gh-112567: Add _PyTimeFraction C API (#112568)

Use a fraction internally in the _PyTime API to reduce the risk of
integer overflow: simplify the fraction using Greatest Common
Divisor (GCD). The fraction API is used by time functions:
perf_counter(), monotonic() and process_time().

For example, QueryPerformanceFrequency() usually returns 10 MHz on
Windows 10 and newer. The fraction SEC_TO_NS / frequency =
1_000_000_000 / 10_000_000 can be simplified to 100 / 1.

* Add _PyTimeFraction type.
* Add functions:

  * _PyTimeFraction_Set()
  * _PyTimeFraction_Mul()
  * _PyTimeFraction_Resolution()

* No longer check "numer * denom <= _PyTime_MAX" in
  _PyTimeFraction_Set(). _PyTimeFraction_Mul() uses _PyTime_Mul()
  which handles integer overflow.
This commit is contained in:
Victor Stinner 2023-12-01 19:50:10 +01:00 committed by GitHub
parent 05a370abd6
commit 5c5022b862
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 130 additions and 107 deletions

View File

@ -253,13 +253,6 @@ PyAPI_FUNC(void) _PyTime_AsTimespec_clamp(_PyTime_t t, struct timespec *ts);
// Compute t1 + t2. Clamp to [_PyTime_MIN; _PyTime_MAX] on overflow. // Compute t1 + t2. Clamp to [_PyTime_MIN; _PyTime_MAX] on overflow.
extern _PyTime_t _PyTime_Add(_PyTime_t t1, _PyTime_t t2); extern _PyTime_t _PyTime_Add(_PyTime_t t1, _PyTime_t t2);
// Compute ticks * mul / div.
// Clamp to [_PyTime_MIN; _PyTime_MAX] on overflow.
// The caller must ensure that ((div - 1) * mul) cannot overflow.
extern _PyTime_t _PyTime_MulDiv(_PyTime_t ticks,
_PyTime_t mul,
_PyTime_t div);
// Structure used by time.get_clock_info() // Structure used by time.get_clock_info()
typedef struct { typedef struct {
const char *implementation; const char *implementation;
@ -355,6 +348,32 @@ PyAPI_FUNC(_PyTime_t) _PyDeadline_Init(_PyTime_t timeout);
PyAPI_FUNC(_PyTime_t) _PyDeadline_Get(_PyTime_t deadline); PyAPI_FUNC(_PyTime_t) _PyDeadline_Get(_PyTime_t deadline);
// --- _PyTimeFraction -------------------------------------------------------
typedef struct {
_PyTime_t numer;
_PyTime_t denom;
} _PyTimeFraction;
// Set a fraction.
// Return 0 on success.
// Return -1 if the fraction is invalid.
extern int _PyTimeFraction_Set(
_PyTimeFraction *frac,
_PyTime_t numer,
_PyTime_t denom);
// Compute ticks * frac.numer / frac.denom.
// Clamp to [_PyTime_MIN; _PyTime_MAX] on overflow.
extern _PyTime_t _PyTimeFraction_Mul(
_PyTime_t ticks,
const _PyTimeFraction *frac);
// Compute a clock resolution: frac.numer / frac.denom / 1e9.
extern double _PyTimeFraction_Resolution(
const _PyTimeFraction *frac);
#ifdef __cplusplus #ifdef __cplusplus
} }
#endif #endif

View File

@ -69,25 +69,6 @@ module time
/*[clinic end generated code: output=da39a3ee5e6b4b0d input=a668a08771581f36]*/ /*[clinic end generated code: output=da39a3ee5e6b4b0d input=a668a08771581f36]*/
#if defined(HAVE_TIMES) || defined(HAVE_CLOCK)
static int
check_ticks_per_second(long tps, const char *context)
{
/* Effectively, check that _PyTime_MulDiv(t, SEC_TO_NS, tps)
cannot overflow. */
if (tps >= 0 && (_PyTime_t)tps > _PyTime_MAX / SEC_TO_NS) {
PyErr_Format(PyExc_OverflowError, "%s is too large", context);
return -1;
}
if (tps < 1) {
PyErr_Format(PyExc_RuntimeError, "invalid %s", context);
return -1;
}
return 0;
}
#endif /* HAVE_TIMES || HAVE_CLOCK */
/* Forward declarations */ /* Forward declarations */
static int pysleep(_PyTime_t timeout); static int pysleep(_PyTime_t timeout);
@ -96,11 +77,11 @@ typedef struct {
PyTypeObject *struct_time_type; PyTypeObject *struct_time_type;
#ifdef HAVE_TIMES #ifdef HAVE_TIMES
// times() clock frequency in hertz // times() clock frequency in hertz
long ticks_per_second; _PyTimeFraction times_base;
#endif #endif
#ifdef HAVE_CLOCK #ifdef HAVE_CLOCK
// clock() frequency in hertz // clock() frequency in hertz
long clocks_per_second; _PyTimeFraction clock_base;
#endif #endif
} time_module_state; } time_module_state;
@ -174,10 +155,11 @@ Return the current time in nanoseconds since the Epoch.");
static int static int
py_clock(time_module_state *state, _PyTime_t *tp, _Py_clock_info_t *info) py_clock(time_module_state *state, _PyTime_t *tp, _Py_clock_info_t *info)
{ {
long clocks_per_second = state->clocks_per_second; _PyTimeFraction *base = &state->clock_base;
if (info) { if (info) {
info->implementation = "clock()"; info->implementation = "clock()";
info->resolution = 1.0 / (double)clocks_per_second; info->resolution = _PyTimeFraction_Resolution(base);
info->monotonic = 1; info->monotonic = 1;
info->adjustable = 0; info->adjustable = 0;
} }
@ -189,7 +171,7 @@ py_clock(time_module_state *state, _PyTime_t *tp, _Py_clock_info_t *info)
"or its value cannot be represented"); "or its value cannot be represented");
return -1; return -1;
} }
_PyTime_t ns = _PyTime_MulDiv(ticks, SEC_TO_NS, clocks_per_second); _PyTime_t ns = _PyTimeFraction_Mul(ticks, base);
*tp = _PyTime_FromNanoseconds(ns); *tp = _PyTime_FromNanoseconds(ns);
return 0; return 0;
} }
@ -1257,7 +1239,7 @@ static int
process_time_times(time_module_state *state, _PyTime_t *tp, process_time_times(time_module_state *state, _PyTime_t *tp,
_Py_clock_info_t *info) _Py_clock_info_t *info)
{ {
long ticks_per_second = state->ticks_per_second; _PyTimeFraction *base = &state->times_base;
struct tms process; struct tms process;
if (times(&process) == (clock_t)-1) { if (times(&process) == (clock_t)-1) {
@ -1266,14 +1248,14 @@ process_time_times(time_module_state *state, _PyTime_t *tp,
if (info) { if (info) {
info->implementation = "times()"; info->implementation = "times()";
info->resolution = _PyTimeFraction_Resolution(base);
info->monotonic = 1; info->monotonic = 1;
info->adjustable = 0; info->adjustable = 0;
info->resolution = 1.0 / (double)ticks_per_second;
} }
_PyTime_t ns; _PyTime_t ns;
ns = _PyTime_MulDiv(process.tms_utime, SEC_TO_NS, ticks_per_second); ns = _PyTimeFraction_Mul(process.tms_utime, base);
ns += _PyTime_MulDiv(process.tms_stime, SEC_TO_NS, ticks_per_second); ns += _PyTimeFraction_Mul(process.tms_stime, base);
*tp = _PyTime_FromNanoseconds(ns); *tp = _PyTime_FromNanoseconds(ns);
return 1; return 1;
} }
@ -1395,8 +1377,7 @@ py_process_time(time_module_state *state, _PyTime_t *tp,
// times() failed, ignore failure // times() failed, ignore failure
#endif #endif
/* clock */ /* clock(). Python 3 requires clock() to build (see gh-66814) */
/* Currently, Python 3 requires clock() to build: see issue #22624 */
return py_clock(state, tp, info); return py_clock(state, tp, info);
#endif #endif
} }
@ -2110,20 +2091,23 @@ time_exec(PyObject *module)
#endif #endif
#ifdef HAVE_TIMES #ifdef HAVE_TIMES
if (_Py_GetTicksPerSecond(&state->ticks_per_second) < 0) { long ticks_per_second;
if (_Py_GetTicksPerSecond(&ticks_per_second) < 0) {
PyErr_SetString(PyExc_RuntimeError, PyErr_SetString(PyExc_RuntimeError,
"cannot read ticks_per_second"); "cannot read ticks_per_second");
return -1; return -1;
} }
if (_PyTimeFraction_Set(&state->times_base, SEC_TO_NS,
if (check_ticks_per_second(state->ticks_per_second, "_SC_CLK_TCK") < 0) { ticks_per_second) < 0) {
PyErr_Format(PyExc_OverflowError, "ticks_per_second is too large");
return -1; return -1;
} }
#endif #endif
#ifdef HAVE_CLOCK #ifdef HAVE_CLOCK
state->clocks_per_second = CLOCKS_PER_SEC; if (_PyTimeFraction_Set(&state->clock_base, SEC_TO_NS,
if (check_ticks_per_second(state->clocks_per_second, "CLOCKS_PER_SEC") < 0) { CLOCKS_PER_SEC) < 0) {
PyErr_Format(PyExc_OverflowError, "CLOCKS_PER_SEC is too large");
return -1; return -1;
} }
#endif #endif

View File

@ -55,6 +55,43 @@
#endif #endif
static _PyTime_t
_PyTime_GCD(_PyTime_t x, _PyTime_t y)
{
// Euclidean algorithm
assert(x >= 1);
assert(y >= 1);
while (y != 0) {
_PyTime_t tmp = y;
y = x % y;
x = tmp;
}
assert(x >= 1);
return x;
}
int
_PyTimeFraction_Set(_PyTimeFraction *frac, _PyTime_t numer, _PyTime_t denom)
{
if (numer < 1 || denom < 1) {
return -1;
}
_PyTime_t gcd = _PyTime_GCD(numer, denom);
frac->numer = numer / gcd;
frac->denom = denom / gcd;
return 0;
}
double
_PyTimeFraction_Resolution(const _PyTimeFraction *frac)
{
return (double)frac->numer / (double)frac->denom / 1e9;
}
static void static void
pytime_time_t_overflow(void) pytime_time_t_overflow(void)
{ {
@ -152,11 +189,17 @@ _PyTime_Mul(_PyTime_t t, _PyTime_t k)
} }
_PyTime_t _PyTime_t
_PyTime_MulDiv(_PyTime_t ticks, _PyTime_t mul, _PyTime_t div) _PyTimeFraction_Mul(_PyTime_t ticks, const _PyTimeFraction *frac)
{ {
const _PyTime_t mul = frac->numer;
const _PyTime_t div = frac->denom;
if (div == 1) {
// Fast-path taken by mach_absolute_time() with 1/1 time base.
return _PyTime_Mul(ticks, mul);
}
/* Compute (ticks * mul / div) in two parts to reduce the risk of integer /* Compute (ticks * mul / div) in two parts to reduce the risk of integer
overflow: compute the integer part, and then the remaining part. overflow: compute the integer part, and then the remaining part.
@ -1016,51 +1059,34 @@ _PyTime_GetSystemClockWithInfo(_PyTime_t *t, _Py_clock_info_t *info)
#ifdef __APPLE__ #ifdef __APPLE__
static int static int
py_mach_timebase_info(_PyTime_t *pnumer, _PyTime_t *pdenom, int raise) py_mach_timebase_info(_PyTimeFraction *base, int raise)
{ {
static mach_timebase_info_data_t timebase; mach_timebase_info_data_t timebase;
/* According to the Technical Q&A QA1398, mach_timebase_info() cannot // According to the Technical Q&A QA1398, mach_timebase_info() cannot
fail: https://developer.apple.com/library/mac/#qa/qa1398/ */ // fail: https://developer.apple.com/library/mac/#qa/qa1398/
(void)mach_timebase_info(&timebase); (void)mach_timebase_info(&timebase);
/* Sanity check: should never occur in practice */ // Check that timebase.numer and timebase.denom can be casted to
if (timebase.numer < 1 || timebase.denom < 1) { // _PyTime_t. In practice, timebase uses uint32_t, so casting cannot
// overflow. At the end, only make sure that the type is uint32_t
// (_PyTime_t is 64-bit long).
Py_BUILD_ASSERT(sizeof(timebase.numer) <= sizeof(_PyTime_t));
Py_BUILD_ASSERT(sizeof(timebase.denom) <= sizeof(_PyTime_t));
_PyTime_t numer = (_PyTime_t)timebase.numer;
_PyTime_t denom = (_PyTime_t)timebase.denom;
// Known time bases:
//
// * (1, 1) on Intel: 1 ns
// * (1000000000, 33333335) on PowerPC: ~30 ns
// * (1000000000, 25000000) on PowerPC: 40 ns
if (_PyTimeFraction_Set(base, numer, denom) < 0) {
if (raise) { if (raise) {
PyErr_SetString(PyExc_RuntimeError, PyErr_SetString(PyExc_RuntimeError,
"invalid mach_timebase_info"); "invalid mach_timebase_info");
} }
return -1; return -1;
} }
/* Check that timebase.numer and timebase.denom can be casted to
_PyTime_t. In practice, timebase uses uint32_t, so casting cannot
overflow. At the end, only make sure that the type is uint32_t
(_PyTime_t is 64-bit long). */
static_assert(sizeof(timebase.numer) <= sizeof(_PyTime_t),
"timebase.numer is larger than _PyTime_t");
static_assert(sizeof(timebase.denom) <= sizeof(_PyTime_t),
"timebase.denom is larger than _PyTime_t");
/* Make sure that _PyTime_MulDiv(ticks, timebase_numer, timebase_denom)
cannot overflow.
Known time bases:
* (1, 1) on Intel
* (1000000000, 33333335) or (1000000000, 25000000) on PowerPC
None of these time bases can overflow with 64-bit _PyTime_t, but
check for overflow, just in case. */
if ((_PyTime_t)timebase.numer > _PyTime_MAX / (_PyTime_t)timebase.denom) {
if (raise) {
PyErr_SetString(PyExc_OverflowError,
"mach_timebase_info is too large");
}
return -1;
}
*pnumer = (_PyTime_t)timebase.numer;
*pdenom = (_PyTime_t)timebase.denom;
return 0; return 0;
} }
#endif #endif
@ -1109,17 +1135,16 @@ py_get_monotonic_clock(_PyTime_t *tp, _Py_clock_info_t *info, int raise_exc)
} }
#elif defined(__APPLE__) #elif defined(__APPLE__)
static _PyTime_t timebase_numer = 0; static _PyTimeFraction base = {0, 0};
static _PyTime_t timebase_denom = 0; if (base.denom == 0) {
if (timebase_denom == 0) { if (py_mach_timebase_info(&base, raise_exc) < 0) {
if (py_mach_timebase_info(&timebase_numer, &timebase_denom, raise_exc) < 0) {
return -1; return -1;
} }
} }
if (info) { if (info) {
info->implementation = "mach_absolute_time()"; info->implementation = "mach_absolute_time()";
info->resolution = (double)timebase_numer / (double)timebase_denom * 1e-9; info->resolution = _PyTimeFraction_Resolution(&base);
info->monotonic = 1; info->monotonic = 1;
info->adjustable = 0; info->adjustable = 0;
} }
@ -1129,7 +1154,7 @@ py_get_monotonic_clock(_PyTime_t *tp, _Py_clock_info_t *info, int raise_exc)
assert(uticks <= (uint64_t)_PyTime_MAX); assert(uticks <= (uint64_t)_PyTime_MAX);
_PyTime_t ticks = (_PyTime_t)uticks; _PyTime_t ticks = (_PyTime_t)uticks;
_PyTime_t ns = _PyTime_MulDiv(ticks, timebase_numer, timebase_denom); _PyTime_t ns = _PyTimeFraction_Mul(ticks, &base);
*tp = pytime_from_nanoseconds(ns); *tp = pytime_from_nanoseconds(ns);
#elif defined(__hpux) #elif defined(__hpux)
@ -1213,7 +1238,7 @@ _PyTime_GetMonotonicClockWithInfo(_PyTime_t *tp, _Py_clock_info_t *info)
#ifdef MS_WINDOWS #ifdef MS_WINDOWS
static int static int
py_win_perf_counter_frequency(LONGLONG *pfrequency, int raise) py_win_perf_counter_frequency(_PyTimeFraction *base, int raise)
{ {
LONGLONG frequency; LONGLONG frequency;
@ -1225,25 +1250,20 @@ py_win_perf_counter_frequency(LONGLONG *pfrequency, int raise)
// Since Windows XP, frequency cannot be zero. // Since Windows XP, frequency cannot be zero.
assert(frequency >= 1); assert(frequency >= 1);
/* Make also sure that (ticks * SEC_TO_NS) cannot overflow in Py_BUILD_ASSERT(sizeof(_PyTime_t) == sizeof(frequency));
_PyTime_MulDiv(), with ticks < frequency. _PyTime_t denom = (_PyTime_t)frequency;
Known QueryPerformanceFrequency() values: // Known QueryPerformanceFrequency() values:
//
* 10,000,000 (10 MHz): 100 ns resolution // * 10,000,000 (10 MHz): 100 ns resolution
* 3,579,545 Hz (3.6 MHz): 279 ns resolution // * 3,579,545 Hz (3.6 MHz): 279 ns resolution
if (_PyTimeFraction_Set(base, SEC_TO_NS, denom) < 0) {
None of these frequencies can overflow with 64-bit _PyTime_t, but
check for integer overflow just in case. */
if (frequency > _PyTime_MAX / SEC_TO_NS) {
if (raise) { if (raise) {
PyErr_SetString(PyExc_OverflowError, PyErr_SetString(PyExc_RuntimeError,
"QueryPerformanceFrequency is too large"); "invalid QueryPerformanceFrequency");
} }
return -1; return -1;
} }
*pfrequency = frequency;
return 0; return 0;
} }
@ -1253,16 +1273,16 @@ py_get_win_perf_counter(_PyTime_t *tp, _Py_clock_info_t *info, int raise_exc)
{ {
assert(info == NULL || raise_exc); assert(info == NULL || raise_exc);
static LONGLONG frequency = 0; static _PyTimeFraction base = {0, 0};
if (frequency == 0) { if (base.denom == 0) {
if (py_win_perf_counter_frequency(&frequency, raise_exc) < 0) { if (py_win_perf_counter_frequency(&base, raise_exc) < 0) {
return -1; return -1;
} }
} }
if (info) { if (info) {
info->implementation = "QueryPerformanceCounter()"; info->implementation = "QueryPerformanceCounter()";
info->resolution = 1.0 / (double)frequency; info->resolution = _PyTimeFraction_Resolution(&base);
info->monotonic = 1; info->monotonic = 1;
info->adjustable = 0; info->adjustable = 0;
} }
@ -1278,7 +1298,7 @@ py_get_win_perf_counter(_PyTime_t *tp, _Py_clock_info_t *info, int raise_exc)
"LONGLONG is larger than _PyTime_t"); "LONGLONG is larger than _PyTime_t");
ticks = (_PyTime_t)ticksll; ticks = (_PyTime_t)ticksll;
_PyTime_t ns = _PyTime_MulDiv(ticks, SEC_TO_NS, (_PyTime_t)frequency); _PyTime_t ns = _PyTimeFraction_Mul(ticks, &base);
*tp = pytime_from_nanoseconds(ns); *tp = pytime_from_nanoseconds(ns);
return 0; return 0;
} }