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 {
  78         tdb_off_t off;
  79         tdb_len_t len;
  80         tdb_len_t unc = 0;
  81
  82         for (off = sizeof(struct tdb_header);
  83              off < tdb->file->map_size;
  84              off += len) {
  85                 const union {
  86                         struct tdb_used_record u;
  87                         struct tdb_free_record f;
  88                         struct tdb_recovery_record r;
  89                 } *p;
  90                 /* We might not be able to get the whole thing. */
  91                 p = tdb_access_read(tdb, off, sizeof(p->f), true);
  92                 if (TDB_PTR_IS_ERR(p)) {
  93                         return TDB_PTR_ERR(p);
  94                 }
  95                 if (frec_magic(&p->f) != TDB_FREE_MAGIC) {
  96                         if (unc > 1) {
  97                                 tally_add(uncoal, unc);
  98                                 unc = 0;
  99                         }
 100                 }
 101
 102                 if (p->r.magic == TDB_RECOVERY_INVALID_MAGIC
 103                     || p->r.magic == TDB_RECOVERY_MAGIC) {
 104                         len = sizeof(p->r) + p->r.max_len;
 105                 } else if (frec_magic(&p->f) == TDB_FREE_MAGIC) {
 106                         len = frec_len(&p->f);
 107                         tally_add(fr, len);
 108                         len += sizeof(p->u);
 109                         unc++;
 110                 } else if (rec_magic(&p->u) == TDB_USED_MAGIC) {
 111                         len = sizeof(p->u)
 112                                 + rec_key_length(&p->u)
 113                                 + rec_data_length(&p->u)
 114                                 + rec_extra_padding(&p->u);
 115
 116                         tally_add(keys, rec_key_length(&p->u));
 117                         tally_add(data, rec_data_length(&p->u));
 118                         tally_add(extra, rec_extra_padding(&p->u));
 119                 } else if (rec_magic(&p->u) == TDB_HTABLE_MAGIC) {
 120                         tdb_off_t count = count_hash(tdb,
 121                                                      off + sizeof(p->u),
 122                                                      TDB_SUBLEVEL_HASH_BITS);
 123                         if (TDB_OFF_IS_ERR(count)) {
 124                                 return TDB_OFF_TO_ERR(count);
 125                         }
 126                         tally_add(hashes, count);
 127                         tally_add(extra, rec_extra_padding(&p->u));
 128                         len = sizeof(p->u)
 129                                 + rec_data_length(&p->u)
 130                                 + rec_extra_padding(&p->u);
 131                 } else if (rec_magic(&p->u) == TDB_FTABLE_MAGIC) {
 132                         len = sizeof(p->u)
 133                                 + rec_data_length(&p->u)
 134                                 + rec_extra_padding(&p->u);
 135                         tally_add(ftables, rec_data_length(&p->u));
 136                         tally_add(extra, rec_extra_padding(&p->u));
 137                 } else if (rec_magic(&p->u) == TDB_CHAIN_MAGIC) {
 138                         len = sizeof(p->u)
 139                                 + rec_data_length(&p->u)
 140                                 + rec_extra_padding(&p->u);
 141                         tally_add(chains, 1);
 142                         tally_add(extra, rec_extra_padding(&p->u));
 143                 } else {
 144                         len = dead_space(tdb, off);
 145                         if (TDB_OFF_IS_ERR(len)) {
 146                                 return TDB_OFF_TO_ERR(len);
 147                         }
 148                 }
 149                 tdb_access_release(tdb, p);
 150         }
 151         if (unc)
 152                 tally_add(uncoal, unc);
 153         return TDB_SUCCESS;
 154 }
 155
 156 static size_t num_capabilities(struct tdb_context *tdb)
 157 {
 158         tdb_off_t off, next;
 159         const struct tdb_capability *cap;
 160         size_t count = 0;
 161
 162         off = tdb_read_off(tdb, offsetof(struct tdb_header, capabilities));
 163         if (TDB_OFF_IS_ERR(off))
 164                 return count;
 165
 166         /* Count capability list. */
 167         for (; off; off = next) {
 168                 cap = tdb_access_read(tdb, off, sizeof(*cap), true);
 169                 if (TDB_PTR_IS_ERR(cap)) {
 170                         break;
 171                 }
 172                 count++;
 173                 next = cap->next;
 174                 tdb_access_release(tdb, cap);
 175         }
 176         return count;
 177 }
 178
 179 static void add_capabilities(struct tdb_context *tdb, size_t num, char *summary)
 180 {
 181         tdb_off_t off, next;
 182         const struct tdb_capability *cap;
 183         size_t count = 0;
 184
 185         /* Append to summary. */
 186         summary += strlen(summary);
 187
 188         off = tdb_read_off(tdb, offsetof(struct tdb_header, capabilities));
 189         if (TDB_OFF_IS_ERR(off))
 190                 return;
 191
 192         /* Walk capability list. */
 193         for (; off; off = next) {
 194                 cap = tdb_access_read(tdb, off, sizeof(*cap), true);
 195                 if (TDB_PTR_IS_ERR(cap)) {
 196                         break;
 197                 }
 198                 count++;
 199                 sprintf(summary, CAPABILITY_FORMAT,
 200                         cap->type & TDB_CAP_TYPE_MASK,
 201                         /* Noopen?  How did we get here? */
 202                         (cap->type & TDB_CAP_NOOPEN) ? " (unopenable)"
 203                         : ((cap->type & TDB_CAP_NOWRITE)
 204                            && (cap->type & TDB_CAP_NOCHECK)) ? " (uncheckable,read-only)"
 205                         : (cap->type & TDB_CAP_NOWRITE) ? " (read-only)"
 206                         : (cap->type & TDB_CAP_NOCHECK) ? " (uncheckable)"
 207                         : "");
 208                 summary += strlen(summary);
 209                 next = cap->next;
 210                 tdb_access_release(tdb, cap);
 211         }
 212 }
 213
 214 enum TDB_ERROR tdb_summary(struct tdb_context *tdb,
 215                            enum tdb_summary_flags flags,
 216                            char **summary)
 217 {
 218         tdb_len_t len;
 219         size_t num_caps;
 220         struct tally *ftables, *hashes, *freet, *keys, *data, *extra, *uncoal,
 221                 *chains;
 222         char *hashesg, *freeg, *keysg, *datag, *extrag, *uncoalg;
 223         enum TDB_ERROR ecode;
 224
 225         if (tdb->flags & TDB_VERSION1) {
 226                 /* tdb1 doesn't do graphs. */
 227                 *summary = tdb1_summary(tdb);
 228                 if (!*summary)
 229                         return tdb->last_error;
 230                 return TDB_SUCCESS;
 231         }
 232
 233         hashesg = freeg = keysg = datag = extrag = uncoalg = NULL;
 234
 235         ecode = tdb_allrecord_lock(tdb, F_RDLCK, TDB_LOCK_WAIT, false);
 236         if (ecode != TDB_SUCCESS) {
 237                 return tdb->last_error = ecode;
 238         }
 239
 240         ecode = tdb_lock_expand(tdb, F_RDLCK);
 241         if (ecode != TDB_SUCCESS) {
 242                 tdb_allrecord_unlock(tdb, F_RDLCK);
 243                 return tdb->last_error = ecode;
 244         }
 245
 246         /* Start stats off empty. */
 247         ftables = tally_new(HISTO_HEIGHT);
 248         hashes = tally_new(HISTO_HEIGHT);
 249         freet = tally_new(HISTO_HEIGHT);
 250         keys = tally_new(HISTO_HEIGHT);
 251         data = tally_new(HISTO_HEIGHT);
 252         extra = tally_new(HISTO_HEIGHT);
 253         uncoal = tally_new(HISTO_HEIGHT);
 254         chains = tally_new(HISTO_HEIGHT);
 255         if (!ftables || !hashes || !freet || !keys || !data || !extra
 256             || !uncoal || !chains) {
 257                 ecode = tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_ERROR,
 258                                    "tdb_summary: failed to allocate"
 259                                    " tally structures");
 260                 goto unlock;
 261         }
 262
 263         ecode = summarize(tdb, hashes, ftables, freet, keys, data, extra,
 264                           uncoal, chains);
 265         if (ecode != TDB_SUCCESS) {
 266                 goto unlock;
 267         }
 268
 269         if (flags & TDB_SUMMARY_HISTOGRAMS) {
 270                 hashesg = tally_histogram(hashes, HISTO_WIDTH, HISTO_HEIGHT);
 271                 freeg = tally_histogram(freet, HISTO_WIDTH, HISTO_HEIGHT);
 272                 keysg = tally_histogram(keys, HISTO_WIDTH, HISTO_HEIGHT);
 273                 datag = tally_histogram(data, HISTO_WIDTH, HISTO_HEIGHT);
 274                 extrag = tally_histogram(extra, HISTO_WIDTH, HISTO_HEIGHT);
 275                 uncoalg = tally_histogram(uncoal, HISTO_WIDTH, HISTO_HEIGHT);
 276         }
 277
 278         num_caps = num_capabilities(tdb);
 279
 280         /* 20 is max length of a %llu. */
 281         len = strlen(SUMMARY_FORMAT) + 33*20 + 1
 282                 + (hashesg ? strlen(hashesg) : 0)
 283                 + (freeg ? strlen(freeg) : 0)
 284                 + (keysg ? strlen(keysg) : 0)
 285                 + (datag ? strlen(datag) : 0)
 286                 + (extrag ? strlen(extrag) : 0)
 287                 + (uncoalg ? strlen(uncoalg) : 0)
 288                 + num_caps * (strlen(CAPABILITY_FORMAT) + 20*4);
 289
 290         *summary = malloc(len);
 291         if (!*summary) {
 292                 ecode = tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_ERROR,
 293                                    "tdb_summary: failed to allocate string");
 294                 goto unlock;
 295         }
 296
 297         sprintf(*summary, SUMMARY_FORMAT,
 298                 (size_t)tdb->file->map_size,
 299                 tally_total(keys, NULL) + tally_total(data, NULL),
 300                 tally_num(keys),
 301                 tally_min(keys), tally_mean(keys), tally_max(keys),
 302                 keysg ? keysg : "",
 303                 tally_min(data), tally_mean(data), tally_max(data),
 304                 datag ? datag : "",
 305                 tally_min(extra), tally_mean(extra), tally_max(extra),
 306                 extrag ? extrag : "",
 307                 tally_num(freet),
 308                 tally_min(freet), tally_mean(freet), tally_max(freet),
 309                 freeg ? freeg : "",
 310                 tally_total(uncoal, NULL),
 311                 tally_min(uncoal), tally_mean(uncoal), tally_max(uncoal),
 312                 uncoalg ? uncoalg : "",
 313                 (unsigned)count_hash(tdb, offsetof(struct tdb_header,
 314                                                    hashtable),
 315                                      TDB_TOPLEVEL_HASH_BITS),
 316                 1 << TDB_TOPLEVEL_HASH_BITS,
 317                 tally_num(chains),
 318                 tally_num(hashes),
 319                 tally_min(hashes), tally_mean(hashes), tally_max(hashes),
 320                 hashesg ? hashesg : "",
 321                 tally_total(keys, NULL) * 100.0 / tdb->file->map_size,
 322                 tally_total(data, NULL) * 100.0 / tdb->file->map_size,
 323                 tally_total(extra, NULL) * 100.0 / tdb->file->map_size,
 324                 tally_total(freet, NULL) * 100.0 / tdb->file->map_size,
 325                 (tally_num(keys) + tally_num(freet) + tally_num(hashes))
 326                 * sizeof(struct tdb_used_record) * 100.0 / tdb->file->map_size,
 327                 tally_num(ftables) * sizeof(struct tdb_freetable)
 328                 * 100.0 / tdb->file->map_size,
 329                 (tally_num(hashes)
 330                  * (sizeof(tdb_off_t) << TDB_SUBLEVEL_HASH_BITS)
 331                  + (sizeof(tdb_off_t) << TDB_TOPLEVEL_HASH_BITS)
 332                  + sizeof(struct tdb_chain) * tally_num(chains))
 333                 * 100.0 / tdb->file->map_size);
 334
 335         add_capabilities(tdb, num_caps, *summary);
 336
 337 unlock:
 338         free(hashesg);
 339         free(freeg);
 340         free(keysg);
 341         free(datag);
 342         free(extrag);
 343         free(uncoalg);
 344         free(hashes);
 345         free(freet);
 346         free(keys);
 347         free(data);
 348         free(extra);
 349         free(uncoal);
 350         free(ftables);
 351         free(chains);
 352
 353         tdb_allrecord_unlock(tdb, F_RDLCK);
 354         tdb_unlock_expand(tdb, F_RDLCK);
 355         return tdb->last_error = ecode;
 356 }