Add code to perform overflow-checking additions and multiplication

Most processors have carry flags which they set on addition overflow, so
it's a good idea to access them whenever possible. Most of them also
have widening multiply instructions that can be used to detect overflow
of the non-widening version.

Tested to compile on:
 Architecture     Compiler
  x86             GCC 4.9, GCC 5*, Clang 3.6*, ICC 16 beta
  x86-64          GCC 4.9, GCC 5*, Clang 3.6*, ICC 16 beta
  x86-64 ILP32    GCC 4.9, GCC 5*, Clang 3.6*
  IA-64 LP64      GCC 4.8
  ARMv7-A         GCC 4.9, Clang 3.6*
  AArch64         Clang 3.6*
  MIPS            GCC 4.9, Clang 3.6*
  MIPS64          GCC 4.9, Clang 3.6*
  PowerPC         GCC 4.9, Clang 3.6*
  PowerPC 64      GCC 4.9, Clang 3.6*
  SPARC           Clang 3.6*
  SPARCv9         Clang 3.6*

[*] supports the intrinsics

If the compiler does not offer a way to detect an overflow, we do it by
hand. For unsigned additions, that's easy, since the C++ language
specifies the behavior of the overflow. That's also the reason why this
code is implemented only for unsigned integers.

For the multiplication, if the compiler does not support widening
multiplications, we do it with a division instead. This is necessary for
GCC < 4.5 and compilers not compatible with GCC or MSVC.

Change-Id: I049a653beeb5454c9539ffff13e637de0f1338c1
Reviewed-by: Ulf Hermann <ulf.hermann@theqtcompany.com>
Reviewed-by: Alex Trotsenko <alex1973tr@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

2 files changed, 161 insertions, 1 deletions

diff --git a/tests/auto/corelib/global/qnumeric/qnumeric.pro b/tests/auto/corelib/global/qnumeric/qnumeric.pro
index 00f3635be9..0772ce6aab 100644
--- a/tests/auto/corelib/global/qnumeric/qnumeric.pro
+++ b/tests/auto/corelib/global/qnumeric/qnumeric.pro
@@ -1,6 +1,6 @@
 CONFIG += testcase parallel_test
 TARGET = tst_qnumeric
-QT = core testlib
+QT = core-private testlib
 SOURCES = tst_qnumeric.cpp
 DEFINES += QT_DISABLE_DEPRECATED_BEFORE=0
 intel_icc: QMAKE_CXXFLAGS += -fp-model strict
diff --git a/tests/auto/corelib/global/qnumeric/tst_qnumeric.cpp b/tests/auto/corelib/global/qnumeric/tst_qnumeric.cpp
index fdc8bc6aab..59a536ed25 100644
--- a/tests/auto/corelib/global/qnumeric/tst_qnumeric.cpp
+++ b/tests/auto/corelib/global/qnumeric/tst_qnumeric.cpp
@@ -34,6 +34,7 @@
 
 #include <QtTest/QtTest>
 #include <QtGlobal>
+#include "private/qnumeric_p.h"
 
 #include <math.h>
 #include <float.h>
@@ -50,6 +51,10 @@ private slots:
     void floatDistance();
     void floatDistance_double_data();
     void floatDistance_double();
+    void addOverflow_data();
+    void addOverflow();
+    void mulOverflow_data();
+    void mulOverflow();
 };
 
 void tst_QNumeric::fuzzyCompare_data()
@@ -206,5 +211,160 @@ void tst_QNumeric::floatDistance_double()
     QCOMPARE(qFloatDistance(val1, val2), expectedDistance);
 }
 
+void tst_QNumeric::addOverflow_data()
+{
+    QTest::addColumn<int>("size");
+    QTest::newRow("quint8") << 8;
+    QTest::newRow("quint16") << 16;
+    QTest::newRow("quint32") << 32;
+    QTest::newRow("quint64") << 64;
+    QTest::newRow("ulong") << 48;   // it's either 32- or 64-bit, so on average it's 48 :-)
+}
+
+// Note: in release mode, all the tests may be statically determined and only the calls
+// to QTest::toString and QTest::qCompare will remain.
+template <typename Int> static void addOverflow_template()
+{
+#if defined(Q_CC_MSVC) && Q_CC_MSVC < 1900
+    QSKIP("Test disabled, this test generates an Internal Compiler Error compiling");
+#else
+    const Int max = std::numeric_limits<Int>::max();
+    Int r;
+
+    // basic values
+    QCOMPARE(add_overflow(Int(0), Int(0), &r), false);
+    QCOMPARE(r, Int(0));
+    QCOMPARE(add_overflow(Int(1), Int(0), &r), false);
+    QCOMPARE(r, Int(1));
+    QCOMPARE(add_overflow(Int(0), Int(1), &r), false);
+    QCOMPARE(r, Int(1));
+
+    // half-way through max
+    QCOMPARE(add_overflow(Int(max/2), Int(max/2), &r), false);
+    QCOMPARE(r, Int(max / 2 * 2));
+    QCOMPARE(add_overflow(Int(max/2 - 1), Int(max/2 + 1), &r), false);
+    QCOMPARE(r, Int(max / 2 * 2));
+    QCOMPARE(add_overflow(Int(max/2 + 1), Int(max/2), &r), false);
+    QCOMPARE(r, max);
+    QCOMPARE(add_overflow(Int(max/2), Int(max/2 + 1), &r), false);
+    QCOMPARE(r, max);
+
+    // more than half
+    QCOMPARE(add_overflow(Int(max/4 * 3), Int(max/4), &r), false);
+    QCOMPARE(r, Int(max / 4 * 4));
+
+    // max
+    QCOMPARE(add_overflow(max, Int(0), &r), false);
+    QCOMPARE(r, max);
+    QCOMPARE(add_overflow(Int(0), max, &r), false);
+    QCOMPARE(r, max);
+
+    // 64-bit issues
+    if (max > std::numeric_limits<uint>::max()) {
+        QCOMPARE(add_overflow(Int(std::numeric_limits<uint>::max()), Int(std::numeric_limits<uint>::max()), &r), false);
+        QCOMPARE(r, Int(2 * Int(std::numeric_limits<uint>::max())));
+    }
+
+    // overflows
+    QCOMPARE(add_overflow(max, Int(1), &r), true);
+    QCOMPARE(add_overflow(Int(1), max, &r), true);
+    QCOMPARE(add_overflow(Int(max/2 + 1), Int(max/2 + 1), &r), true);
+#endif
+}
+
+void tst_QNumeric::addOverflow()
+{
+    QFETCH(int, size);
+    if (size == 8)
+        addOverflow_template<quint8>();
+    if (size == 16)
+        addOverflow_template<quint16>();
+    if (size == 32)
+        addOverflow_template<quint32>();
+    if (size == 48)
+        addOverflow_template<ulong>();  // not really 48-bit
+    if (size == 64)
+        addOverflow_template<quint64>();
+}
+
+void tst_QNumeric::mulOverflow_data()
+{
+    addOverflow_data();
+}
+
+// Note: in release mode, all the tests may be statically determined and only the calls
+// to QTest::toString and QTest::qCompare will remain.
+template <typename Int> static void mulOverflow_template()
+{
+#if defined(Q_CC_MSVC) && Q_CC_MSVC < 1900
+    QSKIP("Test disabled, this test generates an Internal Compiler Error compiling");
+#else
+    const Int max = std::numeric_limits<Int>::max();
+    const Int middle = Int(max >> (sizeof(Int) * CHAR_BIT / 2));
+    Int r;
+
+    // basic multiplications
+    QCOMPARE(mul_overflow(Int(0), Int(0), &r), false);
+    QCOMPARE(r, Int(0));
+    QCOMPARE(mul_overflow(Int(1), Int(0), &r), false);
+    QCOMPARE(r, Int(0));
+    QCOMPARE(mul_overflow(Int(0), Int(1), &r), false);
+    QCOMPARE(r, Int(0));
+    QCOMPARE(mul_overflow(max, Int(0), &r), false);
+    QCOMPARE(r, Int(0));
+    QCOMPARE(mul_overflow(Int(0), max, &r), false);
+    QCOMPARE(r, Int(0));
+
+    QCOMPARE(mul_overflow(Int(1), Int(1), &r), false);
+    QCOMPARE(r, Int(1));
+    QCOMPARE(mul_overflow(Int(1), max, &r), false);
+    QCOMPARE(r, max);
+    QCOMPARE(mul_overflow(max, Int(1), &r), false);
+    QCOMPARE(r, max);
+
+    // almost max
+    QCOMPARE(mul_overflow(middle, middle, &r), false);
+    QCOMPARE(r, Int(max - 2 * middle));
+    QCOMPARE(mul_overflow(Int(middle + 1), middle, &r), false);
+    QCOMPARE(r, Int(middle << (sizeof(Int) * CHAR_BIT / 2)));
+    QCOMPARE(mul_overflow(middle, Int(middle + 1), &r), false);
+    QCOMPARE(r, Int(middle << (sizeof(Int) * CHAR_BIT / 2)));
+    QCOMPARE(mul_overflow(Int(max / 2), Int(2), &r), false);
+    QCOMPARE(r, Int(max & ~Int(1)));
+    QCOMPARE(mul_overflow(Int(max / 4), Int(4), &r), false);
+    QCOMPARE(r, Int(max & ~Int(3)));
+
+    // overflows
+    QCOMPARE(mul_overflow(max, Int(2), &r), true);
+    QCOMPARE(mul_overflow(Int(max / 2), Int(3), &r), true);
+    QCOMPARE(mul_overflow(Int(middle + 1), Int(middle + 1), &r), true);
+#endif
+}
+
+template <typename Int, bool enabled = sizeof(Int) <= sizeof(void*)> struct MulOverflowDispatch;
+template <typename Int> struct MulOverflowDispatch<Int, true>
+{
+    void operator()() { mulOverflow_template<Int>(); }
+};
+template <typename Int> struct MulOverflowDispatch<Int, false>
+{
+    void operator()() { QSKIP("This type is too big for this architecture"); }
+};
+
+void tst_QNumeric::mulOverflow()
+{
+    QFETCH(int, size);
+    if (size == 8)
+        MulOverflowDispatch<quint8>()();
+    if (size == 16)
+        MulOverflowDispatch<quint16>()();
+    if (size == 32)
+        MulOverflowDispatch<quint32>()();
+    if (size == 48)
+        MulOverflowDispatch<ulong>()();     // not really 48-bit
+    if (size == 64)
+        MulOverflowDispatch<quint64>()();
+}
+
 QTEST_APPLESS_MAIN(tst_QNumeric)
 #include "tst_qnumeric.moc"