/usr/share/systemtap/runtime/stat.c is in systemtap-common 3.1-3ubuntu0.1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 | /* -*- linux-c -*-
* Statistics Aggregation
* Copyright (C) 2005-2016 Red Hat Inc.
* Copyright (C) 2006 Intel Corporation
*
* This file is part of systemtap, and is free software. You can
* redistribute it and/or modify it under the terms of the GNU General
* Public License (GPL); either version 2, or (at your option) any
* later version.
*/
#ifndef _STAT_C_
#define _STAT_C_
/** @file stat.c
* @brief Statistics Aggregation
*/
/** @addtogroup stat Statistics Aggregation
* The Statistics aggregations keep per-cpu statistics. You
* must create all aggregations at probe initialization and it is
* best to not read them until probe exit. If you must read them
* while probes are running, the values may be slightly off due
* to a probe updating the statistics of one cpu while another cpu attempts
* to read the same data. This will also negatively impact performance.
*
* Stats keep track of count, sum, min, max, avg, and variance. Bit-shift
* can be optionally specified, scaling the numbers, in order to improve the
* accuracy of the integer arithmetics.
*
* Histograms are optional. If you want a histogram, you must set "type"
* to HIST_LOG or HIST_LINEAR when you call _stp_stat_init().
*
* @{
*/
#include "stat-common.c"
/** Initialize a Stat.
* Call this during probe initialization to create a Stat.
*
* @param type (KEY_HIST_TYPE and associated parameters)
*
* For HIST_LOG, the following additional parametrs are required:
* @param buckets - An integer specifying the number of buckets.
*
* For HIST_LINEAR, the following additional parametrs are required:
* @param start - An integer. The start of the histogram.
* @param stop - An integer. The stopping value. Should be > start.
* @param interval - An integer. The interval.
*
* @param stat_ops (STAT_OP_* and associated parameter bit_shift for STAT_OP_VARIANCE)
*/
static Stat _stp_stat_init (int first_arg, ...)
{
int size, buckets=0, start=0, stop=0, interval=0, bit_shift=0;
int stat_ops=0, htype=0;
int arg = first_arg;
Stat st;
va_list ap;
va_start (ap, first_arg);
do {
switch (arg) {
case KEY_HIST_TYPE:
htype = va_arg(ap, int);
if (htype == HIST_LINEAR) {
start = va_arg(ap, int);
stop = va_arg(ap, int);
interval = va_arg(ap, int);
buckets = _stp_stat_calc_buckets(stop, start, interval);
if (!buckets) {
va_end (ap);
return NULL;
}
}
if (htype == HIST_LOG)
buckets = HIST_LOG_BUCKETS;
break;
case STAT_OP_COUNT:
stat_ops |= STAT_OP_COUNT;
break;
case STAT_OP_SUM:
stat_ops |= STAT_OP_SUM;
break;
case STAT_OP_MIN:
stat_ops |= STAT_OP_MIN;
break;
case STAT_OP_MAX:
stat_ops |= STAT_OP_MAX;
break;
case STAT_OP_AVG:
stat_ops |= STAT_OP_AVG;
break;
case STAT_OP_VARIANCE:
stat_ops |= STAT_OP_VARIANCE;
bit_shift = va_arg(ap, int);
break;
default:
_stp_warn ("Unknown argument %d\n", arg);
}
arg = va_arg(ap, int);
} while (arg);
va_end (ap);
size = buckets * sizeof(int64_t) + sizeof(stat_data);
st = _stp_stat_alloc (size);
if (st == NULL)
return NULL;
st->hist.type = htype;
st->hist.start = start;
st->hist.stop = stop;
st->hist.interval = interval;
st->hist.buckets = buckets;
st->hist.bit_shift = bit_shift;
st->hist.stat_ops = stat_ops;
return st;
}
/** Delete Stat.
* Call this to free all memory allocated during initialization.
*
* @param st Stat
*/
static void _stp_stat_del (Stat st)
{
if (st)
_stp_stat_free(st);
}
/** Add to a Stat.
* Add an int64 to a Stat, and for optimization purposes specify which
* statistical operators are bound to given Stat. Set all of stat_op*
* to 1 if unsure. Note that @avg() is being evaluated separately based
* on @sum and @count within the code directly generated by the translator.
*
* @param st Stat
* @param val Value to add
* @param stat_op_count int
* @param stat_op_sum int
* @param stat_op_min int
* @param stat_op_max int
* @param stat_op_variance int
*
*/
static inline void _stp_stat_add (Stat st, int64_t val, int stat_op_count,
int stat_op_sum, int stat_op_min,
int stat_op_max, int stat_op_variance)
{
stat_data *sd = _stp_stat_per_cpu_ptr (st, STAT_GET_CPU());
STAT_LOCK(sd);
__stp_stat_add (&st->hist, sd, val, stat_op_count, stat_op_sum,
stat_op_min, stat_op_max, stat_op_variance);
STAT_UNLOCK(sd);
STAT_PUT_CPU();
}
static void _stp_stat_clear_data (Stat st, stat_data *sd)
{
int j;
sd->count = sd->sum = sd->min = sd->max = 0;
sd->avg_s = sd->variance = sd->variance_s = 0;
if (st->hist.type != HIST_NONE) {
for (j = 0; j < st->hist.buckets; j++)
sd->histogram[j] = 0;
}
}
/** Get Stats.
* Gets the aggregated Stats for all CPUs.
*
* @param st Stat
* @param clear Set if you want the data cleared after the read. Useful
* for polling.
* @returns A pointer to a stat.
*/
static stat_data *_stp_stat_get (Stat st, int clear)
{
int i, j;
int64_t S1, S2;
stat_data *agg = _stp_stat_get_agg(st);
stat_data *sd;
STAT_LOCK(agg);
_stp_stat_clear_data (st, agg);
S1 = S2 = 0;
for_each_possible_cpu(i) {
stat_data *sd = _stp_stat_per_cpu_ptr (st, i);
STAT_LOCK(sd);
if (sd->count) {
agg->shift = sd->shift;
if (agg->count == 0) {
agg->min = sd->min;
agg->max = sd->max;
}
agg->count += sd->count;
agg->sum += sd->sum;
if (sd->max > agg->max)
agg->max = sd->max;
if (sd->min < agg->min)
agg->min = sd->min;
if (st->hist.type != HIST_NONE) {
for (j = 0; j < st->hist.buckets; j++)
agg->histogram[j] += sd->histogram[j];
}
}
STAT_UNLOCK(sd);
}
agg->avg_s = _stp_div64(NULL, agg->sum << agg->shift, agg->count);
/*
* For aggregating variance over available CPUs, the Total Variance
* formula is being used. This formula is mentioned in following
* paper: Niranjan Kamat, Arnab Nandi: A Closer Look at Variance
* Implementations In Modern Database Systems: SIGMOD Record 2015.
* Available at: http://web.cse.ohio-state.edu/~kamatn/variance.pdf
*/
for_each_possible_cpu(i) {
sd = _stp_stat_per_cpu_ptr (st, i);
STAT_LOCK(sd);
if (sd->count) {
S1 += sd->count * (sd->avg_s - agg->avg_s) * (sd->avg_s - agg->avg_s);
S2 += (sd->count - 1) * sd->variance_s;
}
if (clear)
_stp_stat_clear_data (st, sd);
STAT_UNLOCK(sd);
}
agg->variance_s = _stp_div64(NULL, (S1 + S2), (agg->count - 1));
agg->variance = agg->variance_s >> (2 * agg->shift);
/*
* Originally this function returned the aggregate still
* locked and it was the caller's responsibility to unlock the
* aggregate. However the translator generated code that called
* this function wasn't unlocking it...
*
* But, the translator generates its own locks for global
* variables (like stats), so we don't need to return the
* aggregate still locked.
*
* It is possible we could even skip locking the aggregate in
* this function, but to be a bit paranoid lets keep the
* locking.
*/
STAT_UNLOCK(agg);
return agg;
}
/** Clear Stats.
* Clears the Stats.
*
* @param st Stat
*/
static void _stp_stat_clear (Stat st)
{
int i;
for_each_possible_cpu(i) {
stat_data *sd = _stp_stat_per_cpu_ptr (st, i);
STAT_LOCK(sd);
_stp_stat_clear_data (st, sd);
STAT_UNLOCK(sd);
}
}
/** @} */
#endif /* _STAT_C_ */
|