#ifndef _WIN32
#include <pthread.h>
+#ifdef __APPLE__
+#include <semaphore.h>
+#endif
+#endif
+
+#ifndef BYTE_ORDER
+#define BYTE_ORDER __BYTE_ORDER
+#endif
+#ifndef LITTLE_ENDIAN
+#define LITTLE_ENDIAN __LITTLE_ENDIAN
+#endif
+#ifndef BIG_ENDIAN
+#define BIG_ENDIAN __BIG_ENDIAN
#endif
#include "mdb.h"
#include "midl.h"
-#if (__BYTE_ORDER == __LITTLE_ENDIAN) == (__BYTE_ORDER == __BIG_ENDIAN)
-# error "Unknown or unsupported endianness (__BYTE_ORDER)"
+#if (BYTE_ORDER == LITTLE_ENDIAN) == (BYTE_ORDER == BIG_ENDIAN)
+# error "Unknown or unsupported endianness (BYTE_ORDER)"
#elif (-6 & 5) || CHAR_BIT != 8 || UINT_MAX < 0xffffffff || ULONG_MAX % 0xFFFF
# error "Two's complement, reasonably sized integer types, please"
#endif
#define GET_PAGESIZE(x) {SYSTEM_INFO si; GetSystemInfo(&si); (x) = si.dwPageSize;}
#define close(fd) CloseHandle(fd)
#define munmap(ptr,len) UnmapViewOfFile(ptr)
+#else
+#ifdef __APPLE__
+#define LOCK_MUTEX_R(env) sem_wait((env)->me_rmutex)
+#define UNLOCK_MUTEX_R(env) sem_post((env)->me_rmutex)
+#define LOCK_MUTEX_W(env) sem_wait((env)->me_wmutex)
+#define UNLOCK_MUTEX_W(env) sem_post((env)->me_wmutex)
+#define fdatasync(fd) fsync(fd)
#else
/** Lock the reader mutex.
*/
/** Unlock the writer mutex.
*/
#define UNLOCK_MUTEX_W(env) pthread_mutex_unlock(&(env)->me_txns->mti_wmutex)
+#endif /* __APPLE__ */
/** Get the error code for the last failed system function.
*/
#define GET_PAGESIZE(x) ((x) = sysconf(_SC_PAGE_SIZE))
#endif
+#if defined(_WIN32) || defined(__APPLE__)
+#define MNAME_LEN 32
+#endif
+
/** @} */
#ifndef _WIN32
uint32_t mtb_magic;
/** Version number of this lock file. Must be set to #MDB_VERSION. */
uint32_t mtb_version;
-#ifdef _WIN32
- char mtb_rmname[32];
+#if defined(_WIN32) || defined(__APPLE__)
+ char mtb_rmname[MNAME_LEN];
#else
/** Mutex protecting access to this table.
* This is the reader lock that #LOCK_MUTEX_R acquires.
char pad[(sizeof(MDB_txbody)+CACHELINE-1) & ~(CACHELINE-1)];
} mt1;
union {
-#ifdef _WIN32
- char mt2_wmname[32];
+#if defined(_WIN32) || defined(__APPLE__)
+ char mt2_wmname[MNAME_LEN];
#define mti_wmname mt2.mt2_wmname
#else
pthread_mutex_t mt2_wmutex;
/** @} */
/** Common header for all page types.
- * Overflow pages occupy a number of contiguous pages with no
+ * Overflow records occupy a number of contiguous pages with no
* headers on any page after the first.
*/
typedef struct MDB_page {
#define mp_pgno mp_p.p_pgno
#define mp_next mp_p.p_next
- union padded {
+ union {
pgno_t p_pgno; /**< page number */
void * p_next; /**< for in-memory list of freed structs */
} mp_p;
+/** @defgroup mdb_page Page Flags
+ * @ingroup internal
+ * Flags for the page headers.
+ * @{
+ */
#define P_BRANCH 0x01 /**< branch page */
#define P_LEAF 0x02 /**< leaf page */
#define P_OVERFLOW 0x04 /**< overflow page */
#define P_META 0x08 /**< meta page */
#define P_DIRTY 0x10 /**< dirty page */
#define P_LEAF2 0x20 /**< for #MDB_DUPFIXED records */
- uint32_t mp_flags;
+/** @} */
+ uint32_t mp_flags; /**< @ref mdb_page */
#define mp_lower mp_pb.pb.pb_lower
#define mp_upper mp_pb.pb.pb_upper
#define mp_pages mp_pb.pb_pages
- union page_bounds {
+ union {
struct {
indx_t pb_lower; /**< lower bound of free space */
indx_t pb_upper; /**< upper bound of free space */
/** lo and hi are used for data size on leaf nodes and for
* child pgno on branch nodes. On 64 bit platforms, flags
* is also used for pgno. (Branch nodes have no flags).
- * They are in in host byte order in case that lets some
+ * They are in host byte order in case that lets some
* accesses be optimized into a 32-bit word access.
*/
-#define mn_lo mn_offset[__BYTE_ORDER!=__LITTLE_ENDIAN]
-#define mn_hi mn_offset[__BYTE_ORDER==__LITTLE_ENDIAN] /**< part of dsize or pgno */
- unsigned short mn_offset[2];
- unsigned short mn_flags; /**< flags for special node types */
+#define mn_lo mn_offset[BYTE_ORDER!=LITTLE_ENDIAN]
+#define mn_hi mn_offset[BYTE_ORDER==LITTLE_ENDIAN] /**< part of dsize or pgno */
+ unsigned short mn_offset[2]; /**< storage for #mn_lo and #mn_hi */
+/** @defgroup mdb_node Node Flags
+ * @ingroup internal
+ * Flags for node headers.
+ * @{
+ */
#define F_BIGDATA 0x01 /**< data put on overflow page */
#define F_SUBDATA 0x02 /**< data is a sub-database */
#define F_DUPDATA 0x04 /**< data has duplicates */
+/** @} */
+ unsigned short mn_flags; /**< @ref mdb_node */
unsigned short mn_ksize; /**< key size */
char mn_data[1]; /**< key and data are appended here */
} MDB_node;
*/
MDB_dbi mt_numdbs;
+/** @defgroup mdb_txn Transaction Flags
+ * @ingroup internal
+ * @{
+ */
#define MDB_TXN_RDONLY 0x01 /**< read-only transaction */
#define MDB_TXN_ERROR 0x02 /**< an error has occurred */
- unsigned int mt_flags;
+/** @} */
+ unsigned int mt_flags; /**< @ref mdb_txn */
/** Tracks which of the two meta pages was used at the start
* of this transaction.
*/
MDB_dbx *mc_dbx;
unsigned short mc_snum; /**< number of pushed pages */
unsigned short mc_top; /**< index of top page, mc_snum-1 */
- unsigned int mc_flags;
+/** @defgroup mdb_cursor Cursor Flags
+ * @ingroup internal
+ * Cursor state flags.
+ * @{
+ */
#define C_INITIALIZED 0x01 /**< cursor has been initialized and is valid */
#define C_EOF 0x02 /**< No more data */
-#define C_XDIRTY 0x04 /**< @deprecated mc_xcursor needs to be flushed */
+/** @} */
+ unsigned int mc_flags; /**< @ref mdb_cursor */
MDB_page *mc_pg[CURSOR_STACK]; /**< stack of pushed pages */
indx_t mc_ki[CURSOR_STACK]; /**< stack of page indices */
};
HANDLE me_mfd; /**< just for writing the meta pages */
/** Failed to update the meta page. Probably an I/O error. */
#define MDB_FATAL_ERROR 0x80000000U
- uint32_t me_flags;
+ uint32_t me_flags; /**< @ref mdb_env */
uint32_t me_extrapad; /**< unused for now */
unsigned int me_maxreaders; /**< size of the reader table */
MDB_dbi me_numdbs; /**< number of DBs opened */
HANDLE me_rmutex; /* Windows mutexes don't reside in shared mem */
HANDLE me_wmutex;
#endif
+#ifdef __APPLE__
+ sem_t *me_rmutex; /* Apple doesn't support shared mutexes */
+ sem_t *me_wmutex;
+#endif
};
/** max number of pages to commit in one writev() call */
#define MDB_COMMIT_PAGES 64
static MDB_page *mdb_page_alloc(MDB_cursor *mc, int num);
+static MDB_page *mdb_page_new(MDB_cursor *mc, uint32_t flags, int num);
static int mdb_page_touch(MDB_cursor *mc);
+static int mdb_page_get(MDB_txn *txn, pgno_t pgno, MDB_page **mp);
static int mdb_page_search_root(MDB_cursor *mc,
MDB_val *key, int modify);
static int mdb_page_search(MDB_cursor *mc,
MDB_val *key, int modify);
+static int mdb_page_merge(MDB_cursor *csrc, MDB_cursor *cdst);
+static int mdb_page_split(MDB_cursor *mc, MDB_val *newkey, MDB_val *newdata,
+ pgno_t newpgno);
static int mdb_env_read_header(MDB_env *env, MDB_meta *meta);
static int mdb_env_read_meta(MDB_env *env, int *which);
static int mdb_env_write_meta(MDB_txn *txn);
-static int mdb_page_get(MDB_txn *txn, pgno_t pgno, MDB_page **mp);
static MDB_node *mdb_node_search(MDB_cursor *mc, MDB_val *key, int *exactp);
static int mdb_node_add(MDB_cursor *mc, indx_t indx,
MDB_val *key, MDB_val *data, pgno_t pgno, uint8_t flags);
static void mdb_node_del(MDB_page *mp, indx_t indx, int ksize);
-static int mdb_del0(MDB_cursor *mc, MDB_node *leaf);
+static int mdb_node_move(MDB_cursor *csrc, MDB_cursor *cdst);
static int mdb_node_read(MDB_txn *txn, MDB_node *leaf, MDB_val *data);
+static size_t mdb_leaf_size(MDB_env *env, MDB_val *key, MDB_val *data);
+static size_t mdb_branch_size(MDB_env *env, MDB_val *key);
static int mdb_rebalance(MDB_cursor *mc);
static int mdb_update_key(MDB_page *mp, indx_t indx, MDB_val *key);
-static int mdb_node_move(MDB_cursor *csrc, MDB_cursor *cdst);
-static int mdb_page_merge(MDB_cursor *csrc, MDB_cursor *cdst);
-static int mdb_page_split(MDB_cursor *mc, MDB_val *newkey, MDB_val *newdata,
- pgno_t newpgno);
-static MDB_page *mdb_page_new(MDB_cursor *mc, uint32_t flags, int num);
static void mdb_cursor_pop(MDB_cursor *mc);
static int mdb_cursor_push(MDB_cursor *mc, MDB_page *mp);
+static int mdb_cursor_del0(MDB_cursor *mc, MDB_node *leaf);
static int mdb_cursor_sibling(MDB_cursor *mc, int move_right);
static int mdb_cursor_next(MDB_cursor *mc, MDB_val *key, MDB_val *data, MDB_cursor_op op);
static int mdb_cursor_prev(MDB_cursor *mc, MDB_val *key, MDB_val *data, MDB_cursor_op op);
static void mdb_xcursor_init0(MDB_cursor *mc);
static void mdb_xcursor_init1(MDB_cursor *mc, MDB_node *node);
-static size_t mdb_leaf_size(MDB_env *env, MDB_val *key, MDB_val *data);
-static size_t mdb_branch_size(MDB_env *env, MDB_val *key);
-
static void mdb_default_cmp(MDB_txn *txn, MDB_dbi dbi);
/** @cond */
}
#ifndef _WIN32
-/* Windows doesn't support destructor callbacks for thread-specific storage */
+/** Release a reader thread's slot in the reader lock table.
+ * This function is called automatically when a thread exits.
+ * Windows doesn't support destructor callbacks for thread-specific storage,
+ * so this function is not compiled there.
+ * @param[in] ptr This points to the slot in the reader lock table.
+ */
static void
mdb_env_reader_dest(void *ptr)
{
}
#endif
-/* downgrade the exclusive lock on the region back to shared */
+/** Downgrade the exclusive lock on the region back to shared */
static void
mdb_env_share_locks(MDB_env *env)
{
}
#endif
}
+#if defined(_WIN32) || defined(__APPLE__)
+/*
+ * hash_64 - 64 bit Fowler/Noll/Vo-0 FNV-1a hash code
+ *
+ * @(#) $Revision: 5.1 $
+ * @(#) $Id: hash_64a.c,v 5.1 2009/06/30 09:01:38 chongo Exp $
+ * @(#) $Source: /usr/local/src/cmd/fnv/RCS/hash_64a.c,v $
+ *
+ * http://www.isthe.com/chongo/tech/comp/fnv/index.html
+ *
+ ***
+ *
+ * Please do not copyright this code. This code is in the public domain.
+ *
+ * LANDON CURT NOLL DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
+ * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO
+ * EVENT SHALL LANDON CURT NOLL BE LIABLE FOR ANY SPECIAL, INDIRECT OR
+ * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF
+ * USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR
+ * OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
+ * PERFORMANCE OF THIS SOFTWARE.
+ *
+ * By:
+ * chongo <Landon Curt Noll> /\oo/\
+ * http://www.isthe.com/chongo/
+ *
+ * Share and Enjoy! :-)
+ */
+
+typedef unsigned long long mdb_hash_t;
+#define MDB_HASH_INIT ((mdb_hash_t)0xcbf29ce484222325ULL)
+
+/** perform a 64 bit Fowler/Noll/Vo FNV-1a hash on a buffer
+ * @param[in] str string to hash
+ * @param[in] hval initial value for hash
+ * @return 64 bit hash
+ *
+ * NOTE: To use the recommended 64 bit FNV-1a hash, use MDB_HASH_INIT as the
+ * hval arg on the first call.
+ */
+static inline mdb_hash_t
+mdb_hash_str(char *str, mdb_hash_t hval)
+{
+ unsigned char *s = (unsigned char *)str; /* unsigned string */
+ /*
+ * FNV-1a hash each octet of the string
+ */
+ while (*s) {
+ /* xor the bottom with the current octet */
+ hval ^= (mdb_hash_t)*s++;
+
+ /* multiply by the 64 bit FNV magic prime mod 2^64 */
+ hval += (hval << 1) + (hval << 4) + (hval << 5) +
+ (hval << 7) + (hval << 8) + (hval << 40);
+ }
+ /* return our new hash value */
+ return hval;
+}
+/** Hash the string and output the hash in hex.
+ * @param[in] str string to hash
+ * @param[out] hexbuf an array of 17 chars to hold the hash
+ */
+static void
+mdb_hash_hex(char *str, char *hexbuf)
+{
+ int i;
+ mdb_hash_t h = mdb_hash_str(str, MDB_HASH_INIT);
+ for (i=0; i<8; i++) {
+ hexbuf += sprintf(hexbuf, "%02x", (unsigned int)h & 0xff);
+ h >>= 8;
+ }
+}
+#endif
+
+/** Open and/or initialize the lock region for the environment.
+ * @param[in] env The MDB environment.
+ * @param[in] lpath The pathname of the file used for the lock region.
+ * @param[in] mode The Unix permissions for the file, if we create it.
+ * @param[out] excl Set to true if we got an exclusive lock on the region.
+ * @return 0 on success, non-zero on failure.
+ */
static int
mdb_env_setup_locks(MDB_env *env, char *lpath, int mode, int *excl)
{
#endif
if (*excl) {
#ifdef _WIN32
- char *ptr;
+ char hexbuf[17];
if (!mdb_sec_inited) {
InitializeSecurityDescriptor(&mdb_null_sd,
SECURITY_DESCRIPTOR_REVISION);
mdb_all_sa.lpSecurityDescriptor = &mdb_null_sd;
mdb_sec_inited = 1;
}
- /* FIXME: only using up to 20 characters of the env path here,
- * probably not enough to assure uniqueness...
- */
- sprintf(env->me_txns->mti_rmname, "Global\\MDBr%.20s", lpath);
- ptr = env->me_txns->mti_rmname + sizeof("Global\\MDBr");
- while ((ptr = strchr(ptr, '\\')))
- *ptr++ = '/';
+ mdb_hash_hex(lpath, hexbuf);
+ sprintf(env->me_txns->mti_rmname, "Global\\MDBr%s", hexbuf);
env->me_rmutex = CreateMutex(&mdb_all_sa, FALSE, env->me_txns->mti_rmname);
if (!env->me_rmutex) {
rc = ErrCode();
goto fail;
}
- sprintf(env->me_txns->mti_rmname, "Global\\MDBw%.20s", lpath);
- ptr = env->me_txns->mti_rmname + sizeof("Global\\MDBw");
- while ((ptr = strchr(ptr, '\\')))
- *ptr++ = '/';
- env->me_wmutex = CreateMutex(&mdb_all_sa, FALSE, env->me_txns->mti_rmname);
+ sprintf(env->me_txns->mti_wmname, "Global\\MDBw%s", hexbuf);
+ env->me_wmutex = CreateMutex(&mdb_all_sa, FALSE, env->me_txns->mti_wmname);
if (!env->me_wmutex) {
rc = ErrCode();
goto fail;
}
-#else
+#else /* _WIN32 */
+#ifdef __APPLE__
+ char hexbuf[17];
+ mdb_hash_hex(lpath, hexbuf);
+ sprintf(env->me_txns->mti_rmname, "MDBr%s", hexbuf);
+ if (sem_unlink(env->me_txns->mti_rmname)) {
+ rc = ErrCode();
+ if (rc != ENOENT && rc != EINVAL)
+ goto fail;
+ }
+ env->me_rmutex = sem_open(env->me_txns->mti_rmname, O_CREAT, mode, 1);
+ if (!env->me_rmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+ sprintf(env->me_txns->mti_wmname, "MDBw%s", hexbuf);
+ if (sem_unlink(env->me_txns->mti_wmname)) {
+ rc = ErrCode();
+ if (rc != ENOENT && rc != EINVAL)
+ goto fail;
+ }
+ env->me_wmutex = sem_open(env->me_txns->mti_wmname, O_CREAT, mode, 1);
+ if (!env->me_wmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+#else /* __APPLE__ */
pthread_mutexattr_t mattr;
pthread_mutexattr_init(&mattr);
}
pthread_mutex_init(&env->me_txns->mti_mutex, &mattr);
pthread_mutex_init(&env->me_txns->mti_wmutex, &mattr);
-#endif
+#endif /* __APPLE__ */
+#endif /* _WIN32 */
env->me_txns->mti_version = MDB_VERSION;
env->me_txns->mti_magic = MDB_MAGIC;
env->me_txns->mti_txnid = 0;
rc = ErrCode();
goto fail;
}
+#endif
+#ifdef __APPLE__
+ env->me_rmutex = sem_open(env->me_txns->mti_rmname, 0);
+ if (!env->me_rmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+ env->me_wmutex = sem_open(env->me_txns->mti_wmname, 0);
+ if (!env->me_wmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
#endif
}
return MDB_SUCCESS;
free(env);
}
-/* only for aligned size_t's */
+/** Compare two items pointing at aligned size_t's */
static int
mdb_cmp_long(const MDB_val *a, const MDB_val *b)
{
*(size_t *)a->mv_data > *(size_t *)b->mv_data;
}
-/* only for aligned ints */
+/** Compare two items pointing at aligned int's */
static int
mdb_cmp_int(const MDB_val *a, const MDB_val *b)
{
*(unsigned int *)a->mv_data > *(unsigned int *)b->mv_data;
}
-/* ints must always be the same size */
+/** Compare two items pointing at ints of unknown alignment.
+ * Nodes and keys are guaranteed to be 2-byte aligned.
+ */
static int
mdb_cmp_cint(const MDB_val *a, const MDB_val *b)
{
-#if __BYTE_ORDER == __LITTLE_ENDIAN
+#if BYTE_ORDER == LITTLE_ENDIAN
unsigned short *u, *c;
int x;
#endif
}
+/** Compare two items lexically */
static int
mdb_cmp_memn(const MDB_val *a, const MDB_val *b)
{
return diff ? diff : len_diff<0 ? -1 : len_diff;
}
+/** Compare two items in reverse byte order */
static int
mdb_cmp_memnr(const MDB_val *a, const MDB_val *b)
{
return len_diff<0 ? -1 : len_diff;
}
-/* Search for key within a leaf page, using binary search.
+/** Search for key within a page, using binary search.
* Returns the smallest entry larger or equal to the key.
* If exactp is non-null, stores whether the found entry was an exact match
* in *exactp (1 or 0).
- * If kip is non-null, stores the index of the found entry in *kip.
+ * Updates the cursor index with the index of the found entry.
* If no entry larger or equal to the key is found, returns NULL.
*/
static MDB_node *
return node;
}
+/** Pop a page off the top of the cursor's stack. */
static void
mdb_cursor_pop(MDB_cursor *mc)
{
}
}
+/** Push a page onto the top of the cursor's stack. */
static int
mdb_cursor_push(MDB_cursor *mc, MDB_page *mp)
{
return MDB_SUCCESS;
}
+/** Find the address of the page corresponding to a given page number.
+ * @param[in] txn the transaction for this access.
+ * @param[in] pgno the page number for the page to retrieve.
+ * @param[out] ret address of a pointer where the page's address will be stored.
+ * @return 0 on success, non-zero on failure.
+ */
static int
mdb_page_get(MDB_txn *txn, pgno_t pgno, MDB_page **ret)
{
return (p != NULL) ? MDB_SUCCESS : MDB_PAGE_NOTFOUND;
}
+/** Search for the page a given key should be in.
+ * Pushes parent pages on the cursor stack. This function continues a
+ * search on a cursor that has already been initialized. (Usually by
+ * #mdb_page_search() but also by #mdb_node_move().)
+ * @param[in,out] mc the cursor for this operation.
+ * @param[in] key the key to search for. If NULL, search for the lowest
+ * page. (This is used by #mdb_cursor_first().)
+ * @param[in] modify If true, visited pages are updated with new page numbers.
+ * @return 0 on success, non-zero on failure.
+ */
static int
mdb_page_search_root(MDB_cursor *mc, MDB_val *key, int modify)
{
return MDB_SUCCESS;
}
-/* Search for the page a given key should be in.
- * Pushes parent pages on the cursor stack.
- * If key is NULL, search for the lowest page (used by mdb_cursor_first).
- * If modify is true, visited pages are updated with new page numbers.
+/** Search for the page a given key should be in.
+ * Pushes parent pages on the cursor stack. This function just sets up
+ * the search; it finds the root page for \b mc's database and sets this
+ * as the root of the cursor's stack. Then #mdb_page_search_root() is
+ * called to complete the search.
+ * @param[in,out] mc the cursor for this operation.
+ * @param[in] key the key to search for. If NULL, search for the lowest
+ * page. (This is used by #mdb_cursor_first().)
+ * @param[in] modify If true, visited pages are updated with new page numbers.
+ * @return 0 on success, non-zero on failure.
*/
static int
mdb_page_search(MDB_cursor *mc, MDB_val *key, int modify)
return mdb_page_search_root(mc, key, modify);
}
+/** Return the data associated with a given node.
+ * @param[in] txn The transaction for this operation.
+ * @param[in] leaf The node being read.
+ * @param[out] data Updated to point to the node's data.
+ * @return 0 on success, non-zero on failure.
+ */
static int
mdb_node_read(MDB_txn *txn, MDB_node *leaf, MDB_val *data)
{
- MDB_page *omp; /* overflow mpage */
+ MDB_page *omp; /* overflow page */
pgno_t pgno;
int rc;
return mdb_cursor_set(&mc, key, data, MDB_SET, &exact);
}
+/** Find a sibling for a page.
+ * Replaces the page at the top of the cursor's stack with the
+ * specified sibling, if one exists.
+ * @param[in] mc The cursor for this operation.
+ * @param[in] move_right Non-zero if the right sibling is requested,
+ * otherwise the left sibling.
+ * @return 0 on success, non-zero on failure.
+ */
static int
mdb_cursor_sibling(MDB_cursor *mc, int move_right)
{
return MDB_SUCCESS;
}
+/** Move the cursor to the next data item. */
static int
mdb_cursor_next(MDB_cursor *mc, MDB_val *key, MDB_val *data, MDB_cursor_op op)
{
return MDB_SUCCESS;
}
+/** Move the cursor to the previous data item. */
static int
mdb_cursor_prev(MDB_cursor *mc, MDB_val *key, MDB_val *data, MDB_cursor_op op)
{
return MDB_SUCCESS;
}
+/** Set the cursor on a specific data item. */
static int
mdb_cursor_set(MDB_cursor *mc, MDB_val *key, MDB_val *data,
MDB_cursor_op op, int *exactp)
return rc;
}
+/** Move the cursor to the first item in the database. */
static int
mdb_cursor_first(MDB_cursor *mc, MDB_val *key, MDB_val *data)
{
return MDB_SUCCESS;
}
+/** Move the cursor to the last item in the database. */
static int
mdb_cursor_last(MDB_cursor *mc, MDB_val *key, MDB_val *data)
{
return rc;
}
+/** Touch all the pages in the cursor stack.
+ * Makes sure all the pages are writable, before attempting a write operation.
+ * @param[in] mc The cursor to operate on.
+ */
static int
mdb_cursor_touch(MDB_cursor *mc)
{
}
}
- return mdb_del0(mc, leaf);
+ return mdb_cursor_del0(mc, leaf);
}
-/* Allocate a page and initialize it
+/** Allocate and initialize new pages for a database.
+ * @param[in] mc a cursor on the database being added to.
+ * @param[in] flags flags defining what type of page is being allocated.
+ * @param[in] num the number of pages to allocate. This is usually 1,
+ * unless allocating overflow pages for a large record.
+ * @return Address of a page, or NULL on failure.
*/
static MDB_page *
mdb_page_new(MDB_cursor *mc, uint32_t flags, int num)
return np;
}
+/** Calculate the size of a leaf node.
+ * The size depends on the environment's page size; if a data item
+ * is too large it will be put onto an overflow page and the node
+ * size will only include the key and not the data. Sizes are always
+ * rounded up to an even number of bytes, to guarantee 2-byte alignment
+ * of the #MDB_node headers.
+ * @param[in] env The environment handle.
+ * @param[in] key The key for the node.
+ * @param[in] data The data for the node.
+ * @return The number of bytes needed to store the node.
+ */
static size_t
mdb_leaf_size(MDB_env *env, MDB_val *key, MDB_val *data)
{
return sz + sizeof(indx_t);
}
+/** Calculate the size of a branch node.
+ * The size should depend on the environment's page size but since
+ * we currently don't support spilling large keys onto overflow
+ * pages, it's simply the size of the #MDB_node header plus the
+ * size of the key. Sizes are always rounded up to an even number
+ * of bytes, to guarantee 2-byte alignment of the #MDB_node headers.
+ * @param[in] env The environment handle.
+ * @param[in] key The key for the node.
+ * @return The number of bytes needed to store the node.
+ */
static size_t
mdb_branch_size(MDB_env *env, MDB_val *key)
{
return sz + sizeof(indx_t);
}
+/** Add a node to the page pointed to by the cursor.
+ * @param[in] mc The cursor for this operation.
+ * @param[in] indx The index on the page where the new node should be added.
+ * @param[in] key The key for the new node.
+ * @param[in] data The data for the new node, if any.
+ * @param[in] pgno The page number, if adding a branch node.
+ * @param[in] flags Flags for the node.
+ * @return 0 on success, non-zero on failure. Possible errors are:
+ * <ul>
+ * <li>ENOMEM - failed to allocate overflow pages for the node.
+ * <li>ENOSPC - there is insufficient room in the page. This error
+ * should never happen since all callers already calculate the
+ * page's free space before calling this function.
+ * </ul>
+ */
static int
mdb_node_add(MDB_cursor *mc, indx_t indx,
MDB_val *key, MDB_val *data, pgno_t pgno, uint8_t flags)
return MDB_SUCCESS;
}
+/** Delete the specified node from a page.
+ * @param[in] mp The page to operate on.
+ * @param[in] indx The index of the node to delete.
+ * @param[in] ksize The size of a node. Only used if the page is
+ * part of a #MDB_DUPFIXED database.
+ */
static void
mdb_node_del(MDB_page *mp, indx_t indx, int ksize)
{
mp->mp_upper += sz;
}
+/** Initial setup of a sorted-dups cursor.
+ * Sorted duplicates are implemented as a sub-database for the given key.
+ * The duplicate data items are actually keys of the sub-database.
+ * Operations on the duplicate data items are performed using a sub-cursor
+ * initialized when the sub-database is first accessed. This function does
+ * the preliminary setup of the sub-cursor, filling in the fields that
+ * depend only on the parent DB.
+ * @param[in] mc The main cursor whose sorted-dups cursor is to be initialized.
+ */
static void
mdb_xcursor_init0(MDB_cursor *mc)
{
mx->mx_dbx.md_dirty = 0;
}
+/** Final setup of a sorted-dups cursor.
+ * Sets up the fields that depend on the data from the main cursor.
+ * @param[in] mc The main cursor whose sorted-dups cursor is to be initialized.
+ * @param[in] node The data containing the #MDB_db record for the
+ * sorted-dup database.
+ */
static void
mdb_xcursor_init1(MDB_cursor *mc, MDB_node *node)
{
mx->mx_dbx.md_cmp = mdb_cmp_long;
}
+/** Initialize a cursor for a given transaction and database. */
static void
mdb_cursor_init(MDB_cursor *mc, MDB_txn *txn, MDB_dbi dbi, MDB_xcursor *mx)
{
}
}
+/** Replace the key for a node with a new key.
+ * @param[in] mp The page containing the node to operate on.
+ * @param[in] indx The index of the node to operate on.
+ * @param[in] key The new key to use.
+ * @return 0 on success, non-zero on failure.
+ */
static int
mdb_update_key(MDB_page *mp, indx_t indx, MDB_val *key)
{
return MDB_SUCCESS;
}
-/* Move a node from csrc to cdst.
+/** Move a node from csrc to cdst.
*/
static int
mdb_node_move(MDB_cursor *csrc, MDB_cursor *cdst)
return MDB_SUCCESS;
}
+/** Merge one page into another.
+ * The nodes from the page pointed to by \b csrc will
+ * be copied to the page pointed to by \b cdst and then
+ * the \b csrc page will be freed.
+ * @param[in] csrc Cursor pointing to the source page.
+ * @param[in] cdst Cursor pointing to the destination page.
+ */
static int
mdb_page_merge(MDB_cursor *csrc, MDB_cursor *cdst)
{
return mdb_rebalance(csrc);
}
+/** Copy the contents of a cursor.
+ * @param[in] csrc The cursor to copy from.
+ * @param[out] cdst The cursor to copy to.
+ */
static void
mdb_cursor_copy(const MDB_cursor *csrc, MDB_cursor *cdst)
{
}
}
+/** Rebalance the tree after a delete operation.
+ * @param[in] mc Cursor pointing to the page where rebalancing
+ * should begin.
+ * @return 0 on success, non-zero on failure.
+ */
static int
mdb_rebalance(MDB_cursor *mc)
{
}
}
+/** Complete a delete operation started by #mdb_cursor_del(). */
static int
-mdb_del0(MDB_cursor *mc, MDB_node *leaf)
+mdb_cursor_del0(MDB_cursor *mc, MDB_node *leaf)
{
int rc;
return rc;
}
-/* Split page <mc->top>, and insert <key,(data|newpgno)> in either left or
- * right sibling, at index <mc->ki> (as if unsplit). Updates mc->top and
- * mc->ki with the actual values after split, ie if mc->top and mc->ki
- * refer to a node in the new right sibling page.
+/** Split a page and insert a new node.
+ * @param[in,out] mc Cursor pointing to the page and desired insertion index.
+ * The cursor will be updated to point to the actual page and index where
+ * the node got inserted after the split.
+ * @param[in] newkey The key for the newly inserted node.
+ * @param[in] newdata The data for the newly inserted node.
+ * @param[in] newpgno The page number, if the new node is a branch node.
+ * @return 0 on success, non-zero on failure.
*/
static int
mdb_page_split(MDB_cursor *mc, MDB_val *newkey, MDB_val *newdata, pgno_t newpgno)
return MDB_SUCCESS;
}
+/** Common code for #mdb_stat() and #mdb_env_stat().
+ * @param[in] env the environment to operate in.
+ * @param[in] db the #MDB_db record containing the stats to return.
+ * @param[out] arg the address of an #MDB_stat structure to receive the stats.
+ * @return 0, this function always succeeds.
+ */
static int
mdb_stat0(MDB_env *env, MDB_db *db, MDB_stat *arg)
{
return mdb_stat0(env, &env->me_metas[toggle]->mm_dbs[MAIN_DBI], arg);
}
+/** Set the default comparison functions for a database.
+ * Called immediately after a database is opened to set the defaults.
+ * The user can then override them with #mdb_set_compare() or
+ * #mdb_set_dupsort().
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ */
static void
mdb_default_cmp(MDB_txn *txn, MDB_dbi dbi)
{