mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-01-18 21:12:26 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
mariadb/storage/maria/ma_create.c
unknown e27890cab0 WL#3072 Maria recovery 
* create page cache before initializing engine and not after, because
Maria's recovery needs a page cache
* make the creation of a bitmap page more crash-resistent
* bugfix (see ma_blockrec.c)
* back to old way: create an 8k bitmap page when creating table
* preparations for the UNDO phase: recreate TRNs
* preparations for Checkpoint: list of dirty pages, testing
of rec_lsn to know if page should be skipped during Recovery
(unused in this patch as no Checkpoint module pushed yet)
* maria_chk tags repaired table with a special LSN
* reworking all around in ma_recovery.c (less duplication)


mysys/my_realloc.c:
  noted an issue in my_realloc()
sql/mysqld.cc:
  page cache needs to be created before engines are initialized,
  because Maria's initialization may do a recovery which needs
  the page cache.
storage/maria/ha_maria.cc:
  update to new prototype
storage/maria/ma_bitmap.c:
  when creating the first bitmap page we used chsize to 8192 bytes then 
  pwrite (overwrite) the last 2 bytes (8191-8192). If crash between
  the two operations, this leaves a bitmap page full without its end
  marker. A later recovery may try to read this page and find it
  exists and misses a marker and conclude it's corrupted and fail.
  Changing the chsize to only 8190 bytes: recovery will then find
  the page is too short and recreate it entirely.
storage/maria/ma_blockrec.c:
  Fix for a bug: when executing a REDO, if the data page is created,
  data_file_length was increased before _ma_bitmap_set():
  _ma_bitmap_set() called _ma_read_bitmap_page() which, due to the
  increased data_file_length, expected to find a bitmap page on disk
  with a correct end marker; if the bitmap page didn't exist already
  in fact, this failed. Fixed by increasing data_file_length only after
  _ma_read_bitmap_page() has created the new bitmap page correctly.
  This bug could happen every time a REDO is about creating a new
  bitmap page.
storage/maria/ma_check.c:
  empty data file has a bitmap page
storage/maria/ma_control_file.c:
  useless parameter to ma_control_file_create_or_open(), just
  test if this is recovery.
storage/maria/ma_control_file.h:
  new prototype
storage/maria/ma_create.c:
  Back to how it was before: maria_create() creates an 8k bitmap page.
  Thus (bugfix) data_file_length needs to reflect this instead of being 0.
storage/maria/ma_loghandler.c:
  as ma_test1 and ma_test2 now use real transactions and not
  dummy_transaction_object, REDO for INSERT/UPDATE/DELETE are always
  about real transactions, can assert this.
  A function for Recovery to assign a short id to a table.
storage/maria/ma_loghandler.h:
  new function
storage/maria/ma_loghandler_lsn.h:
  maria_chk tags repaired tables with this LSN
storage/maria/ma_open.c:
  * enforce that DMLs on transactional tables use real transactions
  and not dummy_transaction_object.
  * test if table was repaired with maria_chk (which has to been
  seen as an import of an external table into the server), test
  validity of create_rename_lsn (header corruption detection)
  * comments.
storage/maria/ma_recovery.c:
  * preparations for the UNDO phase: recreate TRNs
  * preparations for Checkpoint: list of dirty pages, testing
  of rec_lsn to know if page should be skipped during Recovery
  (unused in this patch as no Checkpoint module pushed yet)
  * reworking all around (less duplication)
storage/maria/ma_recovery.h:
  a parameter to say if the UNDO phase should be skipped
storage/maria/maria_chk.c:
  tag repaired tables with a special LSN
storage/maria/maria_read_log.c:
  * update to new prototype
  * no UNDO phase in maria_read_log for now
storage/maria/trnman.c:
  * a function for Recovery to create a transaction (TRN), needed
  in the UNDO phase
  * a function for Recovery to grab an existing transaction, needed
  in the UNDO phase (rollback all existing transactions)
storage/maria/trnman_public.h:
  new functions
2007-08-29 16:43:01 +02:00

1233 lines
41 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], linkname[FN_REFLEN], *linkname_ptr;
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((uchar*) &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 uchar 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)
{
/*
NOTE1: here we change a field type FIELD_SKIP_ZERO ->
FIELD_NORMAL
*/
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.born_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;
DBUG_PRINT("info",("creating a transactional table"));
}
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 uchar */
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 uchar */
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 uchar */
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 uchar */
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,(uchar*) 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= share.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= LSN_IMPOSSIBLE;
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);
if (share.data_file_type == BLOCK_RECORD)
{
/*
we are going to create a first bitmap page, set data_file_length
to reflect this, before the state goes to disk
*/
share.state.state.data_file_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 && !maria_in_recovery)
{
/*
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);
/*
NOTE: For test_if_reopen() we need a real path name. Hence we need
MY_RETURN_REAL_PATH for every fn_format(filename, ...).
*/
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_RETURN_REAL_PATH | MY_APPEND_EXT);
}
else
{
fn_format(filename, ci->index_file_name, "", MARIA_NAME_IEXT,
MY_UNPACK_FILENAME | MY_RETURN_REAL_PATH |
(have_iext ? MY_REPLACE_EXT : MY_APPEND_EXT));
}
fn_format(linkname, name, "", MARIA_NAME_IEXT,
MY_UNPACK_FILENAME|MY_APPEND_EXT);
linkname_ptr= linkname;
/*
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
{
char *iext= strrchr(name, '.');
int have_iext= iext && !strcmp(iext, MARIA_NAME_IEXT);
fn_format(filename, name, "", MARIA_NAME_IEXT,
MY_UNPACK_FILENAME | MY_RETURN_REAL_PATH |
(have_iext ? MY_REPLACE_EXT : MY_APPEND_EXT));
linkname_ptr= NullS;
/*
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)
NOTE: The filename is compared against unique_file_name of every
open table. Hence we need a real path here.
*/
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(linkname_ptr, filename, 0, create_mode,
MYF(MY_WME|create_flag))) < 0)
goto err;
errpos=1;
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((uchar*) col_order, MYF(0));
goto err;
}
}
my_free((uchar*) 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 + 4];
uint total_rec_length= 0;
uint i;
log_array[TRANSLOG_INTERNAL_PARTS + 1].length= 1 + 2 + 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 + 1].length, MYF(0));
if ((log_data == NULL) ||
my_pread(file, 1 + 2 + 2 + log_data, kfile_size_before_extension,
0, MYF(MY_NABP)))
goto err;
/*
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);
int2store(log_data + 1 + 2, share.base.keystart);
log_array[TRANSLOG_INTERNAL_PARTS + 0].str= (char *)name;
/* we store the end-zero, for Recovery to just pass it to my_create() */
log_array[TRANSLOG_INTERNAL_PARTS + 0].length=
strlen(log_array[TRANSLOG_INTERNAL_PARTS + 0].str) + 1;
log_array[TRANSLOG_INTERNAL_PARTS + 1].str= log_data;
/* symlink description is also needed for re-creation by Recovery: */
log_array[TRANSLOG_INTERNAL_PARTS + 2].str= (char *)
(ci->data_file_name ? ci->data_file_name : empty_string);
log_array[TRANSLOG_INTERNAL_PARTS + 2].length=
strlen(log_array[TRANSLOG_INTERNAL_PARTS + 2].str) + 1;
log_array[TRANSLOG_INTERNAL_PARTS + 3].str= (char *)
(ci->index_file_name ? ci->index_file_name : empty_string);
log_array[TRANSLOG_INTERNAL_PARTS + 3].length=
strlen(log_array[TRANSLOG_INTERNAL_PARTS + 3].str) + 1;
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.
For now this record can serve when we apply logs to a backup,
so we sync it. This happens before the data file is created. If the
data file was created before, and we crashed before writing the log
record, at restart the table may be used, so we would not have a
trustable history in the log (impossible to apply this log to a
backup). The way we do it, if we crash before writing the log record
then there is no data file and the table cannot be used.
@todo Note that in case of TRUNCATE TABLE we also come here; for
Recovery to be able to finish TRUNCATE TABLE, instead of leaving a
half-truncated table, we should log the record at start of
maria_create(); for that we shouldn't write to the index file but to a
buffer (DYNAMIC_STRING), put the buffer into the record, then put the
buffer into the index file (so, change _ma_keydef_write() etc). That
would also enable Recovery to finish a CREATE TABLE. The final result
would be that we would be able to finish what the SQL layer has asked
for: it would be atomic.
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) ||
translog_flush(share.state.create_rename_lsn)))
goto err;
/*
store LSN into file, needed for Recovery to not be confused if a
DROP+CREATE happened (applying REDOs to the wrong table).
*/
share.kfile.file= file;
if (_ma_update_create_rename_lsn_on_disk(&share, FALSE))
goto err;
my_free(log_data, MYF(0));
}
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(linkname, name, "",MARIA_NAME_DEXT,
MY_UNPACK_FILENAME | MY_APPEND_EXT);
linkname_ptr= linkname;
create_flag=0;
}
else
{
fn_format(filename,name,"", MARIA_NAME_DEXT,
MY_UNPACK_FILENAME | MY_APPEND_EXT);
linkname_ptr= NullS;
create_flag=MY_DELETE_OLD;
}
if ((dfile=
my_create_with_symlink(linkname_ptr, filename, 0, create_mode,
MYF(MY_WME | create_flag | sync_dir))) < 0)
goto err;
errpos=3;
if (_ma_initialize_data_file(&share, dfile))
goto err;
}
/* 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
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(MARIA_SHARE *share, File dfile)
{
if (share->data_file_type == BLOCK_RECORD)
{
share->bitmap.block_size= share->base.block_size;
share->bitmap.file.file = dfile;
return _ma_bitmap_create_first(share);
}
/*
So, in BLOCK_RECORD, a freshly created datafile is one page long; while in
other formats it is 0-byte long.
*/
return 0;
}
/**
@brief Writes create_rename_lsn to disk, optionally forces
This is for special cases where:
- we don't want to write the full state to disk (so, not call
_ma_state_info_write()) because some parts of the state may be
currently inconsistent, or because it would be overkill
- we must sync this LSN immediately for correctness.
@param share table's share
@param do_sync if the write should be forced to disk
@return Operation status
@retval 0 ok
@retval 1 error (disk problem)
*/
int _ma_update_create_rename_lsn_on_disk(MARIA_SHARE *share, my_bool do_sync)
{
char buf[LSN_STORE_SIZE];
File file= share->kfile.file;
DBUG_ASSERT(file >= 0);
lsn_store(buf, share->state.create_rename_lsn);
return (my_pwrite(file, buf, sizeof(buf),
sizeof(share->state.header) + 2, MYF(MY_NABP)) ||
(do_sync && my_sync(file, MYF(0))));
}
Reference in a new issue View git blame Copy permalink