i] == s2[i]
3. s1[i] != '\0'

The result of strcmp is then compared to the value of the difference between s1[i] and s2[i]. For thoroughness, the operands are reversed and also checked.

[PaulkaToast]

After comments are addressed LGTM but you might want to wait for approval from sivachandra.

libc/fuzzing/string/strcmp_fuzz.cpp

28

I think its beneficial to avoid dependency on C++ headers for the fuzzing tests, since std::min is only used here maybe just replace this with an if condition? It could in theory produce a circular dependency because libc++ could depend on some libc functions which is what we are trying to fuzz here.

36

nit: using continue; would make this more explicit but its totally up to you. (:

[sivachandra]

OK after addressing Paula's comments.

libc/fuzzing/string/strcmp_fuzz.cpp

13

I see Paula's comment related to this but I would like to drive this more strongly: The only headers not part of LLVM libc that are OK to include are the freestanding C headers, and the headers provided by the compiler (like stdatomic.h).

[cgyurgyik]

Remove use of cpp header. Use a break statement to be more explicit.

[cgyurgyik]

Thanks Paula.

In D82247#2108113, @sivachandra wrote:

OK after addressing Paula's comments.

Thanks Siva, and acknowledged about headers.

[cgyurgyik]

Done. Used a break statement to be more explicit.

libc/fuzzing/string/strcmp_fuzz.cpp

 //===-- strcmp_fuzz.cpp ---------------------------------------------------===//
 //
 // 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
 //
 //===----------------------------------------------------------------------===//
 ///
 /// Fuzzing test for llvm-libc strcmp implementation.
 ///
 //===----------------------------------------------------------------------===//
 #include "src/string/strcmp.h"
-#include <algorithm>
 #include <stdint.h>
 
 extern "C" int LLVMFuzzerTestTwoInputs(const uint8_t *data1, size_t size1,
                                        const uint8_t *data2, size_t size2) {
   // Verify each data source contains at least one character.
   if (!size1 || !size2)
     return 0;
   // Verify that the final character is the null terminator.
   if (data1[size1 - 1] != '\0' || data2[size2 - 1] != '\0')
     return 0;
 
   const char *s1 = reinterpret_cast<const char *>(data1);
   const char *s2 = reinterpret_cast<const char *>(data2);
 
-  const size_t minimum_size = std::min(size1, size2);
+  const size_t minimum_size = size1 < size2 ? size1 : size2;
 
   // Iterate through until either the minimum size is hit,
   // a character is the null terminator, or the first set
   // of differed bytes between s1 and s2 are found.
   // No bytes following a null byte should be compared.
   size_t i;
-  for (i = 0; i < minimum_size && s1[i] && s1[i] == s2[i]; ++i)
-    ;
+  for (i = 0; i < minimum_size; ++i) {
+    if (!s1[i] || s1[i] != s2[i])
+      break;
+  }
 
   int expected_result = s1[i] - s2[i];
   int actual_result = __llvm_libc::strcmp(s1, s2);
 
   // The expected result should be the difference between the first non-equal
   // characters of s1 and s2. If all characters are equal, the expected result
   // should be '\0' - '\0' = 0.
   if (expected_result != actual_result)