public inbox for libstdc++@gcc.gnu.org
 help / color / mirror / Atom feed
* [committed] libstdc++: Check for EOF if extraction avoids buffer overflow [PR106248]
@ 2022-07-12 22:40 Jonathan Wakely
  0 siblings, 0 replies; only message in thread
From: Jonathan Wakely @ 2022-07-12 22:40 UTC (permalink / raw)
  To: libstdc++, gcc-patches

Tested x86_64-linux, pushed to trunk. Backports to gcc-12 and gcc-11 to
follow.

-- >8 --

In r11-2581-g17abcc77341584 (for LWG 2499) I added overflow checks to
the pre-C++20 operator>>(istream&, char*) overload.  Those checks can
cause extraction to stop after filling the buffer, where previously it
would have tried to extract another character and stopped at EOF. When
that happens we no longer set eofbit in the stream state, which is
consistent with the behaviour of the new C++20 overload, but is an
observable and unexpected change in the C++17 behaviour. What makes it
worse is that the behaviour change is dependent on optimization, because
__builtin_object_size is used to detect the buffer size and that only
works when optimizing.

To avoid the unexpected and optimization-dependent change in behaviour,
set eofbit manually if we stopped extracting because of the buffer size
check, but had reached EOF anyway. If the stream's rdstate() != goodbit
or width() is non-zero and smaller than the buffer, there's nothing to
do. Otherwise, we filled the buffer and need to check for EOF, and maybe
set eofbit.

The new check is guarded by #ifdef __OPTIMIZE__ because otherwise
__builtin_object_size is useless. There's no point compiling and
emitting dead code that can't be eliminated because we're not
optimizing.

We could add extra checks that the next character in the buffer is not
whitespace, to detect the case where we stopped early and prevented a
buffer overflow that would have happened otherwise. That would allow us
to assert or set badbit in the stream state when undefined behaviour was
prevented. However, those extra checks would increase the size of the
function, potentially reducing the likelihood of it being inlined, and
so making the buffer size detection less reliable. It seems preferable
to prevent UB and silently truncate, rather than miss the UB and allow
the overflow to happen.

libstdc++-v3/ChangeLog:

	PR libstdc++/106248
	* include/std/istream [C++17] (operator>>(istream&, char*)):
	Set eofbit if we stopped extracting at EOF.
	* testsuite/27_io/basic_istream/extractors_character/char/pr106248.cc:
	New test.
	* testsuite/27_io/basic_istream/extractors_character/wchar_t/pr106248.cc:
	New test.
---
 libstdc++-v3/include/std/istream              | 33 ++++++++++++---
 .../extractors_character/char/pr106248.cc     | 40 +++++++++++++++++++
 .../extractors_character/wchar_t/pr106248.cc  | 40 +++++++++++++++++++
 3 files changed, 107 insertions(+), 6 deletions(-)
 create mode 100644 libstdc++-v3/testsuite/27_io/basic_istream/extractors_character/char/pr106248.cc
 create mode 100644 libstdc++-v3/testsuite/27_io/basic_istream/extractors_character/wchar_t/pr106248.cc

diff --git a/libstdc++-v3/include/std/istream b/libstdc++-v3/include/std/istream
index b506c4f7504..416ef556fa1 100644
--- a/libstdc++-v3/include/std/istream
+++ b/libstdc++-v3/include/std/istream
@@ -784,7 +784,7 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
    *  - if `width()` is greater than zero, `n` is `min(width(), n)`
    *  - otherwise `n` is the number of elements of the array
    *  - (before C++20 the pointer is assumed to point to an array of
-   *  - the largest possible size for an array of `char_type`).
+   *    the largest possible size for an array of `char_type`).
    *
    *  Characters are extracted and stored until one of the following happens:
    *  - `n - 1` characters are stored
@@ -802,19 +802,40 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
     inline basic_istream<_CharT, _Traits>&
     operator>>(basic_istream<_CharT, _Traits>& __in, _CharT* __s)
     {
+#ifdef __OPTIMIZE__
+      // Function inlining might make the buffer size known, allowing us to
+      // prevent overflow.
       size_t __n = __builtin_object_size(__s, 0);
-      if (__builtin_expect(__n < sizeof(_CharT), false))
+      if (__n < sizeof(_CharT))
 	{
 	  // There is not even space for the required null terminator.
 	  __glibcxx_assert(__n >= sizeof(_CharT));
+	  // No point calling __istream_extract, but still need to reset width.
 	  __in.width(0);
 	  __in.setstate(ios_base::failbit);
 	}
-      else
+      else if (__n != (size_t)-1)
 	{
-	  if (__n == (size_t)-1)
-	    __n = __gnu_cxx::__numeric_traits<streamsize>::__max;
-	  std::__istream_extract(__in, __s, __n / sizeof(_CharT));
+	  __n /= sizeof(_CharT);
+	  streamsize __w = __in.width();
+	  std::__istream_extract(__in, __s, __n);
+	  if (__in.good() && (__w <= 0 || __n < __w))
+	    {
+	      // Stopped extracting early to avoid overflowing the buffer,
+	      // but might have stopped anyway (and set eofbit) if at EOF.
+	      const typename _Traits::int_type __c = __in.rdbuf()->sgetc();
+	      const bool __eof = _Traits::eq_int_type(__c, _Traits::eof());
+	      if (__builtin_expect(__eof, true)) // Assume EOF, not overflow.
+		__in.setstate(ios_base::eofbit);
+	    }
+	}
+      else
+#endif // __OPTIMIZE
+	{
+	  // Buffer size is unknown, have to assume it's huge.
+	  streamsize __n = __gnu_cxx::__numeric_traits<streamsize>::__max;
+	  __n /= sizeof(_CharT);
+	  std::__istream_extract(__in, __s, __n);
 	}
       return __in;
     }
diff --git a/libstdc++-v3/testsuite/27_io/basic_istream/extractors_character/char/pr106248.cc b/libstdc++-v3/testsuite/27_io/basic_istream/extractors_character/char/pr106248.cc
new file mode 100644
index 00000000000..6d89a0e5fef
--- /dev/null
+++ b/libstdc++-v3/testsuite/27_io/basic_istream/extractors_character/char/pr106248.cc
@@ -0,0 +1,40 @@
+// { dg-do run }
+
+#include <sstream>
+#include <testsuite_hooks.h>
+
+void
+test_pr106248()
+{
+  char buf[5] = {'x', 'x', 'x', 'x', 'x'};
+  std::string s("  four");
+  std::istringstream in(s);
+  in >> buf;
+#if __cplusplus >= 202002L
+  // Extraction stops because buffer is full.
+  VERIFY( in.good() );
+#else
+  // PR libstdc++/106248
+  // Extraction stops because all input has been consumed and eofbit is set.
+  VERIFY( in.eof() );
+#endif
+  // Extracted string must be null-terminated.
+  VERIFY( buf[4] == '\0' );
+  VERIFY( std::string(buf) == "four" );
+
+  in.clear();
+  in.str(s);
+  for (int i = 0; i < 5; ++i)
+    s[i] = 'x';
+
+  in.width(5);
+  in >> buf;
+  // Extraction stops due to field width, eofbit not set.
+  VERIFY( in.good() );
+  VERIFY( std::string(buf) == "four" );
+}
+
+int main()
+{
+  test_pr106248();
+}
diff --git a/libstdc++-v3/testsuite/27_io/basic_istream/extractors_character/wchar_t/pr106248.cc b/libstdc++-v3/testsuite/27_io/basic_istream/extractors_character/wchar_t/pr106248.cc
new file mode 100644
index 00000000000..7c226600b9e
--- /dev/null
+++ b/libstdc++-v3/testsuite/27_io/basic_istream/extractors_character/wchar_t/pr106248.cc
@@ -0,0 +1,40 @@
+// { dg-do run }
+
+#include <sstream>
+#include <testsuite_hooks.h>
+
+void
+test_pr106248()
+{
+  wchar_t buf[5] = {L'x', L'x', L'x', L'x', L'x'};
+  std::wstring s(L"  four");
+  std::wistringstream in(s);
+  in >> buf;
+#if __cplusplus >= 202002L
+  // Extraction stops because buffer is full.
+  VERIFY( in.good() );
+#else
+  // PR libstdc++/106248
+  // Extraction stops because all input has been consumed and eofbit is set.
+  VERIFY( in.eof() );
+#endif
+  // Extracted string must be null-terminated.
+  VERIFY( buf[4] == L'\0' );
+  VERIFY( std::wstring(buf) == L"four" );
+
+  in.clear();
+  in.str(s);
+  for (int i = 0; i < 5; ++i)
+    s[i] = L'x';
+
+  in.width(5);
+  in >> buf;
+  // Extraction stops due to field width, eofbit not set.
+  VERIFY( in.good() );
+  VERIFY( std::wstring(buf) == L"four" );
+}
+
+int main()
+{
+  test_pr106248();
+}
-- 
2.36.1


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2022-07-12 22:40 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-07-12 22:40 [committed] libstdc++: Check for EOF if extraction avoids buffer overflow [PR106248] Jonathan Wakely

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).