lib/tdb2/summary.c

   1  /*
   2    Trivial Database 2: human-readable summary code
   3    Copyright (C) Rusty Russell 2010
   4
   5    This library is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Lesser General Public
   7    License as published by the Free Software Foundation; either
   8    version 3 of the License, or (at your option) any later version.
   9
  10    This library is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public
  16    License along with this library; if not, see <http://www.gnu.org/licenses/>.
  17 */
  18 #include "private.h"
  19 #include <assert.h>
  20 #include <ccan/tally/tally.h>
  21
  22 #define SUMMARY_FORMAT \
  23         "Size of file/data: %zu/%zu\n" \
  24         "Number of records: %zu\n" \
  25         "Smallest/average/largest keys: %zu/%zu/%zu\n%s" \
  26         "Smallest/average/largest data: %zu/%zu/%zu\n%s" \
  27         "Smallest/average/largest padding: %zu/%zu/%zu\n%s" \
  28         "Number of free records: %zu\n" \
  29         "Smallest/average/largest free records: %zu/%zu/%zu\n%s" \
  30         "Number of uncoalesced records: %zu\n" \
  31         "Smallest/average/largest uncoalesced runs: %zu/%zu/%zu\n%s" \
  32         "Toplevel hash used: %u of %u\n" \
  33         "Number of chains: %zu\n" \
  34         "Number of subhashes: %zu\n" \
  35         "Smallest/average/largest subhash entries: %zu/%zu/%zu\n%s" \
  36         "Percentage keys/data/padding/free/rechdrs/freehdrs/hashes: %.0f/%.0f/%.0f/%.0f/%.0f/%.0f/%.0f\n"
  37
  38 #define BUCKET_SUMMARY_FORMAT_A                                 \
  39         "Free bucket %zu: total entries %zu.\n"                 \
  40         "Smallest/average/largest length: %zu/%zu/%zu\n%s"
  41 #define BUCKET_SUMMARY_FORMAT_B                                 \
  42         "Free bucket %zu-%zu: total entries %zu.\n"             \
  43         "Smallest/average/largest length: %zu/%zu/%zu\n%s"
  44 #define CAPABILITY_FORMAT                                       \
  45         "Capability %llu%s\n"
  46
  47 #define HISTO_WIDTH 70
  48 #define HISTO_HEIGHT 20
  49
  50 static tdb_off_t count_hash(struct tdb_context *tdb,
  51                             tdb_off_t hash_off, unsigned bits)
  52 {
  53         const tdb_off_t *h;
  54         tdb_off_t count = 0;
  55         unsigned int i;
  56
  57         h = tdb_access_read(tdb, hash_off, sizeof(*h) << bits, true);
  58         if (TDB_PTR_IS_ERR(h)) {
  59                 return TDB_ERR_TO_OFF(TDB_PTR_ERR(h));
  60         }
  61         for (i = 0; i < (1 << bits); i++)
  62                 count += (h[i] != 0);
  63
  64         tdb_access_release(tdb, h);
  65         return count;
  66 }
  67
  68 static enum TDB_ERROR summarize(struct tdb_context *tdb,
  69                                 struct tally *hashes,
  70                                 struct tally *ftables,
  71                                 struct tally *fr,
  72                                 struct tally *keys,
  73                                 struct tally *data,
  74                                 struct tally *extra,
  75                                 struct tally *uncoal,
  76                                 struct tally *chains,
  77                                 size_t *num_caps)
  78 {
  79         tdb_off_t off;
  80         tdb_len_t len;
  81         tdb_len_t unc = 0;
  82
  83         for (off = sizeof(struct tdb_header);
  84              off < tdb->file->map_size;
  85              off += len) {
  86                 const union {
  87                         struct tdb_used_record u;
  88                         struct tdb_free_record f;
  89                         struct tdb_recovery_record r;
  90                 } *p;
  91                 /* We might not be able to get the whole thing. */
  92                 p = tdb_access_read(tdb, off, sizeof(p->f), true);
  93                 if (TDB_PTR_IS_ERR(p)) {
  94                         return TDB_PTR_ERR(p);
  95                 }
  96                 if (frec_magic(&p->f) != TDB_FREE_MAGIC) {
  97                         if (unc > 1) {
  98                                 tally_add(uncoal, unc);
  99                                 unc = 0;
 100                         }
 101                 }
 102
 103                 if (p->r.magic == TDB_RECOVERY_INVALID_MAGIC
 104                     || p->r.magic == TDB_RECOVERY_MAGIC) {
 105                         len = sizeof(p->r) + p->r.max_len;
 106                 } else if (frec_magic(&p->f) == TDB_FREE_MAGIC) {
 107                         len = frec_len(&p->f);
 108                         tally_add(fr, len);
 109                         len += sizeof(p->u);
 110                         unc++;
 111                 } else if (rec_magic(&p->u) == TDB_USED_MAGIC) {
 112                         len = sizeof(p->u)
 113                                 + rec_key_length(&p->u)
 114                                 + rec_data_length(&p->u)
 115                                 + rec_extra_padding(&p->u);
 116
 117                         tally_add(keys, rec_key_length(&p->u));
 118                         tally_add(data, rec_data_length(&p->u));
 119                         tally_add(extra, rec_extra_padding(&p->u));
 120                 } else if (rec_magic(&p->u) == TDB_HTABLE_MAGIC) {
 121                         tdb_off_t count = count_hash(tdb,
 122                                                      off + sizeof(p->u),
 123                                                      TDB_SUBLEVEL_HASH_BITS);
 124                         if (TDB_OFF_IS_ERR(count)) {
 125                                 return TDB_OFF_TO_ERR(count);
 126                         }
 127                         tally_add(hashes, count);
 128                         tally_add(extra, rec_extra_padding(&p->u));
 129                         len = sizeof(p->u)
 130                                 + rec_data_length(&p->u)
 131                                 + rec_extra_padding(&p->u);
 132                 } else if (rec_magic(&p->u) == TDB_FTABLE_MAGIC) {
 133                         len = sizeof(p->u)
 134                                 + rec_data_length(&p->u)
 135                                 + rec_extra_padding(&p->u);
 136                         tally_add(ftables, rec_data_length(&p->u));
 137                         tally_add(extra, rec_extra_padding(&p->u));
 138                 } else if (rec_magic(&p->u) == TDB_CHAIN_MAGIC) {
 139                         len = sizeof(p->u)
 140                                 + rec_data_length(&p->u)
 141                                 + rec_extra_padding(&p->u);
 142                         tally_add(chains, 1);
 143                         tally_add(extra, rec_extra_padding(&p->u));
 144                 } else if (rec_magic(&p->u) == TDB_CAP_MAGIC) {
 145                         len = sizeof(p->u)
 146                                 + rec_data_length(&p->u)
 147                                 + rec_extra_padding(&p->u);
 148                         (*num_caps)++;
 149                 } else {
 150                         len = dead_space(tdb, off);
 151                         if (TDB_OFF_IS_ERR(len)) {
 152                                 return TDB_OFF_TO_ERR(len);
 153                         }
 154                 }
 155                 tdb_access_release(tdb, p);
 156         }
 157         if (unc)
 158                 tally_add(uncoal, unc);
 159         return TDB_SUCCESS;
 160 }
 161
 162 static void add_capabilities(struct tdb_context *tdb, char *summary)
 163 {
 164         tdb_off_t off, next;
 165         const struct tdb_capability *cap;
 166         size_t count = 0;
 167
 168         /* Append to summary. */
 169         summary += strlen(summary);
 170
 171         off = tdb_read_off(tdb, offsetof(struct tdb_header, capabilities));
 172         if (TDB_OFF_IS_ERR(off))
 173                 return;
 174
 175         /* Walk capability list. */
 176         for (; off; off = next) {
 177                 cap = tdb_access_read(tdb, off, sizeof(*cap), true);
 178                 if (TDB_PTR_IS_ERR(cap)) {
 179                         break;
 180                 }
 181                 count++;
 182                 sprintf(summary, CAPABILITY_FORMAT,
 183                         cap->type & TDB_CAP_TYPE_MASK,
 184                         /* Noopen?  How did we get here? */
 185                         (cap->type & TDB_CAP_NOOPEN) ? " (unopenable)"
 186                         : ((cap->type & TDB_CAP_NOWRITE)
 187                            && (cap->type & TDB_CAP_NOCHECK)) ? " (uncheckable,read-only)"
 188                         : (cap->type & TDB_CAP_NOWRITE) ? " (read-only)"
 189                         : (cap->type & TDB_CAP_NOCHECK) ? " (uncheckable)"
 190                         : "");
 191                 summary += strlen(summary);
 192                 next = cap->next;
 193                 tdb_access_release(tdb, cap);
 194         }
 195 }
 196
 197 _PUBLIC_ enum TDB_ERROR tdb_summary(struct tdb_context *tdb,
 198                            enum tdb_summary_flags flags,
 199                            char **summary)
 200 {
 201         tdb_len_t len;
 202         size_t num_caps = 0;
 203         struct tally *ftables, *hashes, *freet, *keys, *data, *extra, *uncoal,
 204                 *chains;
 205         char *hashesg, *freeg, *keysg, *datag, *extrag, *uncoalg;
 206         enum TDB_ERROR ecode;
 207
 208         if (tdb->flags & TDB_VERSION1) {
 209                 /* tdb1 doesn't do graphs. */
 210                 *summary = tdb1_summary(tdb);
 211                 if (!*summary)
 212                         return tdb->last_error;
 213                 return TDB_SUCCESS;
 214         }
 215
 216         hashesg = freeg = keysg = datag = extrag = uncoalg = NULL;
 217
 218         ecode = tdb_allrecord_lock(tdb, F_RDLCK, TDB_LOCK_WAIT, false);
 219         if (ecode != TDB_SUCCESS) {
 220                 return tdb->last_error = ecode;
 221         }
 222
 223         ecode = tdb_lock_expand(tdb, F_RDLCK);
 224         if (ecode != TDB_SUCCESS) {
 225                 tdb_allrecord_unlock(tdb, F_RDLCK);
 226                 return tdb->last_error = ecode;
 227         }
 228
 229         /* Start stats off empty. */
 230         ftables = tally_new(HISTO_HEIGHT);
 231         hashes = tally_new(HISTO_HEIGHT);
 232         freet = tally_new(HISTO_HEIGHT);
 233         keys = tally_new(HISTO_HEIGHT);
 234         data = tally_new(HISTO_HEIGHT);
 235         extra = tally_new(HISTO_HEIGHT);
 236         uncoal = tally_new(HISTO_HEIGHT);
 237         chains = tally_new(HISTO_HEIGHT);
 238         if (!ftables || !hashes || !freet || !keys || !data || !extra
 239             || !uncoal || !chains) {
 240                 ecode = tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_ERROR,
 241                                    "tdb_summary: failed to allocate"
 242                                    " tally structures");
 243                 goto unlock;
 244         }
 245
 246         ecode = summarize(tdb, hashes, ftables, freet, keys, data, extra,
 247                           uncoal, chains, &num_caps);
 248         if (ecode != TDB_SUCCESS) {
 249                 goto unlock;
 250         }
 251
 252         if (flags & TDB_SUMMARY_HISTOGRAMS) {
 253                 hashesg = tally_histogram(hashes, HISTO_WIDTH, HISTO_HEIGHT);
 254                 freeg = tally_histogram(freet, HISTO_WIDTH, HISTO_HEIGHT);
 255                 keysg = tally_histogram(keys, HISTO_WIDTH, HISTO_HEIGHT);
 256                 datag = tally_histogram(data, HISTO_WIDTH, HISTO_HEIGHT);
 257                 extrag = tally_histogram(extra, HISTO_WIDTH, HISTO_HEIGHT);
 258                 uncoalg = tally_histogram(uncoal, HISTO_WIDTH, HISTO_HEIGHT);
 259         }
 260
 261         /* 20 is max length of a %llu. */
 262         len = strlen(SUMMARY_FORMAT) + 33*20 + 1
 263                 + (hashesg ? strlen(hashesg) : 0)
 264                 + (freeg ? strlen(freeg) : 0)
 265                 + (keysg ? strlen(keysg) : 0)
 266                 + (datag ? strlen(datag) : 0)
 267                 + (extrag ? strlen(extrag) : 0)
 268                 + (uncoalg ? strlen(uncoalg) : 0)
 269                 + num_caps * (strlen(CAPABILITY_FORMAT) + 20
 270                               + strlen(" (uncheckable,read-only)"));
 271
 272         *summary = malloc(len);
 273         if (!*summary) {
 274                 ecode = tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_ERROR,
 275                                    "tdb_summary: failed to allocate string");
 276                 goto unlock;
 277         }
 278
 279         sprintf(*summary, SUMMARY_FORMAT,
 280                 (size_t)tdb->file->map_size,
 281                 tally_total(keys, NULL) + tally_total(data, NULL),
 282                 tally_num(keys),
 283                 tally_min(keys), tally_mean(keys), tally_max(keys),
 284                 keysg ? keysg : "",
 285                 tally_min(data), tally_mean(data), tally_max(data),
 286                 datag ? datag : "",
 287                 tally_min(extra), tally_mean(extra), tally_max(extra),
 288                 extrag ? extrag : "",
 289                 tally_num(freet),
 290                 tally_min(freet), tally_mean(freet), tally_max(freet),
 291                 freeg ? freeg : "",
 292                 tally_total(uncoal, NULL),
 293                 tally_min(uncoal), tally_mean(uncoal), tally_max(uncoal),
 294                 uncoalg ? uncoalg : "",
 295                 (unsigned)count_hash(tdb, offsetof(struct tdb_header,
 296                                                    hashtable),
 297                                      TDB_TOPLEVEL_HASH_BITS),
 298                 1 << TDB_TOPLEVEL_HASH_BITS,
 299                 tally_num(chains),
 300                 tally_num(hashes),
 301                 tally_min(hashes), tally_mean(hashes), tally_max(hashes),
 302                 hashesg ? hashesg : "",
 303                 tally_total(keys, NULL) * 100.0 / tdb->file->map_size,
 304                 tally_total(data, NULL) * 100.0 / tdb->file->map_size,
 305                 tally_total(extra, NULL) * 100.0 / tdb->file->map_size,
 306                 tally_total(freet, NULL) * 100.0 / tdb->file->map_size,
 307                 (tally_num(keys) + tally_num(freet) + tally_num(hashes))
 308                 * sizeof(struct tdb_used_record) * 100.0 / tdb->file->map_size,
 309                 tally_num(ftables) * sizeof(struct tdb_freetable)
 310                 * 100.0 / tdb->file->map_size,
 311                 (tally_num(hashes)
 312                  * (sizeof(tdb_off_t) << TDB_SUBLEVEL_HASH_BITS)
 313                  + (sizeof(tdb_off_t) << TDB_TOPLEVEL_HASH_BITS)
 314                  + sizeof(struct tdb_chain) * tally_num(chains))
 315                 * 100.0 / tdb->file->map_size);
 316
 317         add_capabilities(tdb, *summary);
 318
 319 unlock:
 320         free(hashesg);
 321         free(freeg);
 322         free(keysg);
 323         free(datag);
 324         free(extrag);
 325         free(uncoalg);
 326         free(hashes);
 327         free(freet);
 328         free(keys);
 329         free(data);
 330         free(extra);
 331         free(uncoal);
 332         free(ftables);
 333         free(chains);
 334
 335         tdb_allrecord_unlock(tdb, F_RDLCK);
 336         tdb_unlock_expand(tdb, F_RDLCK);
 337         return tdb->last_error = ecode;
 338 }