]> git.ozlabs.org Git - ccan/commitdiff
tdb2: shrink free header from 32 to 24 bytes.
authorRusty Russell <rusty@rustcorp.com.au>
Wed, 1 Dec 2010 13:22:43 +0000 (23:52 +1030)
committerRusty Russell <rusty@rustcorp.com.au>
Wed, 1 Dec 2010 13:22:43 +0000 (23:52 +1030)
This reduces our minimum key+data length to 8 bytes; we do this by packing
the prev pointer where we used to put the flist pointer, and storing the
flist as an 8 bit index (meaning we can only have 256 free tables).

Note that this has a perverse result on the size of the database, as our
4-byte key and 4-byte data now fit perfectly in a minimal record, so
appeding causes us to allocate new records which are 50% larger,
since we detect growing.

Current results of speed test:
$ ./speed 1000000
Adding 1000000 records:  23210 ns (59193360 bytes)
Finding 1000000 records:  2387 ns (59193360 bytes)
Traversing 1000000 records:  2150 ns (59193360 bytes)
Deleting 1000000 records:  13392 ns (59193360 bytes)
Re-adding 1000000 records:  11546 ns (59193360 bytes)
Appending 1000000 records:  29327 ns (91193360 bytes)
Churning 1000000 records:  33026 ns (91193360 bytes)

Previous:
$ ./speed 1000000
Adding 1000000 records:  28324 ns (67232528 bytes)
Finding 1000000 records:  2468 ns (67232528 bytes)
Traversing 1000000 records:  2200 ns (67232528 bytes)
Deleting 1000000 records:  13083 ns (67232528 bytes)
Re-adding 1000000 records:  16433 ns (67232528 bytes)
Appending 1000000 records:  2511 ns (67232528 bytes)
Churning 1000000 records:  31068 ns (67570448 bytes)

ccan/tdb2/check.c
ccan/tdb2/free.c
ccan/tdb2/private.h
ccan/tdb2/summary.c
ccan/tdb2/test/layout.c
ccan/tdb2/test/run-03-coalesce.c

index c5450cccb0cd17d9e8ba248db18fb272b152023c..97347920291a6376f3010ec6b7289f7190607827 100644 (file)
@@ -315,37 +315,37 @@ static bool check_hash(struct tdb_context *tdb,
 static bool check_free(struct tdb_context *tdb,
                       tdb_off_t off,
                       const struct tdb_free_record *frec,
 static bool check_free(struct tdb_context *tdb,
                       tdb_off_t off,
                       const struct tdb_free_record *frec,
-                      tdb_off_t prev, tdb_off_t flist_off, unsigned int bucket)
+                      tdb_off_t prev, unsigned int flist, unsigned int bucket)
 {
        if (frec_magic(frec) != TDB_FREE_MAGIC) {
                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                         "tdb_check: offset %llu bad magic 0x%llx\n",
 {
        if (frec_magic(frec) != TDB_FREE_MAGIC) {
                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                         "tdb_check: offset %llu bad magic 0x%llx\n",
-                        (long long)off, (long long)frec->magic_and_meta);
+                        (long long)off, (long long)frec->magic_and_prev);
                return false;
        }
                return false;
        }
-       if (frec_flist(frec) != flist_off) {
+       if (frec_flist(frec) != flist) {
                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
-                        "tdb_check: offset %llu bad freelist 0x%llx\n",
-                        (long long)off, (long long)frec_flist(frec));
+                        "tdb_check: offset %llu bad freelist %u\n",
+                        (long long)off, frec_flist(frec));
                return false;
        }
 
        if (tdb->methods->oob(tdb, off
                return false;
        }
 
        if (tdb->methods->oob(tdb, off
-                             + frec->data_len+sizeof(struct tdb_used_record),
+                             + frec_len(frec) + sizeof(struct tdb_used_record),
                              false))
                return false;
                              false))
                return false;
-       if (size_to_bucket(frec->data_len) != bucket) {
+       if (size_to_bucket(frec_len(frec)) != bucket) {
                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                         "tdb_check: offset %llu in wrong bucket %u vs %u\n",
                         (long long)off,
                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                         "tdb_check: offset %llu in wrong bucket %u vs %u\n",
                         (long long)off,
-                        bucket, size_to_bucket(frec->data_len));
+                        bucket, size_to_bucket(frec_len(frec)));
                return false;
        }
                return false;
        }
-       if (prev != frec->prev) {
+       if (prev != frec_prev(frec)) {
                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                         "tdb_check: offset %llu bad prev %llu vs %llu\n",
                         (long long)off,
                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                         "tdb_check: offset %llu bad prev %llu vs %llu\n",
                         (long long)off,
-                        (long long)prev, (long long)frec->prev);
+                        (long long)prev, (long long)frec_len(frec));
                return false;
        }
        return true;
                return false;
        }
        return true;
@@ -353,6 +353,7 @@ static bool check_free(struct tdb_context *tdb,
                       
 static bool check_free_list(struct tdb_context *tdb,
                            tdb_off_t flist_off,
                       
 static bool check_free_list(struct tdb_context *tdb,
                            tdb_off_t flist_off,
+                           unsigned flist_num,
                            tdb_off_t free[],
                            size_t num_free,
                            size_t *num_found)
                            tdb_off_t free[],
                            size_t num_free,
                            size_t *num_found)
@@ -384,7 +385,7 @@ static bool check_free_list(struct tdb_context *tdb,
                                return false;
                        if (tdb_read_convert(tdb, off, &f, sizeof(f)))
                                return false;
                                return false;
                        if (tdb_read_convert(tdb, off, &f, sizeof(f)))
                                return false;
-                       if (!check_free(tdb, off, &f, prev, flist_off, i))
+                       if (!check_free(tdb, off, &f, prev, flist_num, i))
                                return false;
 
                        /* FIXME: Check hash bits */
                                return false;
 
                        /* FIXME: Check hash bits */
@@ -436,13 +437,17 @@ static bool check_linear(struct tdb_context *tdb,
                        struct tdb_free_record f;
                        struct tdb_recovery_record r;
                } pad, *p;
                        struct tdb_free_record f;
                        struct tdb_recovery_record r;
                } pad, *p;
-               p = tdb_get(tdb, off, &pad, sizeof(pad));
+               /* r is larger: only get that if we need to. */
+               p = tdb_get(tdb, off, &pad, sizeof(pad.f));
                if (!p)
                        return false;
 
                /* If we crash after ftruncate, we can get zeroes or fill. */
                if (p->r.magic == TDB_RECOVERY_INVALID_MAGIC
                    || p->r.magic ==  0x4343434343434343ULL) {
                if (!p)
                        return false;
 
                /* If we crash after ftruncate, we can get zeroes or fill. */
                if (p->r.magic == TDB_RECOVERY_INVALID_MAGIC
                    || p->r.magic ==  0x4343434343434343ULL) {
+                       p = tdb_get(tdb, off, &pad, sizeof(pad.r));
+                       if (!p)
+                               return false;
                        if (recovery == off) {
                                found_recovery = true;
                                len = sizeof(p->r) + p->r.max_len;
                        if (recovery == off) {
                                found_recovery = true;
                                len = sizeof(p->r) + p->r.max_len;
@@ -462,6 +467,9 @@ static bool check_linear(struct tdb_context *tdb,
                                         (size_t)tdb->map_size);
                        }
                } else if (p->r.magic == TDB_RECOVERY_MAGIC) {
                                         (size_t)tdb->map_size);
                        }
                } else if (p->r.magic == TDB_RECOVERY_MAGIC) {
+                       p = tdb_get(tdb, off, &pad, sizeof(pad.r));
+                       if (!p)
+                               return false;
                        if (recovery != off) {
                                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                                         "tdb_check: unexpected recovery"
                        if (recovery != off) {
                                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                                         "tdb_check: unexpected recovery"
@@ -469,11 +477,23 @@ static bool check_linear(struct tdb_context *tdb,
                                         (size_t)off);
                                return false;
                        }
                                         (size_t)off);
                                return false;
                        }
+                       if (p->r.len > p->r.max_len) {
+                               tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
+                                        "tdb_check: invalid recovery length"
+                                        " %zu\n", (size_t)p->r.len);
+                               return false;
+                       }
+                       if (p->r.eof > tdb->map_size) {
+                               tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
+                                        "tdb_check: invalid old EOF"
+                                        " %zu\n", (size_t)p->r.eof);
+                               return false;
+                       }
                        found_recovery = true;
                        len = sizeof(p->r) + p->r.max_len;
                } else if (frec_magic(&p->f) == TDB_FREE_MAGIC
                           || frec_magic(&p->f) == TDB_COALESCING_MAGIC) {
                        found_recovery = true;
                        len = sizeof(p->r) + p->r.max_len;
                } else if (frec_magic(&p->f) == TDB_FREE_MAGIC
                           || frec_magic(&p->f) == TDB_COALESCING_MAGIC) {
-                       len = sizeof(p->u) + p->f.data_len;
+                       len = sizeof(p->u) + frec_len(&p->f);
                        if (off + len > tdb->map_size) {
                                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                                         "tdb_check: free overlength %llu"
                        if (off + len > tdb->map_size) {
                                tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                                         "tdb_check: free overlength %llu"
@@ -560,7 +580,8 @@ int tdb_check(struct tdb_context *tdb,
        for (flist = first_flist(tdb); flist; flist = next_flist(tdb, flist)) {
                if (flist == TDB_OFF_ERR)
                        goto fail;
        for (flist = first_flist(tdb); flist; flist = next_flist(tdb, flist)) {
                if (flist == TDB_OFF_ERR)
                        goto fail;
-               if (!check_free_list(tdb, flist, free, num_free, &num_found))
+               if (!check_free_list(tdb, flist, num_flists, free, num_free,
+                                    &num_found))
                        goto fail;
                num_flists++;
        }
                        goto fail;
                num_flists++;
        }
index 6fd82c5c540d0ee2767d60385c43654606396c8b..13db5933911073577e42d1175fcecd5b7da12efb 100644 (file)
@@ -62,10 +62,11 @@ tdb_off_t next_flist(struct tdb_context *tdb, tdb_off_t flist)
 int tdb_flist_init(struct tdb_context *tdb)
 {
        /* Use reservoir sampling algorithm to select a free list at random. */
 int tdb_flist_init(struct tdb_context *tdb)
 {
        /* Use reservoir sampling algorithm to select a free list at random. */
-       unsigned int rnd, max = 0;
+       unsigned int rnd, max = 0, count = 0;
        tdb_off_t off;
 
        tdb->flist_off = off = first_flist(tdb);
        tdb_off_t off;
 
        tdb->flist_off = off = first_flist(tdb);
+       tdb->flist = 0;
 
        while (off) {
                if (off == TDB_OFF_ERR)
 
        while (off) {
                if (off == TDB_OFF_ERR)
@@ -74,10 +75,12 @@ int tdb_flist_init(struct tdb_context *tdb)
                rnd = random();
                if (rnd >= max) {
                        tdb->flist_off = off;
                rnd = random();
                if (rnd >= max) {
                        tdb->flist_off = off;
+                       tdb->flist = count;
                        max = rnd;
                }
 
                off = next_flist(tdb, off);
                        max = rnd;
                }
 
                off = next_flist(tdb, off);
+               count++;
        }
        return 0;
 }
        }
        return 0;
 }
@@ -107,10 +110,10 @@ static int remove_from_list(struct tdb_context *tdb,
        tdb_off_t off;
 
        /* Front of list? */
        tdb_off_t off;
 
        /* Front of list? */
-       if (r->prev == 0) {
+       if (frec_prev(r) == 0) {
                off = b_off;
        } else {
                off = b_off;
        } else {
-               off = r->prev + offsetof(struct tdb_free_record, next);
+               off = frec_prev(r) + offsetof(struct tdb_free_record, next);
        }
 
 #ifdef DEBUG
        }
 
 #ifdef DEBUG
@@ -128,11 +131,11 @@ static int remove_from_list(struct tdb_context *tdb,
        }
 
        if (r->next != 0) {
        }
 
        if (r->next != 0) {
-               off = r->next + offsetof(struct tdb_free_record, prev);
+               off = r->next + offsetof(struct tdb_free_record,magic_and_prev);
                /* r->next->prev = r->prev */
 
 #ifdef DEBUG
                /* r->next->prev = r->prev */
 
 #ifdef DEBUG
-               if (tdb_read_off(tdb, off) != r_off) {
+               if (tdb_read_off(tdb, off) & TDB_OFF_MASK != r_off) {
                        tdb->log(tdb, TDB_DEBUG_FATAL, tdb->log_priv,
                                 "remove_from_list: %llu bad list %llu\n",
                                 (long long)r_off, (long long)b_off);
                        tdb->log(tdb, TDB_DEBUG_FATAL, tdb->log_priv,
                                 "remove_from_list: %llu bad list %llu\n",
                                 (long long)r_off, (long long)b_off);
@@ -140,7 +143,7 @@ static int remove_from_list(struct tdb_context *tdb,
                }
 #endif
 
                }
 #endif
 
-               if (tdb_write_off(tdb, off, r->prev)) {
+               if (tdb_write_off(tdb, off, r->magic_and_prev)) {
                        return -1;
                }
        }
                        return -1;
                }
        }
@@ -151,58 +154,65 @@ static int remove_from_list(struct tdb_context *tdb,
 static int enqueue_in_free(struct tdb_context *tdb,
                           tdb_off_t b_off,
                           tdb_off_t off,
 static int enqueue_in_free(struct tdb_context *tdb,
                           tdb_off_t b_off,
                           tdb_off_t off,
-                          struct tdb_free_record *new)
+                          tdb_len_t len)
 {
 {
-       new->prev = 0;
+       struct tdb_free_record new;
+       uint64_t magic = (TDB_FREE_MAGIC << (64 - TDB_OFF_UPPER_STEAL));
+
+       /* We only need to set flist_and_len; rest is set in enqueue_in_free */
+       new.flist_and_len = ((uint64_t)tdb->flist << (64 - TDB_OFF_UPPER_STEAL))
+               | len;
+       /* prev = 0. */
+       new.magic_and_prev = magic;
+
        /* new->next = head. */
        /* new->next = head. */
-       new->next = tdb_read_off(tdb, b_off);
-       if (new->next == TDB_OFF_ERR)
+       new.next = tdb_read_off(tdb, b_off);
+       if (new.next == TDB_OFF_ERR)
                return -1;
 
                return -1;
 
-       if (new->next) {
+       if (new.next) {
 #ifdef DEBUG
                if (tdb_read_off(tdb,
 #ifdef DEBUG
                if (tdb_read_off(tdb,
-                                new->next
-                                + offsetof(struct tdb_free_record, prev))
-                   != 0) {
+                                new.next + offsetof(struct tdb_free_record,
+                                                    magic_and_prev))
+                   != magic) {
                        tdb->log(tdb, TDB_DEBUG_FATAL, tdb->log_priv,
                                 "enqueue_in_free: %llu bad head prev %llu\n",
                        tdb->log(tdb, TDB_DEBUG_FATAL, tdb->log_priv,
                                 "enqueue_in_free: %llu bad head prev %llu\n",
-                                (long long)new->next, (long long)b_off);
+                                (long long)new.next, (long long)b_off);
                        return -1;
                }
 #endif
                /* next->prev = new. */
                        return -1;
                }
 #endif
                /* next->prev = new. */
-               if (tdb_write_off(tdb, new->next
-                                 + offsetof(struct tdb_free_record, prev),
-                                 off) != 0)
+               if (tdb_write_off(tdb, new.next
+                                 + offsetof(struct tdb_free_record,
+                                            magic_and_prev),
+                                 off | magic) != 0)
                        return -1;
        }
        /* head = new */
        if (tdb_write_off(tdb, b_off, off) != 0)
                return -1;
 
                        return -1;
        }
        /* head = new */
        if (tdb_write_off(tdb, b_off, off) != 0)
                return -1;
 
-       return tdb_write_convert(tdb, off, new, sizeof(*new));
+       return tdb_write_convert(tdb, off, &new, sizeof(new));
 }
 
 /* List need not be locked. */
 int add_free_record(struct tdb_context *tdb,
                    tdb_off_t off, tdb_len_t len_with_header)
 {
 }
 
 /* List need not be locked. */
 int add_free_record(struct tdb_context *tdb,
                    tdb_off_t off, tdb_len_t len_with_header)
 {
-       struct tdb_free_record new;
        tdb_off_t b_off;
        tdb_off_t b_off;
+       tdb_len_t len;
        int ret;
 
        int ret;
 
-       assert(len_with_header >= sizeof(new));
+       assert(len_with_header >= sizeof(struct tdb_free_record));
 
 
-       new.magic_and_meta = TDB_FREE_MAGIC << (64 - TDB_OFF_UPPER_STEAL)
-               | tdb->flist_off;
-       new.data_len = len_with_header - sizeof(struct tdb_used_record);
+       len = len_with_header - sizeof(struct tdb_used_record);
 
 
-       b_off = bucket_off(tdb->flist_off, size_to_bucket(new.data_len));
+       b_off = bucket_off(tdb->flist_off, size_to_bucket(len));
        if (tdb_lock_free_bucket(tdb, b_off, TDB_LOCK_WAIT) != 0)
                return -1;
 
        if (tdb_lock_free_bucket(tdb, b_off, TDB_LOCK_WAIT) != 0)
                return -1;
 
-       ret = enqueue_in_free(tdb, b_off, off, &new);
+       ret = enqueue_in_free(tdb, b_off, off, len);
        tdb_unlock_free_bucket(tdb, b_off);
        return ret;
 }
        tdb_unlock_free_bucket(tdb, b_off);
        return ret;
 }
@@ -234,6 +244,17 @@ static size_t record_leftover(size_t keylen, size_t datalen,
        return leftover;
 }
 
        return leftover;
 }
 
+/* FIXME: Shortcut common case where tdb->flist == flist */
+static tdb_off_t flist_offset(struct tdb_context *tdb, unsigned int flist)
+{
+       tdb_off_t off = first_flist(tdb);
+       unsigned int i;
+
+       for (i = 0; i < flist; i++)
+               off = next_flist(tdb, off);
+       return off;
+}
+
 /* Note: we unlock the current bucket if we coalesce or fail. */
 static int coalesce(struct tdb_context *tdb,
                    tdb_off_t off, tdb_off_t b_off, tdb_len_t data_len)
 /* Note: we unlock the current bucket if we coalesce or fail. */
 static int coalesce(struct tdb_context *tdb,
                    tdb_off_t off, tdb_off_t b_off, tdb_len_t data_len)
@@ -245,6 +266,7 @@ static int coalesce(struct tdb_context *tdb,
 
        while (end < tdb->map_size) {
                tdb_off_t nb_off;
 
        while (end < tdb->map_size) {
                tdb_off_t nb_off;
+               unsigned flist, bucket;
 
                /* FIXME: do tdb_get here and below really win? */
                r = tdb_get(tdb, end, &pad, sizeof(pad));
 
                /* FIXME: do tdb_get here and below really win? */
                r = tdb_get(tdb, end, &pad, sizeof(pad));
@@ -254,7 +276,9 @@ static int coalesce(struct tdb_context *tdb,
                if (frec_magic(r) != TDB_FREE_MAGIC)
                        break;
 
                if (frec_magic(r) != TDB_FREE_MAGIC)
                        break;
 
-               nb_off = bucket_off(frec_flist(r), size_to_bucket(r->data_len));
+               flist = frec_flist(r);
+               bucket = size_to_bucket(frec_len(r));
+               nb_off = bucket_off(flist_offset(tdb, flist), bucket);
 
                /* We may be violating lock order here, so best effort. */
                if (tdb_lock_free_bucket(tdb, nb_off, TDB_LOCK_NOWAIT) == -1)
 
                /* We may be violating lock order here, so best effort. */
                if (tdb_lock_free_bucket(tdb, nb_off, TDB_LOCK_NOWAIT) == -1)
@@ -272,9 +296,8 @@ static int coalesce(struct tdb_context *tdb,
                        break;
                }
 
                        break;
                }
 
-               if (unlikely(bucket_off(frec_flist(r),
-                                       size_to_bucket(r->data_len))
-                            != nb_off)) {
+               if (unlikely(frec_flist(r) != flist)
+                   || unlikely(size_to_bucket(frec_len(r)) != bucket)) {
                        tdb_unlock_free_bucket(tdb, nb_off);
                        break;
                }
                        tdb_unlock_free_bucket(tdb, nb_off);
                        break;
                }
@@ -284,7 +307,7 @@ static int coalesce(struct tdb_context *tdb,
                        goto err;
                }
 
                        goto err;
                }
 
-               end += sizeof(struct tdb_used_record) + r->data_len;
+               end += sizeof(struct tdb_used_record) + frec_len(r);
                tdb_unlock_free_bucket(tdb, nb_off);
        }
 
                tdb_unlock_free_bucket(tdb, nb_off);
        }
 
@@ -297,11 +320,11 @@ static int coalesce(struct tdb_context *tdb,
        if (!r)
                goto err;
 
        if (!r)
                goto err;
 
-       if (r->data_len != data_len) {
+       if (frec_len(r) != data_len) {
                tdb->ecode = TDB_ERR_CORRUPT;
                tdb->log(tdb, TDB_DEBUG_FATAL, tdb->log_priv,
                         "coalesce: expected data len %llu not %llu\n",
                tdb->ecode = TDB_ERR_CORRUPT;
                tdb->log(tdb, TDB_DEBUG_FATAL, tdb->log_priv,
                         "coalesce: expected data len %llu not %llu\n",
-                        (long long)data_len, (long long)r->data_len);
+                        (long long)data_len, (long long)frec_len(r));
                goto err;
        }
 
                goto err;
        }
 
@@ -314,8 +337,9 @@ static int coalesce(struct tdb_context *tdb,
 
        /* We have to drop this to avoid deadlocks, so make sure record
         * doesn't get coalesced by someone else! */
 
        /* We have to drop this to avoid deadlocks, so make sure record
         * doesn't get coalesced by someone else! */
-       r->magic_and_meta = TDB_COALESCING_MAGIC << (64 - TDB_OFF_UPPER_STEAL);
-       r->data_len = end - off - sizeof(struct tdb_used_record);
+       r->magic_and_prev = TDB_COALESCING_MAGIC << (64 - TDB_OFF_UPPER_STEAL);
+       /* FIXME: Use 255 as invalid free list? */
+       r->flist_and_len = end - off - sizeof(struct tdb_used_record);
        if (tdb_access_commit(tdb, r) != 0)
                goto err;
 
        if (tdb_access_commit(tdb, r) != 0)
                goto err;
 
@@ -353,7 +377,7 @@ again:
                return TDB_OFF_ERR;
        }
 
                return TDB_OFF_ERR;
        }
 
-       best.data_len = -1ULL;
+       best.flist_and_len = -1ULL;
        best_off = 0;
 
        /* Get slack if we're after extra. */
        best_off = 0;
 
        /* Get slack if we're after extra. */
@@ -377,22 +401,22 @@ again:
                if (frec_magic(r) != TDB_FREE_MAGIC) {
                        tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                                 "lock_and_alloc: %llu non-free 0x%llx\n",
                if (frec_magic(r) != TDB_FREE_MAGIC) {
                        tdb->log(tdb, TDB_DEBUG_ERROR, tdb->log_priv,
                                 "lock_and_alloc: %llu non-free 0x%llx\n",
-                                (long long)off, (long long)r->magic_and_meta);
+                                (long long)off, (long long)r->magic_and_prev);
                        goto unlock_err;
                }
 
                        goto unlock_err;
                }
 
-               if (r->data_len >= size && r->data_len < best.data_len) {
+               if (frec_len(r) >= size && frec_len(r) < frec_len(&best)) {
                        best_off = off;
                        best = *r;
                }
 
                        best_off = off;
                        best = *r;
                }
 
-               if (best.data_len < size * multiplier && best_off)
+               if (frec_len(&best) < size * multiplier && best_off)
                        break;
 
                multiplier *= 1.01;
 
                /* Since we're going slow anyway, try coalescing here. */
                        break;
 
                multiplier *= 1.01;
 
                /* Since we're going slow anyway, try coalescing here. */
-               switch (coalesce(tdb, off, b_off, r->data_len)) {
+               switch (coalesce(tdb, off, b_off, frec_len(r))) {
                case -1:
                        /* This has already unlocked on error. */
                        return -1;
                case -1:
                        /* This has already unlocked on error. */
                        return -1;
@@ -413,13 +437,13 @@ again:
                        goto unlock_err;
 
                leftover = record_leftover(keylen, datalen, want_extra,
                        goto unlock_err;
 
                leftover = record_leftover(keylen, datalen, want_extra,
-                                          best.data_len);
+                                          frec_len(&best));
 
 
-               assert(keylen + datalen + leftover <= best.data_len);
+               assert(keylen + datalen + leftover <= frec_len(&best));
                /* We need to mark non-free before we drop lock, otherwise
                 * coalesce() could try to merge it! */
                if (set_used_header(tdb, &rec, keylen, datalen,
                /* We need to mark non-free before we drop lock, otherwise
                 * coalesce() could try to merge it! */
                if (set_used_header(tdb, &rec, keylen, datalen,
-                                   best.data_len - leftover,
+                                   frec_len(&best) - leftover,
                                    hashlow) != 0)
                        goto unlock_err;
 
                                    hashlow) != 0)
                        goto unlock_err;
 
@@ -431,7 +455,7 @@ again:
                if (leftover) {
                        if (add_free_record(tdb,
                                            best_off + sizeof(rec)
                if (leftover) {
                        if (add_free_record(tdb,
                                            best_off + sizeof(rec)
-                                           + best.data_len - leftover,
+                                           + frec_len(&best) - leftover,
                                            leftover))
                                return TDB_OFF_ERR;
                }
                                            leftover))
                                return TDB_OFF_ERR;
                }
@@ -451,8 +475,8 @@ static tdb_off_t get_free(struct tdb_context *tdb,
                          size_t keylen, size_t datalen, bool want_extra,
                          unsigned hashlow)
 {
                          size_t keylen, size_t datalen, bool want_extra,
                          unsigned hashlow)
 {
-       tdb_off_t off, flist;
-       unsigned start_b, b;
+       tdb_off_t off, flist_off;
+       unsigned start_b, b, flist;
        bool wrapped = false;
 
        /* If they are growing, add 50% to get to higher bucket. */
        bool wrapped = false;
 
        /* If they are growing, add 50% to get to higher bucket. */
@@ -462,31 +486,35 @@ static tdb_off_t get_free(struct tdb_context *tdb,
        else
                start_b = size_to_bucket(adjust_size(keylen, datalen));
 
        else
                start_b = size_to_bucket(adjust_size(keylen, datalen));
 
-       flist = tdb->flist_off;
-       while (!wrapped || flist != tdb->flist_off) {
+       flist_off = tdb->flist_off;
+       flist = tdb->flist;
+       while (!wrapped || flist_off != tdb->flist_off) {
                /* Start at exact size bucket, and search up... */
                /* Start at exact size bucket, and search up... */
-               for (b = find_free_head(tdb, flist, start_b);
+               for (b = find_free_head(tdb, flist_off, start_b);
                     b < TDB_FREE_BUCKETS;
                     b < TDB_FREE_BUCKETS;
-                    b = find_free_head(tdb, flist, b + 1)) {
+                    b = find_free_head(tdb, flist_off, b + 1)) {
                        /* Try getting one from list. */
                        /* Try getting one from list. */
-                       off = lock_and_alloc(tdb, flist,
+                       off = lock_and_alloc(tdb, flist_off,
                                             b, keylen, datalen, want_extra,
                                             hashlow);
                        if (off == TDB_OFF_ERR)
                                return TDB_OFF_ERR;
                        if (off != 0) {
                                /* Worked?  Stay using this list. */
                                             b, keylen, datalen, want_extra,
                                             hashlow);
                        if (off == TDB_OFF_ERR)
                                return TDB_OFF_ERR;
                        if (off != 0) {
                                /* Worked?  Stay using this list. */
-                               tdb->flist_off = flist;
+                               tdb->flist_off = flist_off;
+                               tdb->flist = flist;
                                return off;
                        }
                        /* Didn't work.  Try next bucket. */
                }
 
                /* Hmm, try next list. */
                                return off;
                        }
                        /* Didn't work.  Try next bucket. */
                }
 
                /* Hmm, try next list. */
-               flist = next_flist(tdb, flist);
-               if (flist == 0) {
+               flist_off = next_flist(tdb, flist_off);
+               flist++;
+               if (flist_off == 0) {
                        wrapped = true;
                        wrapped = true;
-                       flist = first_flist(tdb);
+                       flist_off = first_flist(tdb);
+                       flist = 0;
                }
        }
 
                }
        }
 
index 387e49f61182af8fe33028840ebc3be073fea70a..a10d1070d29c3d5f84f09107d032fbb0cf0dac1b 100644 (file)
@@ -173,20 +173,30 @@ static inline uint16_t rec_magic(const struct tdb_used_record *r)
 }
 
 struct tdb_free_record {
 }
 
 struct tdb_free_record {
-        uint64_t magic_and_meta; /* TDB_OFF_UPPER_STEAL bits of magic */
-        uint64_t data_len; /* Not counting these two fields. */
-       /* This is why the minimum record size is 16 bytes.  */
-       uint64_t next, prev;
+        uint64_t magic_and_prev; /* TDB_OFF_UPPER_STEAL bits magic, then prev */
+        uint64_t flist_and_len; /* Len not counting these two fields. */
+       /* This is why the minimum record size is 8 bytes.  */
+       uint64_t next;
 };
 
 };
 
+static inline uint64_t frec_prev(const struct tdb_free_record *f)
+{
+       return f->magic_and_prev & ((1ULL << (64 - TDB_OFF_UPPER_STEAL)) - 1);
+}
+
 static inline uint64_t frec_magic(const struct tdb_free_record *f)
 {
 static inline uint64_t frec_magic(const struct tdb_free_record *f)
 {
-       return f->magic_and_meta >> (64 - TDB_OFF_UPPER_STEAL);
+       return f->magic_and_prev >> (64 - TDB_OFF_UPPER_STEAL);
+}
+
+static inline uint64_t frec_len(const struct tdb_free_record *f)
+{
+       return f->flist_and_len & ((1ULL << (64 - TDB_OFF_UPPER_STEAL))-1);
 }
 
 }
 
-static inline uint64_t frec_flist(const struct tdb_free_record *f)
+static inline unsigned frec_flist(const struct tdb_free_record *f)
 {
 {
-       return f->magic_and_meta & ((1ULL << (64 - TDB_OFF_UPPER_STEAL)) - 1);
+       return f->flist_and_len >> (64 - TDB_OFF_UPPER_STEAL);
 }
 
 struct tdb_recovery_record {
 }
 
 struct tdb_recovery_record {
@@ -311,6 +321,7 @@ struct tdb_context {
        
        /* What freelist are we using? */
        uint64_t flist_off;
        
        /* What freelist are we using? */
        uint64_t flist_off;
+       unsigned int flist;
 
        /* IO methods: changes for transactions. */
        const struct tdb_methods *methods;
 
        /* IO methods: changes for transactions. */
        const struct tdb_methods *methods;
index 97052608fc866735156ab5938bf827ea8f653242..c04fa76f945b809a701b77f1df428dc9b693b76b 100644 (file)
@@ -63,7 +63,7 @@ static bool summarize(struct tdb_context *tdb,
                    || p->r.magic == TDB_RECOVERY_MAGIC) {
                        len = sizeof(p->r) + p->r.max_len;
                } else if (rec_magic(&p->u) != TDB_MAGIC) {
                    || p->r.magic == TDB_RECOVERY_MAGIC) {
                        len = sizeof(p->r) + p->r.max_len;
                } else if (rec_magic(&p->u) != TDB_MAGIC) {
-                       len = p->f.data_len;
+                       len = frec_len(&p->f);
                        tally_add(free, len);
                        tally_add(buckets, size_to_bucket(len));
                        len += sizeof(p->u);
                        tally_add(free, len);
                        tally_add(buckets, size_to_bucket(len));
                        len += sizeof(p->u);
index 3049d3214e62736ec95ddbe979cafe7784034593..25f41e25c3923875acfec2cd00364b0cf6718895 100644 (file)
@@ -136,9 +136,11 @@ static void set_freelist(void *mem, struct tdb_context *tdb,
 static void add_to_freetable(struct tdb_context *tdb,
                             tdb_off_t eoff,
                             tdb_off_t elen,
 static void add_to_freetable(struct tdb_context *tdb,
                             tdb_off_t eoff,
                             tdb_off_t elen,
+                            unsigned flist,
                             struct tle_freelist *freelist)
 {
        tdb->flist_off = freelist->base.off;
                             struct tle_freelist *freelist)
 {
        tdb->flist_off = freelist->base.off;
+       tdb->flist = flist;
        add_free_record(tdb, eoff, sizeof(struct tdb_used_record) + elen);
 }
 
        add_free_record(tdb, eoff, sizeof(struct tdb_used_record) + elen);
 }
 
@@ -288,6 +290,7 @@ struct tdb_context *tdb_layout_get(struct tdb_layout *layout)
                switch (e->base.type) {
                case FREE:
                        add_to_freetable(tdb, e->base.off, e->free.len,
                switch (e->base.type) {
                case FREE:
                        add_to_freetable(tdb, e->base.off, e->free.len,
+                                        e->free.flist_num,
                                         find_flist(layout, e->free.flist_num));
                        break;
                case DATA:
                                         find_flist(layout, e->free.flist_num));
                        break;
                case DATA:
index 5d55577c4885cdfce372a1be4f0f3f0cdeb63533..4def490ffdb281c9e8e830512d90f51a38bce4eb 100644 (file)
@@ -17,7 +17,7 @@ static tdb_len_t free_record_length(struct tdb_context *tdb, tdb_off_t off)
                return TDB_OFF_ERR;
        if (frec_magic(&f) != TDB_FREE_MAGIC)
                return TDB_OFF_ERR;
                return TDB_OFF_ERR;
        if (frec_magic(&f) != TDB_FREE_MAGIC)
                return TDB_OFF_ERR;
-       return f.data_len;
+       return frec_len(&f);
 }
 
 int main(int argc, char *argv[])
 }
 
 int main(int argc, char *argv[])