summaryrefslogtreecommitdiffstats
path: root/block/accounting.c
blob: 2030851d79d65de0bfd4cecebe20c1b2e8a1815c (plain) (blame)
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
273
274
275
276
277
278
279
280
/*
 * QEMU System Emulator block accounting
 *
 * Copyright (c) 2011 Christoph Hellwig
 * Copyright (c) 2015 Igalia, S.L.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 * THE SOFTWARE.
 */

#include "qemu/osdep.h"
#include "block/accounting.h"
#include "block/block_int.h"
#include "qemu/timer.h"
#include "sysemu/qtest.h"

static QEMUClockType clock_type = QEMU_CLOCK_REALTIME;
static const int qtest_latency_ns = NANOSECONDS_PER_SECOND / 1000;

void block_acct_init(BlockAcctStats *stats)
{
    qemu_mutex_init(&stats->lock);
    if (qtest_enabled()) {
        clock_type = QEMU_CLOCK_VIRTUAL;
    }
}

void block_acct_setup(BlockAcctStats *stats, bool account_invalid,
                      bool account_failed)
{
    stats->account_invalid = account_invalid;
    stats->account_failed = account_failed;
}

void block_acct_cleanup(BlockAcctStats *stats)
{
    BlockAcctTimedStats *s, *next;
    QSLIST_FOREACH_SAFE(s, &stats->intervals, entries, next) {
        g_free(s);
    }
    qemu_mutex_destroy(&stats->lock);
}

void block_acct_add_interval(BlockAcctStats *stats, unsigned interval_length)
{
    BlockAcctTimedStats *s;
    unsigned i;

    s = g_new0(BlockAcctTimedStats, 1);
    s->interval_length = interval_length;
    s->stats = stats;
    qemu_mutex_lock(&stats->lock);
    QSLIST_INSERT_HEAD(&stats->intervals, s, entries);

    for (i = 0; i < BLOCK_MAX_IOTYPE; i++) {
        timed_average_init(&s->latency[i], clock_type,
                           (uint64_t) interval_length * NANOSECONDS_PER_SECOND);
    }
    qemu_mutex_unlock(&stats->lock);
}

BlockAcctTimedStats *block_acct_interval_next(BlockAcctStats *stats,
                                              BlockAcctTimedStats *s)
{
    if (s == NULL) {
        return QSLIST_FIRST(&stats->intervals);
    } else {
        return QSLIST_NEXT(s, entries);
    }
}

void block_acct_start(BlockAcctStats *stats, BlockAcctCookie *cookie,
                      int64_t bytes, enum BlockAcctType type)
{
    assert(type < BLOCK_MAX_IOTYPE);

    cookie->bytes = bytes;
    cookie->start_time_ns = qemu_clock_get_ns(clock_type);
    cookie->type = type;
}

/* block_latency_histogram_compare_func:
 * Compare @key with interval [@it[0], @it[1]).
 * Return: -1 if @key < @it[0]
 *          0 if @key in [@it[0], @it[1])
 *         +1 if @key >= @it[1]
 */
static int block_latency_histogram_compare_func(const void *key, const void *it)
{
    uint64_t k = *(uint64_t *)key;
    uint64_t a = ((uint64_t *)it)[0];
    uint64_t b = ((uint64_t *)it)[1];

    return k < a ? -1 : (k < b ? 0 : 1);
}

static void block_latency_histogram_account(BlockLatencyHistogram *hist,
                                            int64_t latency_ns)
{
    uint64_t *pos;

    if (hist->bins == NULL) {
        /* histogram disabled */
        return;
    }


    if (latency_ns < hist->boundaries[0]) {
        hist->bins[0]++;
        return;
    }

    if (latency_ns >= hist->boundaries[hist->nbins - 2]) {
        hist->bins[hist->nbins - 1]++;
        return;
    }

    pos = bsearch(&latency_ns, hist->boundaries, hist->nbins - 2,
                  sizeof(hist->boundaries[0]),
                  block_latency_histogram_compare_func);
    assert(pos != NULL);

    hist->bins[pos - hist->boundaries + 1]++;
}

int block_latency_histogram_set(BlockAcctStats *stats, enum BlockAcctType type,
                                uint64List *boundaries)
{
    BlockLatencyHistogram *hist = &stats->latency_histogram[type];
    uint64List *entry;
    uint64_t *ptr;
    uint64_t prev = 0;
    int new_nbins = 1;

    for (entry = boundaries; entry; entry = entry->next) {
        if (entry->value <= prev) {
            return -EINVAL;
        }
        new_nbins++;
        prev = entry->value;
    }

    hist->nbins = new_nbins;
    g_free(hist->boundaries);
    hist->boundaries = g_new(uint64_t, hist->nbins - 1);
    for (entry = boundaries, ptr = hist->boundaries; entry;
         entry = entry->next, ptr++)
    {
        *ptr = entry->value;
    }

    g_free(hist->bins);
    hist->bins = g_new0(uint64_t, hist->nbins);

    return 0;
}

void block_latency_histograms_clear(BlockAcctStats *stats)
{
    int i;

    for (i = 0; i < BLOCK_MAX_IOTYPE; i++) {
        BlockLatencyHistogram *hist = &stats->latency_histogram[i];
        g_free(hist->bins);
        g_free(hist->boundaries);
        memset(hist, 0, sizeof(*hist));
    }
}

static void block_account_one_io(BlockAcctStats *stats, BlockAcctCookie *cookie,
                                 bool failed)
{
    BlockAcctTimedStats *s;
    int64_t time_ns = qemu_clock_get_ns(clock_type);
    int64_t latency_ns = time_ns - cookie->start_time_ns;

    if (qtest_enabled()) {
        latency_ns = qtest_latency_ns;
    }

    assert(cookie->type < BLOCK_MAX_IOTYPE);

    if (cookie->type == BLOCK_ACCT_NONE) {
        return;
    }

    WITH_QEMU_LOCK_GUARD(&stats->lock) {
        if (failed) {
            stats->failed_ops[cookie->type]++;
        } else {
            stats->nr_bytes[cookie->type] += cookie->bytes;
            stats->nr_ops[cookie->type]++;
        }

        block_latency_histogram_account(&stats->latency_histogram[cookie->type],
                                        latency_ns);

        if (!failed || stats->account_failed) {
            stats->total_time_ns[cookie->type] += latency_ns;
            stats->last_access_time_ns = time_ns;

            QSLIST_FOREACH(s, &stats->intervals, entries) {
                timed_average_account(&s->latency[cookie->type], latency_ns);
            }
        }
    }

    cookie->type = BLOCK_ACCT_NONE;
}

void block_acct_done(BlockAcctStats *stats, BlockAcctCookie *cookie)
{
    block_account_one_io(stats, cookie, false);
}

void block_acct_failed(BlockAcctStats *stats, BlockAcctCookie *cookie)
{
    block_account_one_io(stats, cookie, true);
}

void block_acct_invalid(BlockAcctStats *stats, enum BlockAcctType type)
{
    assert(type < BLOCK_MAX_IOTYPE);

    /* block_account_one_io() updates total_time_ns[], but this one does
     * not.  The reason is that invalid requests are accounted during their
     * submission, therefore there's no actual I/O involved.
     */
    qemu_mutex_lock(&stats->lock);
    stats->invalid_ops[type]++;

    if (stats->account_invalid) {
        stats->last_access_time_ns = qemu_clock_get_ns(clock_type);
    }
    qemu_mutex_unlock(&stats->lock);
}

void block_acct_merge_done(BlockAcctStats *stats, enum BlockAcctType type,
                      int num_requests)
{
    assert(type < BLOCK_MAX_IOTYPE);

    qemu_mutex_lock(&stats->lock);
    stats->merged[type] += num_requests;
    qemu_mutex_unlock(&stats->lock);
}

int64_t block_acct_idle_time_ns(BlockAcctStats *stats)
{
    return qemu_clock_get_ns(clock_type) - stats->last_access_time_ns;
}

double block_acct_queue_depth(BlockAcctTimedStats *stats,
                              enum BlockAcctType type)
{
    uint64_t sum, elapsed;

    assert(type < BLOCK_MAX_IOTYPE);

    qemu_mutex_lock(&stats->stats->lock);
    sum = timed_average_sum(&stats->latency[type], &elapsed);
    qemu_mutex_unlock(&stats->stats->lock);

    return (double) sum / elapsed;
}