Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(519)

Side by Side Diff: base/numerics/safe_math_impl.h

Issue 141583008: Add support for safe math operations in base/numerics (Closed) Base URL: svn://svn.chromium.org/chrome/trunk/src/
Patch Set: style fixes Created 6 years, 10 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View unified diff | Download patch | Annotate | Revision Log
OLDNEW
(Empty)
1 // Copyright 2014 The Chromium Authors. All rights reserved.
2 // Use of this source code is governed by a BSD-style license that can be
3 // found in the LICENSE file.
4
5 #ifndef SAFE_MATH_IMPL_H_
6 #define SAFE_MATH_IMPL_H_
7
8 #include <stdint.h>
9
10 #include <cmath>
11 #include <cstdlib>
12 #include <limits>
13
14 #include "base/compiler_specific.h"
15 #include "base/macros.h"
16 #include "base/template_util.h"
17
18 namespace base {
19 namespace internal {
20
21 using std::numeric_limits;
22
23 // Everything from here up to the floating point operations is portable C++,
24 // but it may not be fast. This code could be split based on
25 // platform/architecture and replaced with potentially faster implementations.
26
27 // Integer promotion templates used by the portable checked integer arithmetic.
28 template <size_t Size, bool IsSigned>
29 struct IntegerForSizeAndSign {};
awong 2014/02/11 20:14:12 Here and elsewhere, don't provide a definition for
jschuh 2014/02/21 19:22:28 Done.
30 template <>
31 struct IntegerForSizeAndSign<1, true> {
32 typedef int8_t type;
33 };
34 template <>
35 struct IntegerForSizeAndSign<1, false> {
36 typedef uint8_t type;
37 };
38 template <>
39 struct IntegerForSizeAndSign<2, true> {
40 typedef int16_t type;
41 };
42 template <>
43 struct IntegerForSizeAndSign<2, false> {
44 typedef uint16_t type;
45 };
46 template <>
47 struct IntegerForSizeAndSign<4, true> {
48 typedef int32_t type;
49 };
50 template <>
51 struct IntegerForSizeAndSign<4, false> {
52 typedef uint32_t type;
53 };
54 template <>
55 struct IntegerForSizeAndSign<8, true> {
56 typedef int64_t type;
57 };
58 template <>
59 struct IntegerForSizeAndSign<8, false> {
60 typedef uint64_t type;
61 };
62 // The ArithmeticPromotion template below will need to be updated (or more
awong 2014/02/11 20:14:12 This looks like a TODO. Make it into a TODO? Also
jschuh 2014/02/21 19:22:28 We don't support 128-bit math today, so it's a war
63 // likely replaced with decltype) if we add support for 128-bit integers).
64
65 template <typename Integer>
66 struct UnsignedIntegerForSize {
67 typedef typename enable_if<
awong 2014/02/11 20:14:12 Why do we need an enable_if here? Normally you use
jschuh 2014/02/21 19:22:28 I'm preventing this: UnsignedIntegerForSize<flo
68 numeric_limits<Integer>::is_integer,
69 typename IntegerForSizeAndSign<sizeof(Integer), false>::type>::type type;
70 };
71
72 template <typename Integer>
73 struct SignedIntegerForSize {
74 typedef typename enable_if<
75 numeric_limits<Integer>::is_integer,
76 typename IntegerForSizeAndSign<sizeof(Integer), true>::type>::type type;
77 };
78
79 template <typename Integer>
80 struct TwiceWiderInteger {
81 typedef typename enable_if<
82 numeric_limits<Integer>::is_integer,
83 typename IntegerForSizeAndSign<
84 sizeof(Integer) * 2,
85 numeric_limits<Integer>::is_signed>::type>::type type;
86 };
87
88 template <typename Integer>
89 struct PositionOfSignBit {
90 static const typename enable_if<numeric_limits<Integer>::is_integer,
91 size_t>::type value = 8 * sizeof(Integer) - 1;
92 };
93
94 // Helper templates for integer manipulations.
95
96 template <typename T>
97 bool HasSignBit(T x) {
98 // Cast to unsigned since right shift on signed is undefined.
99 return !!(static_cast<typename UnsignedIntegerForSize<T>::type>(x) >>
100 PositionOfSignBit<T>::value);
101 }
102
103 // This wrapper undoes the standard integer promotions.
104 template <typename T>
105 T BinaryComplement(T x) {
106 return ~x;
107 }
108
109 // Here are the actual portable checked integer implementations.
110
111 template <typename T>
112 typename enable_if<numeric_limits<T>::is_integer, T>::type
113 CheckedAdd(T x, T y, RangeCheckId* validity) {
114 // Since the value of x+y is undefined if we have a signed type, we compute
115 // it using the unsigned type of the same size.
116 typedef typename UnsignedIntegerForSize<T>::type UnsignedDst;
117 UnsignedDst ux = static_cast<UnsignedDst>(x);
118 UnsignedDst uy = static_cast<UnsignedDst>(y);
119 UnsignedDst uresult = ux + uy;
120 // Addition is valid if the sign of (x + y) is equal to either that of x or
121 // that of y.
122 if (numeric_limits<T>::is_signed) {
123 if (HasSignBit(BinaryComplement((uresult ^ ux) & (uresult ^ uy))))
124 *validity = TYPE_VALID;
125 else // Direction of wrap is inverse of result sign.
126 *validity = HasSignBit(uresult) ? TYPE_OVERFLOW : TYPE_UNDERFLOW;
127
128 } else { // Unsigned is either valid or overflow.
129 *validity = BinaryComplement(x) >= y ? TYPE_VALID : TYPE_OVERFLOW;
130 }
131 return static_cast<T>(uresult);
132 }
133
134 template <typename T>
135 typename enable_if<numeric_limits<T>::is_integer, T>::type
136 CheckedSub(T x, T y, RangeCheckId* validity) {
137 // Since the value of x+y is undefined if we have a signed type, we compute
138 // it using the unsigned type of the same size.
139 typedef typename UnsignedIntegerForSize<T>::type UnsignedDst;
140 UnsignedDst ux = static_cast<UnsignedDst>(x);
141 UnsignedDst uy = static_cast<UnsignedDst>(y);
142 UnsignedDst uresult = ux - uy;
143 // Subtraction is valid if either x and y have same sign, or (x-y) and x have
144 // the same sign.
145 if (numeric_limits<T>::is_signed) {
146 if (HasSignBit(BinaryComplement((uresult ^ ux) & (ux ^ uy))))
147 *validity = TYPE_VALID;
148 else // Direction of wrap is inverse of result sign.
149 *validity = HasSignBit(uresult) ? TYPE_OVERFLOW : TYPE_UNDERFLOW;
150
151 } else { // Unsigned is either valid or underflow.
152 *validity = x >= y ? TYPE_VALID : TYPE_UNDERFLOW;
153 }
154 return static_cast<T>(uresult);
155 }
156
157 // Integer multiplication is a bit complicated. In the fast case we just
158 // we just promote to a twice wider type, and range check the result. In the
159 // slow case we need to manually check that the result won't be truncated by
160 // checking with division against the appropriate bound.
161 template <typename T>
162 typename enable_if<
163 numeric_limits<T>::is_integer && sizeof(T) * 2 <= sizeof(uintmax_t),
164 T>::type
165 CheckedMul(T x, T y, RangeCheckId* validity) {
166 typedef typename TwiceWiderInteger<T>::type IntermediateType;
167 IntermediateType tmp =
168 static_cast<IntermediateType>(x) * static_cast<IntermediateType>(y);
169 *validity = RangeCheck<T>(tmp);
170 return static_cast<T>(tmp);
171 }
172
173 template <typename T>
174 typename enable_if<numeric_limits<T>::is_integer&& numeric_limits<
awong 2014/02/11 20:14:12 Yes...it's looking like you're using enable_if to
jschuh 2014/02/21 19:22:28 This is breaking out the specializations. Based on
175 T>::is_signed&&(sizeof(T) * 2 > sizeof(uintmax_t)),
176 T>::type
177 CheckedMul(T x, T y, RangeCheckId* validity) {
178 // if either side is zero then the result will be zero.
179 if (!(x || y)) {
180 return TYPE_VALID;
181
182 } else if (x > 0) {
183 if (y > 0)
184 *validity =
185 x <= numeric_limits<T>::max() / y ? TYPE_VALID : TYPE_OVERFLOW;
186 else
187 *validity =
188 y >= numeric_limits<T>::min() / x ? TYPE_VALID : TYPE_UNDERFLOW;
189
190 } else {
191 if (y > 0)
192 *validity =
193 x >= numeric_limits<T>::min() / y ? TYPE_VALID : TYPE_UNDERFLOW;
194 else
195 *validity =
196 y >= numeric_limits<T>::max() / x ? TYPE_VALID : TYPE_OVERFLOW;
197 }
198
199 return x * y;
200 }
201
202 template <typename T>
203 typename enable_if<numeric_limits<T>::is_integer &&
204 !numeric_limits<T>::is_signed &&
205 (sizeof(T) * 2 > sizeof(uintmax_t)),
206 T>::type
207 CheckedMul(T x, T y, RangeCheckId* validity) {
208 *validity = (y == 0 || x <= numeric_limits<T>::max() / y) ? TYPE_VALID
209 : TYPE_OVERFLOW;
210 return x * y;
211 }
212
213 // Division just requires a check for an invalid negation on signed min/-1.
214 template <typename T>
215 T CheckedDiv(T x,
216 T y,
217 RangeCheckId* validity,
218 typename enable_if<numeric_limits<T>::is_integer, int>::type = 0) {
219 if (numeric_limits<T>::is_signed && x == numeric_limits<T>::min() &&
220 y == static_cast<T>(-1)) {
221 *validity = TYPE_OVERFLOW;
222 return numeric_limits<T>::min();
223 }
224
225 *validity = TYPE_VALID;
226 return x / y;
227 }
228
229 template <typename T>
230 typename enable_if<numeric_limits<T>::is_integer&& numeric_limits<T>::is_signed,
231 T>::type
232 CheckedMod(T x, T y, RangeCheckId* validity) {
233 *validity = y > 0 ? TYPE_VALID : TYPE_INVALID;
234 return x % y;
235 }
236
237 template <typename T>
238 typename enable_if<
239 numeric_limits<T>::is_integer && !numeric_limits<T>::is_signed,
240 T>::type
241 CheckedMod(T x, T y, RangeCheckId* validity) {
242 *validity = TYPE_VALID;
243 return x % y;
244 }
245
246 template <typename T>
247 typename enable_if<numeric_limits<T>::is_integer&& numeric_limits<T>::is_signed,
248 T>::type
249 CheckedNeg(T value, RangeCheckId* validity) {
250 *validity = value != numeric_limits<T>::min() ? TYPE_VALID : TYPE_OVERFLOW;
251 // The negation of signed min is min, so catch that one.
252 return -value;
253 }
254
255 template <typename T>
256 typename enable_if<
257 numeric_limits<T>::is_integer && !numeric_limits<T>::is_signed,
258 T>::type
259 CheckedNeg(T value, RangeCheckId* validity) {
260 // The only legal unsigned negation is zero.
261 *validity = value ? TYPE_UNDERFLOW : TYPE_VALID;
262 return static_cast<T>(
263 -static_cast<typename SignedIntegerForSize<T>::type>(value));
264 }
265
266 template <typename T>
267 typename enable_if<numeric_limits<T>::is_integer&& numeric_limits<T>::is_signed,
268 T>::type
269 CheckedAbs(T value, RangeCheckId* validity) {
270 *validity = value != numeric_limits<T>::min() ? TYPE_VALID : TYPE_OVERFLOW;
271 return std::abs(value);
272 }
273
274 template <typename T>
275 typename enable_if<
276 numeric_limits<T>::is_integer && !numeric_limits<T>::is_signed,
277 T>::type
278 CheckedAbs(T value, RangeCheckId* validity) {
279 // Absolute value of a positive is just its identiy.
280 *validity = TYPE_VALID;
281 return value;
282 }
283
284 // These are the floating point stubs that the compiler needs to see. Only the
285 // negation operation is ever called.
286 #define BASE_FLOAT_ARITHMETIC_STUBS(NAME) \
287 template <typename T> \
288 typename enable_if<numeric_limits<T>::is_iec559, T>::type Checked##NAME( \
289 T, T, RangeCheckId*) { \
290 NOTREACHED(); \
291 return 0; \
292 }
293
294 BASE_FLOAT_ARITHMETIC_STUBS(Add)
295 BASE_FLOAT_ARITHMETIC_STUBS(Sub)
296 BASE_FLOAT_ARITHMETIC_STUBS(Mul)
297 BASE_FLOAT_ARITHMETIC_STUBS(Div)
298 BASE_FLOAT_ARITHMETIC_STUBS(Mod)
299
300 #undef BASE_FLOAT_ARITHMETIC_STUBS
301
302 template <typename T>
303 typename enable_if<numeric_limits<T>::is_iec559, T>::type CheckedNeg(
304 T value,
305 RangeCheckId*) {
306 return -value;
307 }
308
309 template <typename T>
310 typename enable_if<numeric_limits<T>::is_iec559, T>::type CheckedAbs(
311 T value,
312 RangeCheckId*) {
313 return std::abs(value);
314 }
315
316 // Floats carry around their validity state with them, but integers do not. So,
317 // we wrap the underlying value in a specialization in order to hide that detail
318 // and expose an interface via accessors.
319 enum NumericTypeId { NUMERIC_INTEGER, NUMERIC_FLOATING, NUMERIC_UNKNOWN };
320
321 template <typename NumericType>
322 struct GetNumericTypeId {
323 static const NumericTypeId value =
324 numeric_limits<NumericType>::is_integer
325 ? NUMERIC_INTEGER
326 : (numeric_limits<NumericType>::is_iec559 ? NUMERIC_FLOATING
327 : NUMERIC_UNKNOWN);
328 };
329
330 template <typename T, NumericTypeId type = GetNumericTypeId<T>::value>
331 class CheckedNumericState {};
332
333 // Integrals require quite a bit of additional housekeeping to manage state.
334 template <typename T>
335 class CheckedNumericState<T, NUMERIC_INTEGER> {
336 private:
337 T value_;
338 RangeCheckId validity_;
339
340 public:
341 template <typename Src, NumericTypeId type>
342 friend class CheckedNumericState;
343
344 CheckedNumericState() : value_(0), validity_(TYPE_VALID) {}
345
346 template <typename Src>
347 CheckedNumericState(Src value, RangeCheckId validity)
348 : value_(value),
349 validity_(static_cast<RangeCheckId>(validity | RangeCheck<T>(value))) {
350 COMPILE_ASSERT(numeric_limits<Src>::is_specialized,
351 argument_must_be_numeric);
352 }
353
354 // Copy constructor.
355 template <typename Src>
356 CheckedNumericState(const CheckedNumericState<Src>& rhs)
357 : value_(static_cast<T>(rhs.value())),
358 validity_(static_cast<RangeCheckId>(rhs.validity() |
359 RangeCheck<T>(rhs.value()))) {}
360
361 template <typename Src>
362 explicit CheckedNumericState(
363 Src value,
364 typename enable_if<numeric_limits<Src>::is_specialized, int>::type = 0)
365 : value_(static_cast<T>(value)), validity_(RangeCheck<T>(value)) {}
366
367 RangeCheckId validity() const { return validity_; }
368 T value() const { return value_; }
369 };
370
371 // Floating points maintain their own validity, but need translation wrappers.
372 template <typename T>
373 class CheckedNumericState<T, NUMERIC_FLOATING> {
374 private:
375 T value_;
376
377 public:
378 template <typename Src, NumericTypeId type>
379 friend class CheckedNumericState;
380
381 CheckedNumericState() : value_(0.0) {}
382
383 template <typename Src>
384 CheckedNumericState(
385 Src value,
386 RangeCheckId validity,
387 typename enable_if<numeric_limits<Src>::is_integer, int>::type = 0) {
388 switch (RangeCheck<T>(value)) {
389 case TYPE_VALID:
390 value_ = static_cast<T>(value);
391 break;
392
393 case TYPE_UNDERFLOW:
394 value_ = -numeric_limits<T>::infinity();
395 break;
396
397 case TYPE_OVERFLOW:
398 value_ = numeric_limits<T>::infinity();
399 break;
400
401 case TYPE_INVALID:
402 value_ = numeric_limits<T>::quiet_NaN();
403 break;
404
405 default:
406 NOTREACHED();
407 }
408 }
409
410 template <typename Src>
411 explicit CheckedNumericState(
412 Src value,
413 typename enable_if<numeric_limits<Src>::is_specialized, int>::type = 0)
414 : value_(static_cast<T>(value)) {}
415
416 // Copy constructor.
417 template <typename Src>
418 CheckedNumericState(const CheckedNumericState<Src>& rhs)
419 : value_(static_cast<T>(rhs.value())) {}
420
421 RangeCheckId validity() const {
422 return BASE_NUMERIC_RANGE_CHECK_RESULT(value_ <= numeric_limits<T>::max(),
423 value_ >= -numeric_limits<T>::max());
424 }
425 T value() const { return value_; }
426 };
427
428 // For integers less than 128-bit and floats 32-bit or larger, we can distil
429 // C/C++ arithmetic promotions down to two simple rules:
430 // 1. The type with the larger maximum exponent always takes precedence.
431 // 2. The resulting type must be promoted to at least an int.
432 // The following template specializations implement that promotion logic.
433 enum ArithmeticPromotionId {
434 LEFT_PROMOTION,
435 RIGHT_PROMOTION,
436 DEFAULT_PROMOTION
437 };
438
439 template <typename Lhs,
440 typename Rhs = Lhs,
441 ArithmeticPromotionId Promotion =
442 (MaxExponent<Lhs>::value > MaxExponent<Rhs>::value)
443 ? (MaxExponent<Lhs>::value > MaxExponent<int>::value
444 ? LEFT_PROMOTION
445 : DEFAULT_PROMOTION)
446 : (MaxExponent<Rhs>::value > MaxExponent<int>::value
447 ? RIGHT_PROMOTION
448 : DEFAULT_PROMOTION) > struct ArithmeticPromotion {};
449
450 template <typename Lhs, typename Rhs>
451 struct ArithmeticPromotion<Lhs, Rhs, LEFT_PROMOTION> {
452 typedef Lhs type;
453 };
454
455 template <typename Lhs, typename Rhs>
456 struct ArithmeticPromotion<Lhs, Rhs, RIGHT_PROMOTION> {
457 typedef Rhs type;
458 };
459
460 template <typename Lhs, typename Rhs>
461 struct ArithmeticPromotion<Lhs, Rhs, DEFAULT_PROMOTION> {
462 typedef int type;
463 };
464
465 // We can statically check if operations on the provided types can wrap, so we
466 // can skip the checked operations if they're not needed. So, for an integer we
467 // care if the destination type preserves the sign and is twice the width of
468 // the source.
469 template <typename T, typename Lhs, typename Rhs>
470 struct IsIntegerArithmeticSafe {
471 static const bool value = !numeric_limits<T>::is_iec559 &&
472 StaticRangeCheck<T, Lhs>::value == CONTAINS_RANGE &&
473 sizeof(T) >= (2 * sizeof(Lhs)) &&
474 StaticRangeCheck<T, Rhs>::value != CONTAINS_RANGE &&
475 sizeof(T) >= (2 * sizeof(Rhs));
476 };
477
478 } // namespace internal
479 } // namespace base
480
481 #endif // SAFE_MATH_IMPL_H_
OLDNEW

Powered by Google App Engine
This is Rietveld 408576698