mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-01-17 20:42:30 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
mariadb/storage/maria/ma_create.c
unknown 1a96259191 - WL#3239 "log CREATE TABLE in Maria" 
- WL#3240 "log DROP TABLE in Maria"
- similarly, log RENAME TABLE, REPAIR/OPTIMIZE TABLE, and
DELETE no_WHERE_clause (== the DELETE which just truncates the files)
- create_rename_lsn added to MARIA_SHARE's state
- all these operations (except DROP TABLE) also update the table's
create_rename_lsn, which is needed for the correctness of
Recovery (see function comment of _ma_repair_write_log_record()
in ma_check.c)
- write a COMMIT record when transaction commits.
- don't log REDOs/UNDOs if this is an internal temporary table
like inside ALTER TABLE (I expect this to be a big win). There was
already no logging for user-created "CREATE TEMPORARY" tables.
- don't fsync files/directories if the table is not transactional
- in translog_write_record(), autogenerate a 2-byte-id for the table
and log the "id->name" pair (LOGREC_FILE_ID); log
LOGREC_LONG_TRANSACTION_ID; automatically store
the table's 2-byte-id in any log record.
- preparations for Checkpoint: translog_get_horizon(); pausing Checkpoint
when some dirty pages are unknown; capturing trn->rec_lsn,
trn->first_undo_lsn for Checkpoint and log's low-water-mark computing.
- assertions, comments.


storage/maria/Makefile.am:
  more files to build
storage/maria/ha_maria.cc:
  - logging a REPAIR log record if REPAIR/OPTIMIZE was successful.
  - ha_maria::data_file_type does not have to be set in every info()
  call, just do it once in open().
  - if caller said that transactionality can be disabled (like if
  caller is ALTER TABLE) i.e. thd->transaction.on==FALSE, then we
  temporarily disable transactionality of the table in external_lock();
  that will ensure that no REDOs/UNDOs are logged for this possibly
  massive write operation (they are not needed, as if any write fails,
  the table will be dropped). We re-enable in external_lock(F_UNLCK),
  which in ALTER TABLE happens before the tmp table replaces the original
  one (which is good, as thus the final table will have a REDO RENAME
  and a correct create_rename_lsn).
  - when we commit we also have to write a log record, so
  trnman_commit_trn() calls become ma_commit() calls
  - at end of engine's initialization, we are potentially entering a
  multi-threaded dangerous world (clients are going to be accepted)
  and so some assertions of mutex-owning become enforceable, for that
  we set maria_multi_threaded=TRUE (see ma_control_file.c)
storage/maria/ha_maria.h:
  new member ha_maria::save_transactional (see also ha_maria.cc)
storage/maria/ma_blockrec.c:
  - fixing comments according to discussion with Monty
  - if a table is transactional but temporarily non-transactional
  (like in ALTER TABLE), we need to give a sensible LSN to the pages
  (and, if we give 0, pagecache asserts).
  - translog_write_record() now takes care of storing the share's
  2-byte-id in the log record
storage/maria/ma_blockrec.h:
  fixing comment according to discussion with Monty
storage/maria/ma_check.c:
  When REPAIR/OPTIMIZE modify the data/index file, if this is a
  transactional table, they must sync it; if they remove files or rename
  files, they must sync the directory, so that everything is durable.
  This is just applying to REPAIR/OPTIMIZE the logic already implemented
  in CREATE/DROP/RENAME a few months ago.
  Adding a function to write a LOGREC_REPAIR_TABLE at end of
  REPAIR/OPTIMIZE (called only by ha_maria, not by maria_chk), and
  to update the table's create_rename_lsn.
storage/maria/ma_close.c:
  fix for a future bug
storage/maria/ma_control_file.c:
  ensuring that if Maria is running in multi-threaded mode, anybody
  wanting to write to the control file and update
  last_checkpoint_lsn/last_logno owns the log's lock.
storage/maria/ma_control_file.h:
  see ma_control_file.c
storage/maria/ma_create.c:
  when creating a table:
  - sync it and its directory only if this is a transactional table
  and there is a log (no point in syncing in maria_chk)
  - decouple the two uses of linkname/linkname_ptr (for index file and
  for data file) into more variables, as we need to know all links
  until the moment we write the LOGREC_CREATE_TABLE.
  - set share.data_file_type early so that _ma_initialize_data_file()
  knows it (Monty's bugfix so that a table always has at least a bitmap
  page when it is created; so data-file is not 0 bytes anymore).
  - log a LOGREC_CREATE_TABLE; it contains the bytes which we have
  just written to the index file's header. Update table's
  create_rename_lsn.
  - syncing of kfile had been bugified in a previous merge, correcting
  - syncing of dfile is now needed as it's not empty anymore
  - in _ma_initialize_data_file(), use share's block_size and not the
  global one. This is a gratuitous change, both variables are equal,
  just that I find it more future-proof to use share-bound variable
  rather than global one.
storage/maria/ma_delete_all.c:
  log a LOGREC_DELETE_ALL record when doing ma_delete_all_rows();
  update create_rename_lsn then.
storage/maria/ma_delete_table.c:
  - logging LOGREC_DROP_TABLE; knowing if this is needed, requires
  knowing if the table is transactional, which requires opening the
  table.
  - we need to sync directories only if the table is transactional
storage/maria/ma_extra.c:
  questions
storage/maria/ma_init.c:
  when maria_end() is called, engine is not multithreaded
storage/maria/ma_loghandler.c:
  - translog_inited has to be visible to ma_create() (see how it is used
  in ma_create())
  - checkpoint record will be a single record, not three
  - no REDO for TRUNCATE (TRUNCATE calls ma_create() internally so will
  log a REDO_CREATE)
  - adding REDO for DELETE no_WHERE_clause (fast DELETE of all rows by
  truncating the files), REPAIR.
  - MY_WAIT_IF_FULL to wait&retry if a log write hits a full disk
  - in translog_write_record(), if MARIA_SHARE does not yet have a
  2-byte-id, generate one for it and log LOGREC_FILE_ID; automatically
  store this short id into log records.
  - in translog_write_record(), if transaction has not logged its
  long trid, log LOGREC_LONG_TRANSACTION_ID.
  - For Checkpoint, we need to know the current end-of-log: adding
  translog_get_horizon().
  - For Control File, adding an assertion that the thread owns the
  log's lock (control file is protected by this lock)
storage/maria/ma_loghandler.h:
  Changes in log records (see ma_loghandler.c).
  new prototypes, new functions.
storage/maria/ma_loghandler_lsn.h:
  adding a type LSN_WITH_FLAGS especially for TRN::first_undo_lsn,
  where the most significant byte is used for flags.
storage/maria/ma_open.c:
  storing the create_rename_lsn in the index file's header (in the
  state, precisely) and retrieving it from there.
storage/maria/ma_pagecache.c:
  - my set_if_bigger was wrong, correcting it
  - if the first_in_switch list is not empty, it means that
  changed_blocks misses some dirty pages, so Checkpoint cannot run and
  needs to wait. A variable missing_blocks_in_changed_list is added to
  tell that (should it be named missing_blocks_in_changed_blocks?)
  - pagecache_collect_changed_blocks_with_lsn() now also tells the
  minimum rec_lsn (needed for low-water mark computation).
storage/maria/ma_pagecache.h:
  see ma_pagecache.c
storage/maria/ma_panic.c:
  comment
storage/maria/ma_range.c:
  comment
storage/maria/ma_rename.c:
  - logging LOGREC_RENAME_TABLE; knowing if this is needed, requires
  knowing if the table is transactional, which requires opening the
  table.
  - update create_rename_lsn
  - we need to sync directories only if the table is transactional
storage/maria/ma_static.c:
  comment
storage/maria/ma_test_all.sh:
  - tip for Valgrind-ing ma_test_all
  - do "export maria_path=somepath" before calling ma_test_all,
  if you want to run ma_test_all out of storage/maria (useful
  to have parallel runs, like one normal and one Valgrind, they
  must not use the same tables so need to run in different directories)
storage/maria/maria_def.h:
  - state now contains, in memory and on disk, the create_rename_lsn
  - share now contains a 2-byte-id
storage/maria/trnman.c:
  preparations for Checkpoint: capture trn->rec_lsn, trn->first_undo_lsn;
  minimum first_undo_lsn needed to know log's low-water-mark
storage/maria/trnman.h:
  using most significant byte of first_undo_lsn to hold miscellaneous
  flags, for now TRANSACTION_LOGGED_LONG_ID.
  dummy_transaction_object is already declared in ma_static.c.
storage/maria/trnman_public.h:
  dummy_transaction_object was declared in all files including
  trnman_public.h, while in fact it's a single object.
  new prototype
storage/maria/unittest/ma_test_loghandler-t.c:
  update for new prototype
storage/maria/unittest/ma_test_loghandler_multigroup-t.c:
  update for new prototype
storage/maria/unittest/ma_test_loghandler_multithread-t.c:
  update for new prototype
storage/maria/unittest/ma_test_loghandler_pagecache-t.c:
  update for new prototype
storage/maria/ma_commit.c:
  function which wraps:
  - writing a LOGREC_COMMIT record (==commit on disk)
  - calling trnman_commit_trn() (=commit in memory)
storage/maria/ma_commit.h:
  new header file
.tree-is-private:
  this file is now needed to keep our tree private (don't push it
  to public trees). When 5.1 is merged into mysql-maria, we can abandon
  our maria-specific post-commit trigger; .tree_is_private will take
  care of keeping commit mails private. Don't push this file to public
  trees.
2007-06-22 14:49:37 +02:00

1159 lines
38 KiB
C
Raw Blame History

/* Copyright (C) 2006 MySQL AB & MySQL Finland AB & TCX DataKonsult AB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; version 2 of the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
/* Create a MARIA table */
#include "ma_ftdefs.h"
#include "ma_sp_defs.h"
#include <my_bit.h>
#include "ma_blockrec.h"
#include "trnman_public.h"
#if defined(MSDOS) || defined(__WIN__)
#ifdef __WIN__
#include <fcntl.h>
#else
#include <process.h> /* Prototype for getpid */
#endif
#endif
#include <m_ctype.h>
static int compare_columns(MARIA_COLUMNDEF **a, MARIA_COLUMNDEF **b);
/*
Old options is used when recreating database, from maria_chk
*/
int maria_create(const char *name, enum data_file_type datafile_type,
uint keys,MARIA_KEYDEF *keydefs,
uint columns, MARIA_COLUMNDEF *columndef,
uint uniques, MARIA_UNIQUEDEF *uniquedefs,
MARIA_CREATE_INFO *ci,uint flags)
{
register uint i,j;
File dfile,file;
int errpos,save_errno, create_mode= O_RDWR | O_TRUNC, res;
myf create_flag;
uint length,max_key_length,packed,pack_bytes,pointer,real_length_diff,
key_length,info_length,key_segs,options,min_key_length_skip,
base_pos,long_varchar_count,varchar_length,
unique_key_parts,fulltext_keys,offset, not_block_record_extra_length;
uint max_field_lengths, extra_header_size;
ulong reclength, real_reclength,min_pack_length;
char filename[FN_REFLEN], dlinkname[FN_REFLEN], *dlinkname_ptr= NULL,
klinkname[FN_REFLEN], *klinkname_ptr= NULL;
ulong pack_reclength;
ulonglong tot_length,max_rows, tmp;
enum en_fieldtype type;
enum data_file_type org_datafile_type= datafile_type;
MARIA_SHARE share;
MARIA_KEYDEF *keydef,tmp_keydef;
MARIA_UNIQUEDEF *uniquedef;
HA_KEYSEG *keyseg,tmp_keyseg;
MARIA_COLUMNDEF *column, *end_column;
ulong *rec_per_key_part;
my_off_t key_root[HA_MAX_POSSIBLE_KEY], kfile_size_before_extension;
MARIA_CREATE_INFO tmp_create_info;
my_bool tmp_table= FALSE; /* cache for presence of HA_OPTION_TMP_TABLE */
my_bool forced_packed;
myf sync_dir= 0;
uchar *log_data= NULL;
DBUG_ENTER("maria_create");
DBUG_PRINT("enter", ("keys: %u columns: %u uniques: %u flags: %u",
keys, columns, uniques, flags));
DBUG_ASSERT(maria_block_size && maria_block_size % IO_SIZE == 0);
LINT_INIT(dfile);
LINT_INIT(file);
if (!ci)
{
bzero((char*) &tmp_create_info,sizeof(tmp_create_info));
ci=&tmp_create_info;
}
if (keys + uniques > MARIA_MAX_KEY || columns == 0)
{
DBUG_RETURN(my_errno=HA_WRONG_CREATE_OPTION);
}
errpos=0;
options=0;
bzero((byte*) &share,sizeof(share));
if (flags & HA_DONT_TOUCH_DATA)
{
org_datafile_type= ci->org_data_file_type;
if (!(ci->old_options & HA_OPTION_TEMP_COMPRESS_RECORD))
options=ci->old_options &
(HA_OPTION_COMPRESS_RECORD | HA_OPTION_PACK_RECORD |
HA_OPTION_READ_ONLY_DATA | HA_OPTION_CHECKSUM |
HA_OPTION_TMP_TABLE | HA_OPTION_DELAY_KEY_WRITE);
else
{
/* Uncompressing rows */
options=ci->old_options &
(HA_OPTION_CHECKSUM | HA_OPTION_TMP_TABLE | HA_OPTION_DELAY_KEY_WRITE);
}
}
if (ci->reloc_rows > ci->max_rows)
ci->reloc_rows=ci->max_rows; /* Check if wrong parameter */
if (!(rec_per_key_part=
(ulong*) my_malloc((keys + uniques)*HA_MAX_KEY_SEG*sizeof(long),
MYF(MY_WME | MY_ZEROFILL))))
DBUG_RETURN(my_errno);
/* Start by checking fields and field-types used */
varchar_length=long_varchar_count=packed= not_block_record_extra_length=
pack_reclength= max_field_lengths= 0;
reclength= min_pack_length= ci->null_bytes;
forced_packed= 0;
for (column= columndef, end_column= column + columns ;
column != end_column ;
column++)
{
/* Fill in not used struct parts */
column->offset= reclength;
column->empty_pos= 0;
column->empty_bit= 0;
column->fill_length= column->length;
reclength+= column->length;
type= column->type;
if (type == FIELD_SKIP_PRESPACE && datafile_type == BLOCK_RECORD)
type= FIELD_NORMAL; /* SKIP_PRESPACE not supported */
if (type != FIELD_NORMAL && type != FIELD_CHECK)
{
column->empty_pos= packed/8;
column->empty_bit= (1 << (packed & 7));
if (type == FIELD_BLOB)
{
forced_packed= 1;
packed++;
share.base.blobs++;
if (pack_reclength != INT_MAX32)
{
if (column->length == 4+portable_sizeof_char_ptr)
pack_reclength= INT_MAX32;
else
{
/* Add max possible blob length */
pack_reclength+= (1 << ((column->length-
portable_sizeof_char_ptr)*8));
}
}
max_field_lengths+= (column->length - portable_sizeof_char_ptr);
}
else if (type == FIELD_SKIP_PRESPACE ||
type == FIELD_SKIP_ENDSPACE)
{
forced_packed= 1;
max_field_lengths+= column->length > 255 ? 2 : 1;
not_block_record_extra_length++;
packed++;
}
else if (type == FIELD_VARCHAR)
{
varchar_length+= column->length-1; /* Used for min_pack_length */
pack_reclength++;
not_block_record_extra_length++;
max_field_lengths++;
packed++;
column->fill_length= 1;
/* We must test for 257 as length includes pack-length */
if (test(column->length >= 257))
{
long_varchar_count++;
max_field_lengths++;
column->fill_length= 2;
}
}
else if (type == FIELD_SKIP_ZERO)
packed++;
else
{
if (!column->null_bit)
min_pack_length+= column->length;
else
not_block_record_extra_length+= column->length;
column->empty_pos= 0;
column->empty_bit= 0;
}
}
else /* FIELD_NORMAL */
{
if (!column->null_bit)
{
min_pack_length+= column->length;
share.base.fixed_not_null_fields++;
share.base.fixed_not_null_fields_length+= column->length;
}
else
not_block_record_extra_length+= column->length;
}
}
if (datafile_type == STATIC_RECORD && forced_packed)
{
/* Can't use fixed length records, revert to block records */
datafile_type= BLOCK_RECORD;
}
if (datafile_type == DYNAMIC_RECORD)
options|= HA_OPTION_PACK_RECORD; /* Must use packed records */
if (datafile_type == STATIC_RECORD)
{
/* We can't use checksum with static length rows */
flags&= ~HA_CREATE_CHECKSUM;
options&= ~HA_OPTION_CHECKSUM;
min_pack_length+= varchar_length;
packed= 0;
}
if (datafile_type != BLOCK_RECORD)
min_pack_length+= not_block_record_extra_length;
if ((packed & 7) == 1)
{
/*
Not optimal packing, try to remove a 1 byte length zero-field as
this will get same record length, but smaller pack overhead
*/
while (column != columndef)
{
column--;
if (column->type == (int) FIELD_SKIP_ZERO && column->length == 1)
{
column->type=(int) FIELD_NORMAL;
column->empty_pos= 0;
column->empty_bit= 0;
packed--;
min_pack_length++;
break;
}
}
}
if (flags & HA_CREATE_TMP_TABLE)
{
options|= HA_OPTION_TMP_TABLE;
tmp_table= TRUE;
create_mode|= O_EXCL | O_NOFOLLOW;
/* "CREATE TEMPORARY" tables are not crash-safe (dropped at restart) */
ci->transactional= FALSE;
}
share.base.null_bytes= ci->null_bytes;
share.base.original_null_bytes= ci->null_bytes;
share.base.transactional= ci->transactional;
share.base.max_field_lengths= max_field_lengths;
share.base.field_offsets= 0; /* for future */
if (pack_reclength != INT_MAX32)
pack_reclength+= max_field_lengths + long_varchar_count;
if (flags & HA_CREATE_CHECKSUM || (options & HA_OPTION_CHECKSUM))
{
options|= HA_OPTION_CHECKSUM;
min_pack_length++;
pack_reclength++;
}
if (flags & HA_CREATE_DELAY_KEY_WRITE)
options|= HA_OPTION_DELAY_KEY_WRITE;
if (flags & HA_CREATE_RELIES_ON_SQL_LAYER)
options|= HA_OPTION_RELIES_ON_SQL_LAYER;
pack_bytes= (packed + 7) / 8;
if (pack_reclength != INT_MAX32)
pack_reclength+= reclength+pack_bytes +
test(test_all_bits(options, HA_OPTION_CHECKSUM | HA_PACK_RECORD));
min_pack_length+= pack_bytes;
/* Calculate min possible row length for rows-in-block */
extra_header_size= MAX_FIXED_HEADER_SIZE;
if (ci->transactional)
extra_header_size= TRANS_MAX_FIXED_HEADER_SIZE;
share.base.min_row_length= (extra_header_size + share.base.null_bytes +
pack_bytes);
if (!ci->data_file_length && ci->max_rows)
{
if (pack_reclength == INT_MAX32 ||
(~(ulonglong) 0)/ci->max_rows < (ulonglong) pack_reclength)
ci->data_file_length= ~(ulonglong) 0;
else
ci->data_file_length=(ulonglong) ci->max_rows*pack_reclength;
}
else if (!ci->max_rows)
{
if (datafile_type == BLOCK_RECORD)
{
uint rows_per_page= ((maria_block_size - PAGE_OVERHEAD_SIZE) /
(min_pack_length + extra_header_size +
DIR_ENTRY_SIZE));
ulonglong data_file_length= ci->data_file_length;
if (!data_file_length)
data_file_length= ((((ulonglong) 1 << ((BLOCK_RECORD_POINTER_SIZE-1) *
8)) -1));
if (rows_per_page > 0)
{
set_if_smaller(rows_per_page, MAX_ROWS_PER_PAGE);
ci->max_rows= data_file_length / maria_block_size * rows_per_page;
}
else
ci->max_rows= data_file_length / (min_pack_length +
extra_header_size +
DIR_ENTRY_SIZE);
}
else
ci->max_rows=(ha_rows) (ci->data_file_length/(min_pack_length +
((options &
HA_OPTION_PACK_RECORD) ?
3 : 0)));
}
max_rows= (ulonglong) ci->max_rows;
if (datafile_type == BLOCK_RECORD)
{
/* The + 1 is for record position withing page */
pointer= maria_get_pointer_length((ci->data_file_length /
maria_block_size), 3) + 1;
set_if_smaller(pointer, BLOCK_RECORD_POINTER_SIZE);
if (!max_rows)
max_rows= (((((ulonglong) 1 << ((pointer-1)*8)) -1) * maria_block_size) /
min_pack_length);
}
else
{
if (datafile_type != STATIC_RECORD)
pointer= maria_get_pointer_length(ci->data_file_length,
maria_data_pointer_size);
else
pointer= maria_get_pointer_length(ci->max_rows, maria_data_pointer_size);
if (!max_rows)
max_rows= ((((ulonglong) 1 << (pointer*8)) -1) / min_pack_length);
}
real_reclength=reclength;
if (datafile_type == STATIC_RECORD)
{
if (reclength <= pointer)
reclength=pointer+1; /* reserve place for delete link */
}
else
reclength+= long_varchar_count; /* We need space for varchar! */
max_key_length=0; tot_length=0 ; key_segs=0;
fulltext_keys=0;
share.state.rec_per_key_part=rec_per_key_part;
share.state.key_root=key_root;
share.state.key_del= HA_OFFSET_ERROR;
if (uniques)
max_key_length= MARIA_UNIQUE_HASH_LENGTH + pointer;
for (i=0, keydef=keydefs ; i < keys ; i++ , keydef++)
{
share.state.key_root[i]= HA_OFFSET_ERROR;
min_key_length_skip=length=real_length_diff=0;
key_length=pointer;
if (keydef->flag & HA_SPATIAL)
{
#ifdef HAVE_SPATIAL
/* BAR TODO to support 3D and more dimensions in the future */
uint sp_segs=SPDIMS*2;
keydef->flag=HA_SPATIAL;
if (flags & HA_DONT_TOUCH_DATA)
{
/*
Called by maria_chk - i.e. table structure was taken from
MYI file and SPATIAL key *does have* additional sp_segs keysegs.
keydef->seg here points right at the GEOMETRY segment,
so we only need to decrease keydef->keysegs.
(see maria_recreate_table() in _ma_check.c)
*/
keydef->keysegs-=sp_segs-1;
}
for (j=0, keyseg=keydef->seg ; (int) j < keydef->keysegs ;
j++, keyseg++)
{
if (keyseg->type != HA_KEYTYPE_BINARY &&
keyseg->type != HA_KEYTYPE_VARBINARY1 &&
keyseg->type != HA_KEYTYPE_VARBINARY2)
{
my_errno=HA_WRONG_CREATE_OPTION;
goto err_no_lock;
}
}
keydef->keysegs+=sp_segs;
key_length+=SPLEN*sp_segs;
length++; /* At least one length byte */
min_key_length_skip+=SPLEN*2*SPDIMS;
#else
my_errno= HA_ERR_UNSUPPORTED;
goto err_no_lock;
#endif /*HAVE_SPATIAL*/
}
else if (keydef->flag & HA_FULLTEXT)
{
keydef->flag=HA_FULLTEXT | HA_PACK_KEY | HA_VAR_LENGTH_KEY;
options|=HA_OPTION_PACK_KEYS; /* Using packed keys */
for (j=0, keyseg=keydef->seg ; (int) j < keydef->keysegs ;
j++, keyseg++)
{
if (keyseg->type != HA_KEYTYPE_TEXT &&
keyseg->type != HA_KEYTYPE_VARTEXT1 &&
keyseg->type != HA_KEYTYPE_VARTEXT2)
{
my_errno=HA_WRONG_CREATE_OPTION;
goto err_no_lock;
}
if (!(keyseg->flag & HA_BLOB_PART) &&
(keyseg->type == HA_KEYTYPE_VARTEXT1 ||
keyseg->type == HA_KEYTYPE_VARTEXT2))
{
/* Make a flag that this is a VARCHAR */
keyseg->flag|= HA_VAR_LENGTH_PART;
/* Store in bit_start number of bytes used to pack the length */
keyseg->bit_start= ((keyseg->type == HA_KEYTYPE_VARTEXT1)?
1 : 2);
}
}
fulltext_keys++;
key_length+= HA_FT_MAXBYTELEN+HA_FT_WLEN;
length++; /* At least one length byte */
min_key_length_skip+=HA_FT_MAXBYTELEN;
real_length_diff=HA_FT_MAXBYTELEN-FT_MAX_WORD_LEN_FOR_SORT;
}
else
{
/* Test if prefix compression */
if (keydef->flag & HA_PACK_KEY)
{
/* Can't use space_compression on number keys */
if ((keydef->seg[0].flag & HA_SPACE_PACK) &&
keydef->seg[0].type == (int) HA_KEYTYPE_NUM)
keydef->seg[0].flag&= ~HA_SPACE_PACK;
/* Only use HA_PACK_KEY when first segment is a variable length key */
if (!(keydef->seg[0].flag & (HA_SPACE_PACK | HA_BLOB_PART |
HA_VAR_LENGTH_PART)))
{
/* pack relative to previous key */
keydef->flag&= ~HA_PACK_KEY;
keydef->flag|= HA_BINARY_PACK_KEY | HA_VAR_LENGTH_KEY;
}
else
{
keydef->seg[0].flag|=HA_PACK_KEY; /* for easyer intern test */
keydef->flag|=HA_VAR_LENGTH_KEY;
options|=HA_OPTION_PACK_KEYS; /* Using packed keys */
}
}
if (keydef->flag & HA_BINARY_PACK_KEY)
options|=HA_OPTION_PACK_KEYS; /* Using packed keys */
if (keydef->flag & HA_AUTO_KEY && ci->with_auto_increment)
share.base.auto_key=i+1;
for (j=0, keyseg=keydef->seg ; j < keydef->keysegs ; j++, keyseg++)
{
/* numbers are stored with high by first to make compression easier */
switch (keyseg->type) {
case HA_KEYTYPE_SHORT_INT:
case HA_KEYTYPE_LONG_INT:
case HA_KEYTYPE_FLOAT:
case HA_KEYTYPE_DOUBLE:
case HA_KEYTYPE_USHORT_INT:
case HA_KEYTYPE_ULONG_INT:
case HA_KEYTYPE_LONGLONG:
case HA_KEYTYPE_ULONGLONG:
case HA_KEYTYPE_INT24:
case HA_KEYTYPE_UINT24:
case HA_KEYTYPE_INT8:
keyseg->flag|= HA_SWAP_KEY;
break;
case HA_KEYTYPE_VARTEXT1:
case HA_KEYTYPE_VARTEXT2:
case HA_KEYTYPE_VARBINARY1:
case HA_KEYTYPE_VARBINARY2:
if (!(keyseg->flag & HA_BLOB_PART))
{
/* Make a flag that this is a VARCHAR */
keyseg->flag|= HA_VAR_LENGTH_PART;
/* Store in bit_start number of bytes used to pack the length */
keyseg->bit_start= ((keyseg->type == HA_KEYTYPE_VARTEXT1 ||
keyseg->type == HA_KEYTYPE_VARBINARY1) ?
1 : 2);
}
break;
default:
break;
}
if (keyseg->flag & HA_SPACE_PACK)
{
DBUG_ASSERT(!(keyseg->flag & HA_VAR_LENGTH_PART));
keydef->flag |= HA_SPACE_PACK_USED | HA_VAR_LENGTH_KEY;
options|=HA_OPTION_PACK_KEYS; /* Using packed keys */
length++; /* At least one length byte */
min_key_length_skip+=keyseg->length;
if (keyseg->length >= 255)
{ /* prefix may be 3 bytes */
min_key_length_skip+=2;
length+=2;
}
}
if (keyseg->flag & (HA_VAR_LENGTH_PART | HA_BLOB_PART))
{
DBUG_ASSERT(!test_all_bits(keyseg->flag,
(HA_VAR_LENGTH_PART | HA_BLOB_PART)));
keydef->flag|=HA_VAR_LENGTH_KEY;
length++; /* At least one length byte */
options|=HA_OPTION_PACK_KEYS; /* Using packed keys */
min_key_length_skip+=keyseg->length;
if (keyseg->length >= 255)
{ /* prefix may be 3 bytes */
min_key_length_skip+=2;
length+=2;
}
}
key_length+= keyseg->length;
if (keyseg->null_bit)
{
key_length++;
options|=HA_OPTION_PACK_KEYS;
keyseg->flag|=HA_NULL_PART;
keydef->flag|=HA_VAR_LENGTH_KEY | HA_NULL_PART_KEY;
}
}
} /* if HA_FULLTEXT */
key_segs+=keydef->keysegs;
if (keydef->keysegs > HA_MAX_KEY_SEG)
{
my_errno=HA_WRONG_CREATE_OPTION;
goto err_no_lock;
}
/*
key_segs may be 0 in the case when we only want to be able to
add on row into the table. This can happen with some DISTINCT queries
in MySQL
*/
if ((keydef->flag & (HA_NOSAME | HA_NULL_PART_KEY)) == HA_NOSAME &&
key_segs)
share.state.rec_per_key_part[key_segs-1]=1L;
length+=key_length;
/*
A key can't be longer than than half a index block (as we have
to be able to put at least 2 keys on an index block for the key
algorithms to work).
*/
if (length > maria_max_key_length())
{
my_errno=HA_WRONG_CREATE_OPTION;
goto err_no_lock;
}
keydef->block_length= maria_block_size;
keydef->keylength= (uint16) key_length;
keydef->minlength= (uint16) (length-min_key_length_skip);
keydef->maxlength= (uint16) length;
if (length > max_key_length)
max_key_length= length;
tot_length+= ((max_rows/(ulong) (((uint) maria_block_size-5)/
(length*2))) *
maria_block_size);
}
unique_key_parts=0;
offset=reclength-uniques*MARIA_UNIQUE_HASH_LENGTH;
for (i=0, uniquedef=uniquedefs ; i < uniques ; i++ , uniquedef++)
{
uniquedef->key=keys+i;
unique_key_parts+=uniquedef->keysegs;
share.state.key_root[keys+i]= HA_OFFSET_ERROR;
tot_length+= (max_rows/(ulong) (((uint) maria_block_size-5)/
((MARIA_UNIQUE_HASH_LENGTH + pointer)*2)))*
(ulong) maria_block_size;
}
keys+=uniques; /* Each unique has 1 key */
key_segs+=uniques; /* Each unique has 1 key seg */
base_pos=(MARIA_STATE_INFO_SIZE + keys * MARIA_STATE_KEY_SIZE +
key_segs * MARIA_STATE_KEYSEG_SIZE);
info_length= base_pos+(uint) (MARIA_BASE_INFO_SIZE+
keys * MARIA_KEYDEF_SIZE+
uniques * MARIA_UNIQUEDEF_SIZE +
(key_segs + unique_key_parts)*HA_KEYSEG_SIZE+
columns*MARIA_COLUMNDEF_SIZE);
DBUG_PRINT("info", ("info_length: %u", info_length));
/* There are only 16 bits for the total header length. */
if (info_length > 65535)
{
my_printf_error(0, "Maria table '%s' has too many columns and/or "
"indexes and/or unique constraints.",
MYF(0), name + dirname_length(name));
my_errno= HA_WRONG_CREATE_OPTION;
goto err_no_lock;
}
bmove(share.state.header.file_version,(byte*) maria_file_magic,4);
ci->old_options=options| (ci->old_options & HA_OPTION_TEMP_COMPRESS_RECORD ?
HA_OPTION_COMPRESS_RECORD |
HA_OPTION_TEMP_COMPRESS_RECORD: 0);
mi_int2store(share.state.header.options,ci->old_options);
mi_int2store(share.state.header.header_length,info_length);
mi_int2store(share.state.header.state_info_length,MARIA_STATE_INFO_SIZE);
mi_int2store(share.state.header.base_info_length,MARIA_BASE_INFO_SIZE);
mi_int2store(share.state.header.base_pos,base_pos);
share.state.header.data_file_type= datafile_type;
share.state.header.org_data_file_type= org_datafile_type;
share.state.header.language= (ci->language ?
ci->language : default_charset_info->number);
share.state.dellink = HA_OFFSET_ERROR;
share.state.first_bitmap_with_space= 0;
share.state.create_rename_lsn= 0;
share.state.process= (ulong) getpid();
share.state.unique= (ulong) 0;
share.state.update_count=(ulong) 0;
share.state.version= (ulong) time((time_t*) 0);
share.state.sortkey= (ushort) ~0;
share.state.auto_increment=ci->auto_increment;
share.options=options;
share.base.rec_reflength=pointer;
share.base.block_size= maria_block_size;
/* Get estimate for index file length (this may be wrong for FT keys) */
tmp= (tot_length + maria_block_size * keys *
MARIA_INDEX_BLOCK_MARGIN) / maria_block_size;
/*
use maximum of key_file_length we calculated and key_file_length value we
got from MYI file header (see also mariapack.c:save_state)
*/
share.base.key_reflength=
maria_get_pointer_length(max(ci->key_file_length,tmp),3);
share.base.keys= share.state.header.keys= keys;
share.state.header.uniques= uniques;
share.state.header.fulltext_keys= fulltext_keys;
mi_int2store(share.state.header.key_parts,key_segs);
mi_int2store(share.state.header.unique_key_parts,unique_key_parts);
maria_set_all_keys_active(share.state.key_map, keys);
share.base.keystart = share.state.state.key_file_length=
MY_ALIGN(info_length, maria_block_size);
share.base.max_key_block_length= maria_block_size;
share.base.max_key_length=ALIGN_SIZE(max_key_length+4);
share.base.records=ci->max_rows;
share.base.reloc= ci->reloc_rows;
share.base.reclength=real_reclength;
share.base.pack_reclength=reclength+ test(options & HA_OPTION_CHECKSUM);
share.base.max_pack_length=pack_reclength;
share.base.min_pack_length=min_pack_length;
share.base.pack_bytes= pack_bytes;
share.base.fields= columns;
share.base.pack_fields= packed;
#ifdef USE_RAID
share.base.raid_type=ci->raid_type;
share.base.raid_chunks=ci->raid_chunks;
share.base.raid_chunksize=ci->raid_chunksize;
#endif
/* max_data_file_length and max_key_file_length are recalculated on open */
if (tmp_table)
share.base.max_data_file_length= (my_off_t) ci->data_file_length;
else if (ci->transactional && translog_inited)
{
/*
we have checked translog_inited above, because maria_chk may call us
(via maria_recreate_table()) and it does not have a log.
*/
sync_dir= MY_SYNC_DIR;
}
if (datafile_type == BLOCK_RECORD)
share.base.min_block_length= share.base.min_row_length;
else
{
share.base.min_block_length=
(share.base.pack_reclength+3 < MARIA_EXTEND_BLOCK_LENGTH &&
! share.base.blobs) ?
max(share.base.pack_reclength,MARIA_MIN_BLOCK_LENGTH) :
MARIA_EXTEND_BLOCK_LENGTH;
}
if (! (flags & HA_DONT_TOUCH_DATA))
share.state.create_time= (long) time((time_t*) 0);
pthread_mutex_lock(&THR_LOCK_maria);
if (ci->index_file_name)
{
char *iext= strrchr(ci->index_file_name, '.');
int have_iext= iext && !strcmp(iext, MARIA_NAME_IEXT);
if (tmp_table)
{
char *path;
/* chop off the table name, tempory tables use generated name */
if ((path= strrchr(ci->index_file_name, FN_LIBCHAR)))
*path= '\0';
fn_format(filename, name, ci->index_file_name, MARIA_NAME_IEXT,
MY_REPLACE_DIR | MY_UNPACK_FILENAME | MY_APPEND_EXT);
}
else
{
fn_format(filename, ci->index_file_name, "", MARIA_NAME_IEXT,
MY_UNPACK_FILENAME | (have_iext ? MY_REPLACE_EXT :
MY_APPEND_EXT));
}
fn_format(klinkname, name, "", MARIA_NAME_IEXT,
MY_UNPACK_FILENAME|MY_APPEND_EXT);
klinkname_ptr= klinkname;
/*
Don't create the table if the link or file exists to ensure that one
doesn't accidently destroy another table.
Don't sync dir now if the data file has the same path.
*/
create_flag=
(ci->data_file_name &&
!strcmp(ci->index_file_name, ci->data_file_name)) ? 0 : sync_dir;
}
else
{
fn_format(filename, name, "", MARIA_NAME_IEXT,
(MY_UNPACK_FILENAME |
(flags & HA_DONT_TOUCH_DATA) ? MY_RETURN_REAL_PATH : 0) |
MY_APPEND_EXT);
/*
Replace the current file.
Don't sync dir now if the data file has the same path.
*/
create_flag= MY_DELETE_OLD | (!ci->data_file_name ? 0 : sync_dir);
}
/*
If a MRG_MARIA table is in use, the mapped MARIA tables are open,
but no entry is made in the table cache for them.
A TRUNCATE command checks for the table in the cache only and could
be fooled to believe, the table is not open.
Pull the emergency brake in this situation. (Bug #8306)
*/
if (_ma_test_if_reopen(filename))
{
my_printf_error(0, "MARIA table '%s' is in use "
"(most likely by a MERGE table). Try FLUSH TABLES.",
MYF(0), name + dirname_length(name));
goto err;
}
if ((file= my_create_with_symlink(klinkname_ptr, filename, 0, create_mode,
MYF(MY_WME|create_flag))) < 0)
goto err;
errpos=1;
if (!(flags & HA_DONT_TOUCH_DATA))
{
if (ci->data_file_name)
{
char *dext= strrchr(ci->data_file_name, '.');
int have_dext= dext && !strcmp(dext, MARIA_NAME_DEXT);
if (tmp_table)
{
char *path;
/* chop off the table name, tempory tables use generated name */
if ((path= strrchr(ci->data_file_name, FN_LIBCHAR)))
*path= '\0';
fn_format(filename, name, ci->data_file_name, MARIA_NAME_DEXT,
MY_REPLACE_DIR | MY_UNPACK_FILENAME | MY_APPEND_EXT);
}
else
{
fn_format(filename, ci->data_file_name, "", MARIA_NAME_DEXT,
MY_UNPACK_FILENAME |
(have_dext ? MY_REPLACE_EXT : MY_APPEND_EXT));
}
fn_format(dlinkname, name, "",MARIA_NAME_DEXT,
MY_UNPACK_FILENAME | MY_APPEND_EXT);
dlinkname_ptr= dlinkname;
create_flag=0;
}
else
{
fn_format(filename,name,"", MARIA_NAME_DEXT,
MY_UNPACK_FILENAME | MY_APPEND_EXT);
create_flag=MY_DELETE_OLD;
}
if ((dfile=
my_create_with_symlink(dlinkname_ptr, filename, 0, create_mode,
MYF(MY_WME | create_flag | sync_dir))) < 0)
goto err;
errpos=3;
share.data_file_type= datafile_type;
if (_ma_initialize_data_file(dfile, &share))
goto err;
}
DBUG_PRINT("info", ("write state info and base info"));
if (_ma_state_info_write(file, &share.state, 2) ||
_ma_base_info_write(file, &share.base))
goto err;
DBUG_PRINT("info", ("base_pos: %d base_info_size: %d",
base_pos, MARIA_BASE_INFO_SIZE));
DBUG_ASSERT(my_tell(file,MYF(0)) == base_pos+ MARIA_BASE_INFO_SIZE);
/* Write key and keyseg definitions */
DBUG_PRINT("info", ("write key and keyseg definitions"));
for (i=0 ; i < share.base.keys - uniques; i++)
{
uint sp_segs=(keydefs[i].flag & HA_SPATIAL) ? 2*SPDIMS : 0;
if (_ma_keydef_write(file, &keydefs[i]))
goto err;
for (j=0 ; j < keydefs[i].keysegs-sp_segs ; j++)
if (_ma_keyseg_write(file, &keydefs[i].seg[j]))
goto err;
#ifdef HAVE_SPATIAL
for (j=0 ; j < sp_segs ; j++)
{
HA_KEYSEG sseg;
sseg.type=SPTYPE;
sseg.language= 7; /* Binary */
sseg.null_bit=0;
sseg.bit_start=0;
sseg.bit_end=0;
sseg.bit_length= 0;
sseg.bit_pos= 0;
sseg.length=SPLEN;
sseg.null_pos=0;
sseg.start=j*SPLEN;
sseg.flag= HA_SWAP_KEY;
if (_ma_keyseg_write(file, &sseg))
goto err;
}
#endif
}
/* Create extra keys for unique definitions */
offset=reclength-uniques*MARIA_UNIQUE_HASH_LENGTH;
bzero((char*) &tmp_keydef,sizeof(tmp_keydef));
bzero((char*) &tmp_keyseg,sizeof(tmp_keyseg));
for (i=0; i < uniques ; i++)
{
tmp_keydef.keysegs=1;
tmp_keydef.flag= HA_UNIQUE_CHECK;
tmp_keydef.block_length= (uint16) maria_block_size;
tmp_keydef.keylength= MARIA_UNIQUE_HASH_LENGTH + pointer;
tmp_keydef.minlength=tmp_keydef.maxlength=tmp_keydef.keylength;
tmp_keyseg.type= MARIA_UNIQUE_HASH_TYPE;
tmp_keyseg.length= MARIA_UNIQUE_HASH_LENGTH;
tmp_keyseg.start= offset;
offset+= MARIA_UNIQUE_HASH_LENGTH;
if (_ma_keydef_write(file,&tmp_keydef) ||
_ma_keyseg_write(file,(&tmp_keyseg)))
goto err;
}
/* Save unique definition */
DBUG_PRINT("info", ("write unique definitions"));
for (i=0 ; i < share.state.header.uniques ; i++)
{
HA_KEYSEG *keyseg_end;
keyseg= uniquedefs[i].seg;
if (_ma_uniquedef_write(file, &uniquedefs[i]))
goto err;
for (keyseg= uniquedefs[i].seg, keyseg_end= keyseg+ uniquedefs[i].keysegs;
keyseg < keyseg_end;
keyseg++)
{
switch (keyseg->type) {
case HA_KEYTYPE_VARTEXT1:
case HA_KEYTYPE_VARTEXT2:
case HA_KEYTYPE_VARBINARY1:
case HA_KEYTYPE_VARBINARY2:
if (!(keyseg->flag & HA_BLOB_PART))
{
keyseg->flag|= HA_VAR_LENGTH_PART;
keyseg->bit_start= ((keyseg->type == HA_KEYTYPE_VARTEXT1 ||
keyseg->type == HA_KEYTYPE_VARBINARY1) ?
1 : 2);
}
break;
default:
DBUG_ASSERT((keyseg->flag & HA_VAR_LENGTH_PART) == 0);
break;
}
if (_ma_keyseg_write(file, keyseg))
goto err;
}
}
DBUG_PRINT("info", ("write field definitions"));
if (datafile_type == BLOCK_RECORD)
{
/* Store columns in a more efficent order */
MARIA_COLUMNDEF **col_order, **pos;
if (!(col_order= (MARIA_COLUMNDEF**) my_malloc(share.base.fields *
sizeof(MARIA_COLUMNDEF*),
MYF(MY_WME))))
goto err;
for (column= columndef, pos= col_order ;
column != end_column ;
column++, pos++)
*pos= column;
qsort(col_order, share.base.fields, sizeof(*col_order),
(qsort_cmp) compare_columns);
for (i=0 ; i < share.base.fields ; i++)
{
if (_ma_columndef_write(file, col_order[i]))
{
my_free((gptr) col_order, MYF(0));
goto err;
}
}
my_free((gptr) col_order, MYF(0));
}
else
{
for (i=0 ; i < share.base.fields ; i++)
if (_ma_columndef_write(file, &columndef[i]))
goto err;
}
if ((kfile_size_before_extension= my_tell(file,MYF(0))) == MY_FILEPOS_ERROR)
goto err;
#ifndef DBUG_OFF
if (kfile_size_before_extension != info_length)
DBUG_PRINT("warning",("info_length: %u != used_length: %u",
info_length, (uint)kfile_size_before_extension));
#endif
if (sync_dir)
{
/*
we log the first bytes and then the size to which we extend; this is
not log 1 KB of mostly zeroes if this is a small table.
*/
char empty_string[]= "";
LEX_STRING log_array[TRANSLOG_INTERNAL_PARTS + 3];
uint total_rec_length= 0;
uint i;
log_array[TRANSLOG_INTERNAL_PARTS + 0].length= 1 + 2 +
kfile_size_before_extension;
/* we are needing maybe 64 kB, so don't use the stack */
log_data= my_malloc(log_array[TRANSLOG_INTERNAL_PARTS + 0].length, MYF(0));
if ((log_data == NULL) ||
my_pread(file, 1 + 2 + log_data, kfile_size_before_extension,
0, MYF(MY_NABP)))
goto err_no_lock;
/*
remember if the data file was created or not, to know if Recovery can
do it or not, in the future
*/
log_data[0]= test(flags & HA_DONT_TOUCH_DATA);
int2store(log_data + 1, kfile_size_before_extension);
log_array[TRANSLOG_INTERNAL_PARTS + 0].str= log_data;
/* symlink description is also needed for re-creation by Recovery: */
log_array[TRANSLOG_INTERNAL_PARTS + 1].str=
dlinkname_ptr ? dlinkname : empty_string;
log_array[TRANSLOG_INTERNAL_PARTS + 1].length=
strlen(log_array[TRANSLOG_INTERNAL_PARTS + 1].str);
log_array[TRANSLOG_INTERNAL_PARTS + 2].str=
klinkname_ptr ? klinkname : empty_string;
log_array[TRANSLOG_INTERNAL_PARTS + 2].length=
strlen(log_array[TRANSLOG_INTERNAL_PARTS + 2].str);
for (i= TRANSLOG_INTERNAL_PARTS;
i < (sizeof(log_array)/sizeof(log_array[0])); i++)
total_rec_length+= log_array[i].length;
/*
For this record to be of any use for Recovery, we need the upper
MySQL layer to be crash-safe, which it is not now (that would require
work using the ddl_log of sql/sql_table.cc); when it is, we should
reconsider the moment of writing this log record (before or after op,
under THR_LOCK_maria or not...), how to use it in Recovery, and force
the log. For now this record is just informative.
Note that in case of TRUNCATE TABLE we also come here.
When in CREATE/TRUNCATE (or DROP or RENAME or REPAIR) we have not called
external_lock(), so have no TRN. It does not matter, as all these
operations are non-transactional and sync their files.
*/
if (unlikely(translog_write_record(&share.state.create_rename_lsn,
LOGREC_REDO_CREATE_TABLE,
&dummy_transaction_object, NULL,
total_rec_length,
sizeof(log_array)/sizeof(log_array[0]),
log_array, NULL)))
goto err_no_lock;
/*
store LSN into file, needed for Recovery to not be confused if a
DROP+CREATE happened (applying REDOs to the wrong table).
If such direct my_pwrite() to a fixed offset is too "hackish", I can
call ma_state_info_write() again but it will be less efficient.
*/
lsn_store(log_data, share.state.create_rename_lsn);
if (my_pwrite(file, log_data, LSN_STORE_SIZE,
sizeof(share.state.header) + 2, MYF(MY_NABP)))
goto err_no_lock;
my_free(log_data, MYF(0));
}
/* Enlarge files */
DBUG_PRINT("info", ("enlarge to keystart: %lu",
(ulong) share.base.keystart));
if (my_chsize(file,(ulong) share.base.keystart,0,MYF(0)))
goto err;
if (sync_dir && my_sync(file, MYF(0)))
goto err;
if (! (flags & HA_DONT_TOUCH_DATA))
{
#ifdef USE_RELOC
if (my_chsize(dfile,share.base.min_pack_length*ci->reloc_rows,0,MYF(0)))
goto err;
#endif
errpos=2;
if ((sync_dir && my_sync(dfile, MYF(0))) || my_close(dfile,MYF(0)))
goto err;
}
pthread_mutex_unlock(&THR_LOCK_maria);
res= 0;
my_free((char*) rec_per_key_part,MYF(0));
errpos=0;
if (my_close(file,MYF(0)))
res= my_errno;
DBUG_RETURN(res);
err:
pthread_mutex_unlock(&THR_LOCK_maria);
err_no_lock:
save_errno=my_errno;
switch (errpos) {
case 3:
VOID(my_close(dfile,MYF(0)));
/* fall through */
case 2:
if (! (flags & HA_DONT_TOUCH_DATA))
my_delete_with_symlink(fn_format(filename,name,"",MARIA_NAME_DEXT,
MY_UNPACK_FILENAME | MY_APPEND_EXT),
sync_dir);
/* fall through */
case 1:
VOID(my_close(file,MYF(0)));
if (! (flags & HA_DONT_TOUCH_DATA))
my_delete_with_symlink(fn_format(filename,name,"",MARIA_NAME_IEXT,
MY_UNPACK_FILENAME | MY_APPEND_EXT),
sync_dir);
}
my_free(log_data, MYF(MY_ALLOW_ZERO_PTR));
my_free((char*) rec_per_key_part, MYF(0));
DBUG_RETURN(my_errno=save_errno); /* return the fatal errno */
}
uint maria_get_pointer_length(ulonglong file_length, uint def)
{
DBUG_ASSERT(def >= 2 && def <= 7);
if (file_length) /* If not default */
{
#ifdef NOT_YET_READY_FOR_8_BYTE_POINTERS
if (file_length >= (ULL(1) << 56))
def=8;
else
#endif
if (file_length >= (ULL(1) << 48))
def=7;
else if (file_length >= (ULL(1) << 40))
def=6;
else if (file_length >= (ULL(1) << 32))
def=5;
else if (file_length >= (ULL(1) << 24))
def=4;
else if (file_length >= (ULL(1) << 16))
def=3;
else
def=2;
}
return def;
}
/*
Sort columns for records-in-block
IMPLEMENTATION
Sort columns in following order:
Fixed size, not null columns
Fixed length, null fields
Variable length fields (CHAR, VARCHAR)
Blobs
For same kind of fields, keep fields in original order
*/
static inline int sign(longlong a)
{
return a < 0 ? -1 : (a > 0 ? 1 : 0);
}
static int compare_columns(MARIA_COLUMNDEF **a_ptr, MARIA_COLUMNDEF **b_ptr)
{
MARIA_COLUMNDEF *a= *a_ptr, *b= *b_ptr;
enum en_fieldtype a_type, b_type;
a_type= ((a->type == FIELD_NORMAL || a->type == FIELD_CHECK) ?
FIELD_NORMAL : a->type);
b_type= ((b->type == FIELD_NORMAL || b->type == FIELD_CHECK) ?
FIELD_NORMAL : b->type);
if (a_type == FIELD_NORMAL && !a->null_bit)
{
if (b_type != FIELD_NORMAL || b->null_bit)
return -1;
return sign((long) (a->offset - b->offset));
}
if (b_type == FIELD_NORMAL && !b->null_bit)
return 1;
if (a_type == b_type)
return sign((long) (a->offset - b->offset));
if (a_type == FIELD_NORMAL)
return -1;
if (b_type == FIELD_NORMAL)
return 1;
if (a_type == FIELD_BLOB)
return 1;
if (b_type == FIELD_BLOB)
return -1;
return sign((long) (a->offset - b->offset));
}
/* Initialize data file */
int _ma_initialize_data_file(File dfile, MARIA_SHARE *share)
{
if (share->data_file_type == BLOCK_RECORD)
{
if (my_chsize(dfile, share->base.block_size, 0, MYF(MY_WME)))
return 1;
share->state.state.data_file_length= share->base.block_size;
_ma_bitmap_delete_all(share);
}
return 0;
}
Reference in a new issue View git blame Copy permalink