mariadb/fsp/fsp0fsp.c
inaam 0757affd7b branches/innodb+: Merged revisions 5525:5971 from branches/zip
------------------------------------------------------------------------
  r5971 | marko | 2009-09-23 09:03:51 -0400 (Wed, 23 Sep 2009) | 2 lines
  
  branches/zip: os_file_pwrite(): Make the code compile in InnoDB Hot Backup
  when the pwrite system call is not available.
  ------------------------------------------------------------------------
  r5956 | calvin | 2009-09-22 19:30:10 -0400 (Tue, 22 Sep 2009) | 4 lines
  
  branches/zip: remove handler0vars.h from Makefile.am
  
  Left over from r5950.
  
  ------------------------------------------------------------------------
  r5951 | calvin | 2009-09-22 11:17:01 -0400 (Tue, 22 Sep 2009) | 4 lines
  
  branches/zip: adjust CMake file to work with old versions of MySQL
  
  Tested with MySQL 5.1.38 and 5.1.30.
  
  ------------------------------------------------------------------------
  r5950 | calvin | 2009-09-22 02:42:46 -0400 (Tue, 22 Sep 2009) | 17 lines
  
  branches/zip: adjust Windows loading method for 5.1.38
  
  Starting at 5.1.38, MySQL server exports symbols needed
  for dynamic plugin on Windows. There is no need for
  Windows specific loading. Also, the CMake files are
  simplified in 5.1.38.
  
  When WITH_INNOBASE_STORAGE_ENGINE is specified during
  configuration (win\configure.js), InnoDB is built as
  a static library. Otherwise, a dynamic InnoDB will be
  built (ha_innodb.dll).
  
  CMakeLists.txt requires minor changes in order to work
  with MySQL prior to 5.1.38. The changes will be in a
  separate patch.
  
  This patch addresses Mantis issue#286.
  ------------------------------------------------------------------------
  r5945 | calvin | 2009-09-21 10:53:22 -0400 (Mon, 21 Sep 2009) | 4 lines
  
  branches/zip: fix a type in r5935
  
  Should be innodb_open_files, spotted by Michael.
  
  ------------------------------------------------------------------------
  r5940 | vasil | 2009-09-21 01:26:04 -0400 (Mon, 21 Sep 2009) | 4 lines
  
  branches/zip:
  
  Add ChangeLog entries for c5938.
  
  ------------------------------------------------------------------------
  r5938 | calvin | 2009-09-19 03:14:25 -0400 (Sat, 19 Sep 2009) | 41 lines
  
  branches/zip: Merge revisions 2584:2956 from branches/6.0,
  except c2932.
  
  Bug#37232 and bug#31183 were fixed in the 6.0 branch only.
  They should be fixed in the plugin too, specially MySQL 6.0
  is discontinued at this point.
  
      ------------------------------------------------------------------------
      r2604 | inaam | 2008-08-21 09:37:06 -0500 (Thu, 21 Aug 2008) | 8 lines
  
      branches/6.0     bug#37232
  
      Relax locking behaviour for REPLACE INTO t SELECT ... FROM t1.
      Now SELECT on t1 is performed as a consistent read when the isolation
      level is set to READ COMMITTED.
  
      Reviewed by: Heikki
      ------------------------------------------------------------------------
      r2605 | inaam | 2008-08-21 09:59:33 -0500 (Thu, 21 Aug 2008) | 7 lines
  
      branches/6.0
  
      Added a comment to clarify why distinct calls to read MySQL binary
      log file name and log position do not entail any race condition.
  
      Suggested by: Heikki
      ------------------------------------------------------------------------
      r2956 | inaam | 2008-11-04 04:47:30 -0600 (Tue, 04 Nov 2008) | 11 lines
  
      branches/6.0  bug#31183
  
      If the system tablespace runs out of space because 'autoextend' is
      not specified with innodb_data_file_path there was no error message
      printed to the error log. The client would get 'table full' error.
      This patch prints an appropriate error message to the error log.
  
      rb://43
  
      Approved by: Marko
      ------------------------------------------------------------------------
  
  ------------------------------------------------------------------------
  r5935 | calvin | 2009-09-18 17:08:02 -0400 (Fri, 18 Sep 2009) | 6 lines
  
  branches/zip: fix bug#44338; minor non-functional changes
  
  Bug#44338 innodb has message about non-existing option 
  innodb_max_files_open. Change the option to innodb_open_files.
  The fix was committed into 6.0 branch.
  
  ------------------------------------------------------------------------
  r5934 | vasil | 2009-09-18 13:06:46 -0400 (Fri, 18 Sep 2009) | 4 lines
  
  branches/zip:
  
  Fix typo.
  
  ------------------------------------------------------------------------
  r5924 | vasil | 2009-09-18 00:59:30 -0400 (Fri, 18 Sep 2009) | 4 lines
  
  branches/zip:
  
  White space and formatting cleanup in the ChangeLog
  
  ------------------------------------------------------------------------
  r5922 | marko | 2009-09-17 02:32:08 -0400 (Thu, 17 Sep 2009) | 4 lines
  
  branches/zip: innodb-zip.test: Make the test work with zlib 1.2.3.3.
  Apparently, the definition of compressBound() has slightly changed.
  
  This has been filed as Mantis Issue #345.
  ------------------------------------------------------------------------
  r5920 | vasil | 2009-09-16 14:47:22 -0400 (Wed, 16 Sep 2009) | 4 lines
  
  branches/zip:
  
  Add ChangeLog entries for r5916.
  
  ------------------------------------------------------------------------
  r5919 | vasil | 2009-09-16 14:37:13 -0400 (Wed, 16 Sep 2009) | 4 lines
  
  branches/zip:
  
  Whitespace cleanup in the ChangeLog.
  
  ------------------------------------------------------------------------
  r5917 | marko | 2009-09-16 05:56:23 -0400 (Wed, 16 Sep 2009) | 1 line
  
  branches/zip: innobase_get_cset_width(): Cache the value of current_thd.
  ------------------------------------------------------------------------
  r5916 | marko | 2009-09-16 05:54:43 -0400 (Wed, 16 Sep 2009) | 128 lines
  
  branches/zip: Merge revisions 5622:5912 from branches/5.1, except r5700
  (changes to CMakeLists.txt)
  
    ------------------------------------------------------------------------
    r5622 | vasil | 2009-08-03 15:27:00 +0300 (Mon, 03 Aug 2009) | 20 lines
    Changed paths:
       M /branches/5.1/Makefile.am
  
    branches/5.1:
  
    Merge a change from MySQL:
  
      ------------------------------------------------------------
      revno: 2988
      committer: Satya B <satya.bn@sun.com>
      branch nick: mysql-5.1-bugteam
      timestamp: Wed 2009-07-01 11:06:05 +0530
      message:
        Fix build failure after applying Innodb snapshot 5.1-ss5282
  
        After applying Innodb snapshot 5.1-ss5282, build was broken
        because of missing header file.
  
        Adding the header file to Makefile.am after informing the
        innodb developers.
      modified:
        storage/innobase/Makefile.am
    ------------------------------------------------------------------------
    r5740 | jyang | 2009-09-03 06:33:47 +0300 (Thu, 03 Sep 2009) | 5 lines
    Changed paths:
       M /branches/5.1/handler/ha_innodb.cc
       M /branches/5.1/include/db0err.h
       A /branches/5.1/mysql-test/innodb_bug46000.result
       A /branches/5.1/mysql-test/innodb_bug46000.test
  
    branches/5.1: Disallow creating index with the name of
    "GEN_CLUST_INDEX" which is reserved for the default system
    primary index. (Bug #46000) rb://149 approved by Sunny Bains.
    ------------------------------------------------------------------------
    r5741 | jyang | 2009-09-03 07:16:01 +0300 (Thu, 03 Sep 2009) | 5 lines
    Changed paths:
       M /branches/5.1/dict/dict0dict.c
       M /branches/5.1/handler/ha_innodb.cc
       A /branches/5.1/mysql-test/innodb_bug44369.result
       A /branches/5.1/mysql-test/innodb_bug44369.test
       M /branches/5.1/row/row0mysql.c
  
    branches/5.1: Block creating table with column name conflicting
    with Innodb reserved key words. (Bug #44369) rb://151 approved
    by Sunny Bains.
    ------------------------------------------------------------------------
    r5757 | jyang | 2009-09-04 04:26:13 +0300 (Fri, 04 Sep 2009) | 3 lines
    Changed paths:
       M /branches/5.1/handler/ha_innodb.cc
       M /branches/5.1/include/db0err.h
       D /branches/5.1/mysql-test/innodb_bug46000.result
       D /branches/5.1/mysql-test/innodb_bug46000.test
  
    branches/5.1: Revert change in 5740. Making the fix in a subsequent
    check in.
    ------------------------------------------------------------------------
    r5760 | jyang | 2009-09-04 07:07:34 +0300 (Fri, 04 Sep 2009) | 3 lines
    Changed paths:
       M /branches/5.1/dict/dict0dict.c
       M /branches/5.1/handler/ha_innodb.cc
       D /branches/5.1/mysql-test/innodb_bug44369.result
       D /branches/5.1/mysql-test/innodb_bug44369.test
       M /branches/5.1/row/row0mysql.c
  
    branches/5.1: This is to revert change 5741. A return status for
    create_table_def() needs to be fixed.
    ------------------------------------------------------------------------
    r5797 | calvin | 2009-09-09 18:26:29 +0300 (Wed, 09 Sep 2009) | 3 lines
    Changed paths:
       M /branches/5.1/handler/ha_innodb.cc
  
    branches/5.1: merge change from 5.1.38
  
    HA_ERR_TOO_MANY_CONCURRENT_TRXS is added in 5.1.38.
    ------------------------------------------------------------------------
    r5799 | calvin | 2009-09-09 20:47:31 +0300 (Wed, 09 Sep 2009) | 10 lines
    Changed paths:
       M /branches/5.1/handler/ha_innodb.cc
  
    branches/5.1: fix bug#46256
  
    Allow tables to be dropped even if the collation is not found,
    but issue a warning.
  
    Could not find an easy way to add mysql-test since it requires
    changes to charsets and restarting the server. Tests were
    executed manually.
  
    Approved by: Heikki (on IM)
    ------------------------------------------------------------------------
    r5805 | vasil | 2009-09-10 08:41:48 +0300 (Thu, 10 Sep 2009) | 7 lines
    Changed paths:
       M /branches/5.1/handler/ha_innodb.cc
  
    branches/5.1:
  
    Fix a compilation warning caused by c5799:
  
    handler/ha_innodb.cc: In function 'void innobase_get_cset_width(ulint, ulint*, ulint*)':
    handler/ha_innodb.cc:830: warning: format '%d' expects type 'int', but argument 2 has type 'ulint'
    ------------------------------------------------------------------------
    r5834 | jyang | 2009-09-11 00:43:05 +0300 (Fri, 11 Sep 2009) | 5 lines
    Changed paths:
       M /branches/5.1/dict/dict0dict.c
       M /branches/5.1/handler/ha_innodb.cc
       A /branches/5.1/mysql-test/innodb_bug44369.result
       A /branches/5.1/mysql-test/innodb_bug44369.test
       M /branches/5.1/row/row0mysql.c
  
    branches/5.1: Block creating table with column name conflicting
    with Innodb reserved key words. (Bug #44369) rb://151 approved
    by Sunny Bains.
    ------------------------------------------------------------------------
    r5895 | jyang | 2009-09-15 03:39:21 +0300 (Tue, 15 Sep 2009) | 5 lines
    Changed paths:
       M /branches/5.1/handler/ha_innodb.cc
       A /branches/5.1/mysql-test/innodb_bug46000.result
       A /branches/5.1/mysql-test/innodb_bug46000.test
  
    branches/5.1: Disallow creating index with the name of
    "GEN_CLUST_INDEX" which is reserved for the default system
    primary index. (Bug #46000) rb://149 approved by Marko Makela.
    ------------------------------------------------------------------------
  ------------------------------------------------------------------------
  r5910 | marko | 2009-09-16 04:07:21 -0400 (Wed, 16 Sep 2009) | 9 lines
  
  branches/zip: Introduce UNIV_LOG_LSN_DEBUG and MLOG_LSN for redo log
  diagnostics.  This was written in order to better track down
  Issue #313 in InnoDB Hot Backup.
  
  MLOG_LSN: A new redo log entry type, for recording the current log
  sequence number (LSN).  This will be checked in an assertion in
  recv_parse_log_rec().
  
  rb://161, discussed with Sunny and Vasil.
  ------------------------------------------------------------------------
  r5899 | marko | 2009-09-15 07:26:01 -0400 (Tue, 15 Sep 2009) | 4 lines
  
  branches/zip: ut0ut.h: Do not #include "os0sync.h" #ifdef UNIV_HOTBACKUP.
  Since r5872, the InnoDB Hot Backup build was broken.
  Fix it by not defining any thread synchronization primitives in ut0ut.h.
  InnoDB Hot Backup is a single-threaded program.
  ------------------------------------------------------------------------
  r5898 | marko | 2009-09-15 06:18:50 -0400 (Tue, 15 Sep 2009) | 2 lines
  
  branches/zip: Add */.dirstamp to svn:ignore,
  for https://svn.innodb.com/svn/hotbackup/branches/3.5
  ------------------------------------------------------------------------
  r5897 | marko | 2009-09-15 04:29:00 -0400 (Tue, 15 Sep 2009) | 8 lines
  
  branches/zip: Avoid bogus messages about latching order violations when
  UNIV_SYNC_DEBUG is defined.
  
  sync_thread_levels_g(): Add the parameter "warn".  Do not print
  anything unless it is set.
  
  sync_thread_add_level(): Pass warn=TRUE to sync_thread_levels_g()
  when the check is within an assertion; FALSE if it is not.
  ------------------------------------------------------------------------
  r5893 | inaam | 2009-09-14 11:20:48 -0400 (Mon, 14 Sep 2009) | 10 lines
  
  branches/zip  rb://159
  
  In case of pages that are not made young the counter is incremented
  only when the page in question is 'old'. In case of pages that are
  made young the counter is incremented in case of all pages. For apple
  to apple comparison this patch changes the 'young-making' counter to
  consider only 'old' blocks.
  
  Approved by: Marko
  
  ------------------------------------------------------------------------
  r5889 | vasil | 2009-09-14 05:17:18 -0400 (Mon, 14 Sep 2009) | 5 lines
  
  branches/zip:
  
  Add missing return statement in the test program that could have
  caused a warning.
  
  ------------------------------------------------------------------------
  r5888 | vasil | 2009-09-14 04:38:45 -0400 (Mon, 14 Sep 2009) | 40 lines
  
  branches/zip:
  
  Back-merge c5880 and c5881 from branches/embedded-1.0:
  
    ------------------------------------------------------------------------
    r5880 | vasil | 2009-09-12 17:28:44 +0300 (Sat, 12 Sep 2009) | 18 lines
    Changed paths:
       M /branches/embedded-1.0/configure.in
       M /branches/embedded-1.0/include/os0sync.h
       M /branches/embedded-1.0/srv/srv0start.c
    
    branches/embedded-1.0:
    
    Clean up and simplify the code that surrounds the atomic ops:
    
    * Simplify the code that prints what atomics are used:
    Instead of repeating the same conditions on which each atomics are used
    use just one printf that prints a variable defined by the code which
    chooses what atomics to use.
    
    * In os0sync.h pick up each atomic variant only if it has been selected
    by autoconf (based on IB_ATOMIC_MODE_* macros). Define the startup message
    to be printed.
    
    * In configure.in: check what user has chosen and if he has chosen
    something that is not available, emit an error. If nothing has been chosen
    explicitly by the user, auto select an option according to the described
    logic in configure.in.
    
    ------------------------------------------------------------------------
    r5881 | vasil | 2009-09-12 20:08:27 +0300 (Sat, 12 Sep 2009) | 4 lines
    Changed paths:
       M /branches/embedded-1.0/configure.in
    
    branches/embedded-1.0:
    
    Fix syntax error in test program.
    
    ------------------------------------------------------------------------
  
  ------------------------------------------------------------------------
  r5875 | vasil | 2009-09-12 08:11:25 -0400 (Sat, 12 Sep 2009) | 4 lines
  
  branches/zip:
  
  Remove unnecessary macro.
  
  ------------------------------------------------------------------------
  r5872 | vasil | 2009-09-12 05:35:17 -0400 (Sat, 12 Sep 2009) | 5 lines
  
  branches/zip:
  
  Explicitly include os0sync.h to the places where HAVE_ATOMIC_BUILTINS and
  INNODB_RW_LOCKS_USE_ATOMICS are used to avoid potential problems.
  
  ------------------------------------------------------------------------
  r5871 | vasil | 2009-09-12 05:25:44 -0400 (Sat, 12 Sep 2009) | 6 lines
  
  branches/zip:
  
  Rename HAVE_SOLARIS_ATOMICS to HAVE_IB_SOLARIS_ATOMICS and 
  IB_HAVE_PAUSE_INSTRUCTION to HAVE_IB_PAUSE_INSTRUCTION so they
  all follow the same HAVE_IB_* convention.
  
  ------------------------------------------------------------------------
  r5870 | vasil | 2009-09-12 05:13:44 -0400 (Sat, 12 Sep 2009) | 7 lines
  
  branches/zip:
  
  Define HAVE_ATOMIC_BUILTINS and INNODB_RW_LOCKS_USE_ATOMICS in os0sync.h
  instead of in univ.i. The code expects os_*() macros to be present if
  HAVE_ATOMIC_BUILTINS and INNODB_RW_LOCKS_USE_ATOMICS are defined. So define
  them next to defining the os_*() macros.
  
  ------------------------------------------------------------------------
  r5869 | vasil | 2009-09-12 04:33:11 -0400 (Sat, 12 Sep 2009) | 15 lines
  
  branches/zip:
  
  Include ut0auxconf.h only if none of the macros it would define is defined.
  The check when to include this header was outdated from the time when there
  was only one macro involved.
  
  Move the atomics checks that are in univ.i outside of
  #if windows ... #else ... #endif
  This simplifies the code and removes some duplicates like defining
  HAVE_ATOMIC_BUILTINS if HAVE_WINDOWS_ATOMICS is defined in both branches.
  
  Do not define the same macro HAVE_ATOMIC_PTHREAD_T for different events.
  Instead define HAVE_IB_ATOMIC_PTHREAD_T_GCC and
  HAVE_IB_ATOMIC_PTHREAD_T_SOLARIS.
  
  ------------------------------------------------------------------------
  r5868 | vasil | 2009-09-12 04:01:17 -0400 (Sat, 12 Sep 2009) | 6 lines
  
  branches/zip:
  
  Move the check whether to include ut0auxconf.h before everything because
  we are now even checking for GCC atomics, we relied on MySQL to define
  this macro before.
  
  ------------------------------------------------------------------------
  r5867 | vasil | 2009-09-12 03:43:45 -0400 (Sat, 12 Sep 2009) | 4 lines
  
  branches/zip:
  
  Update comment to reflect reality.
  
  ------------------------------------------------------------------------
  r5866 | vasil | 2009-09-12 03:30:08 -0400 (Sat, 12 Sep 2009) | 5 lines
  
  branches/zip:
  
  Add the check for GCC atomics to ut0auxconf* (copied from plug.in) because
  we no longer rely on MySQL's HAVE_GCC_ATOMIC_BUILTINS.
  
  ------------------------------------------------------------------------
  r5865 | vasil | 2009-09-12 03:26:03 -0400 (Sat, 12 Sep 2009) | 10 lines
  
  branches/zip:
  
  Simplify the compile time checks by splittig them into 5 independent checks:
  
  * Whether GCC atomics are available
  * Whether pthread_t can be used by GCC atomics
  * Whether Solaris libc atomics are available
  * Whether pthread_t can be used by Solaris libs atomics
  * Checking the size of pthread_t
  
  ------------------------------------------------------------------------
  r5864 | vasil | 2009-09-12 03:22:55 -0400 (Sat, 12 Sep 2009) | 4 lines
  
  branches/zip:
  
  Include string.h which is needed for memset().
  
  ------------------------------------------------------------------------
  r5863 | vasil | 2009-09-12 03:07:08 -0400 (Sat, 12 Sep 2009) | 10 lines
  
  branches/zip:
  
  Check that pthread_t can indeed be passed to Solaris atomic functions, instead
  of assuming that it can be passed if 0 can be assigned to it. It could be that:
  * 0 can be assigned, but pthread_t cannot be passed and
  * 0 cannot be assigned but pthread_t can be passed
  
  Better to check what we are interested in, not something else and make
  assumptions.
  
  ------------------------------------------------------------------------
  r5858 | vasil | 2009-09-11 13:46:47 -0400 (Fri, 11 Sep 2009) | 4 lines
  
  branches/zip:
  
  Fix the indentation of the closing bracket.
  
  ------------------------------------------------------------------------
  r5826 | marko | 2009-09-10 07:29:46 -0400 (Thu, 10 Sep 2009) | 12 lines
  
  branches/zip: Roll back recovered dictionary transactions before
  dropping incomplete indexes (Issue #337).
  
  trx_rollback_or_clean_recovered(ibool all): New function, split from
  trx_rollback_or_clean_all_recovered().  all==FALSE will only roll back
  dictionary transactions.
  
  recv_recovery_from_checkpoint_finish(): Call
  trx_rollback_or_clean_recovered(FALSE) before
  row_merge_drop_temp_indexes().
  
  rb://158 approved by Sunny Bains
  ------------------------------------------------------------------------
  r5825 | marko | 2009-09-10 06:47:09 -0400 (Thu, 10 Sep 2009) | 20 lines
  
  branches/zip: Reduce mutex contention that was introduced when
  addressing Bug #45015 (Issue #316), in r5703.
  
  buf_page_set_accessed_make_young(): New auxiliary function, called by
  buf_page_get_zip(), buf_page_get_gen(),
  buf_page_optimistic_get_func(). Call ut_time_ms() outside of
  buf_pool_mutex. Use cached access_time.
  
  buf_page_set_accessed(): Add the parameter time_ms, so that
  ut_time_ms() need not be called while holding buf_pool_mutex.
  
  buf_page_optimistic_get_func(), buf_page_get_known_nowait(): Read
  buf_page_t::access_time without holding buf_pool_mutex. This should be
  OK, because the field is only used for heuristic purposes.
  
  buf_page_peek_if_too_old(): If buf_pool->freed_page_clock == 0, return
  FALSE, so that we will not waste time moving blocks in the LRU list in
  the warm-up phase or when the workload fits in the buffer pool.
  
  rb://156 approved by Sunny Bains
  ------------------------------------------------------------------------
  r5822 | marko | 2009-09-10 06:10:20 -0400 (Thu, 10 Sep 2009) | 1 line
  
  branches/zip: buf_page_release(): De-stutter the function comment.
  ------------------------------------------------------------------------
  r5804 | marko | 2009-09-10 01:29:31 -0400 (Thu, 10 Sep 2009) | 1 line
  
  branches/zip: trx_cleanup_at_db_startup(): Fix a typo in comment.
  ------------------------------------------------------------------------
  r5798 | calvin | 2009-09-09 11:28:10 -0400 (Wed, 09 Sep 2009) | 5 lines
  
  branches/zip:
  
  HA_ERR_TOO_MANY_CONCURRENT_TRXS is added in 5.1.38.
  But the plugin should still work with previous versions
  of MySQL.
  ------------------------------------------------------------------------
  r5792 | vasil | 2009-09-09 09:35:58 -0400 (Wed, 09 Sep 2009) | 32 lines
  
  branches/zip:
  
  Fix a bug in manipulating the variable innodb_old_blocks_pct:
  
  for any value assigned it got that value -1, except for 75. When
  assigned 75, it got 75.
  
    mysql> set global innodb_old_blocks_pct=15;
    Query OK, 0 rows affected (0.00 sec)
    
    mysql> show variables like 'innodb_old_blocks_pct';
    +-----------------------+-------+
    | Variable_name         | Value |
    +-----------------------+-------+
    | innodb_old_blocks_pct | 14    | 
    +-----------------------+-------+
    1 row in set (0.00 sec)
    
    mysql> set global innodb_old_blocks_pct=75;
    Query OK, 0 rows affected (0.00 sec)
    
    mysql> show variables like 'innodb_old_blocks_pct';
    +-----------------------+-------+
    | Variable_name         | Value |
    +-----------------------+-------+
    | innodb_old_blocks_pct | 75    | 
    +-----------------------+-------+
  
  After the fix it gets exactly what was assigned.
  
  Approved by:	Marko (via IM)
  
  ------------------------------------------------------------------------
  r5783 | marko | 2009-09-09 03:25:00 -0400 (Wed, 09 Sep 2009) | 1 line
  
  branches/zip: buf_page_is_accessed(): Correct the function comment.
  ------------------------------------------------------------------------
  r5782 | marko | 2009-09-09 03:00:59 -0400 (Wed, 09 Sep 2009) | 2 lines
  
  branches/zip: buf_page_peek_if_too_old(): Silence a compiler warning
  that was introduced in r5779 on 32-bit systems.
  ------------------------------------------------------------------------
  r5780 | marko | 2009-09-09 02:50:50 -0400 (Wed, 09 Sep 2009) | 1 line
  
  branches/zip: ut_time_ms(): Return ulint, not uint.
  ------------------------------------------------------------------------
  r5779 | marko | 2009-09-09 02:17:19 -0400 (Wed, 09 Sep 2009) | 2 lines
  
  branches/zip: buf_page_peek_if_too_old(): Make the bitmasking work when
  buf_pool->freed_page_clock is wider than 32 bits.
  ------------------------------------------------------------------------
  r5777 | marko | 2009-09-08 11:50:25 -0400 (Tue, 08 Sep 2009) | 2 lines
  
  branches/zip: Remove BUF_LRU_INITIAL_RATIO, which should have been removed
  together with buf_LRU_get_recent_limit().
  ------------------------------------------------------------------------
  r5775 | calvin | 2009-09-07 17:15:05 -0400 (Mon, 07 Sep 2009) | 13 lines
  
  branches/zip: Build InnoDB on Windows with UNIV_HOTBACKUP
  
  The changes are non-functional changes for normal InnoDB,
  but needed for building the Hot Backup on Windows (with
  UNIV_HOTBACKUP defined).
  
  - Define os_aio_use_native_aio for HB.
  - Do not acquire seek mutexes for backup since HB is single threaded.
  - Do not use srv_flush_log_at_trx_commit for HB build
  
  rb://155
  
  Approved by: Marko
  ------------------------------------------------------------------------
  r5752 | marko | 2009-09-03 10:55:51 -0400 (Thu, 03 Sep 2009) | 10 lines
  
  branches/zip: recv_recover_page_func(): Write the log sequence number
  to the compressed page, if there is one.  Previously, the function only
  wrote the LSN to the uncompressed page.
  
  It is not clear why recv_recover_page_func() is updating FIL_PAGE_LSN
  in the buffer pool.  The log sequence number will be stamped on the
  page when it is flushed to disk, in buf_flush_init_for_writing().
  I noticed this inconsistency when analyzing Issue #313, but this patch
  does not fix it.  That is no surprise, since FIL_PAGE_LSN should only
  matter on disk files, not in the buffer pool.
  ------------------------------------------------------------------------
  r5751 | marko | 2009-09-03 10:36:15 -0400 (Thu, 03 Sep 2009) | 7 lines
  
  branches/zip: row_merge(): Remove a bogus debug assertion
  that was triggered when creating an index on an empty table.
  
  row_merge_sort(): Add debug assertions and comments that justify
  the loop termination condition.
  
  The bogus assertion ut_ad(ihalf > 0) was reported by Michael.
  ------------------------------------------------------------------------
  r5748 | marko | 2009-09-03 07:05:44 -0400 (Thu, 03 Sep 2009) | 1 line
  
  branches/zip: MLOG_MULTI_REC_END: Correct the comment.
  ------------------------------------------------------------------------
  r5747 | marko | 2009-09-03 06:46:38 -0400 (Thu, 03 Sep 2009) | 2 lines
  
  branches/zip: recv_scan_log_recs(): Replace while with do...while,
  because the termination condition will always hold on the first iteration.
  ------------------------------------------------------------------------
  r5746 | marko | 2009-09-03 04:55:36 -0400 (Thu, 03 Sep 2009) | 2 lines
  
  branches/zip: log_reserve_and_write_fast(): Do not cache the log_sys pointer
  in a local variable.
  ------------------------------------------------------------------------
  r5745 | marko | 2009-09-03 04:38:22 -0400 (Thu, 03 Sep 2009) | 2 lines
  
  branches/zip: log_check_log_recs(): Enclose in #ifdef UNIV_LOG_DEBUG.
  Add const qualifiers.
  ------------------------------------------------------------------------
  r5744 | marko | 2009-09-03 04:28:35 -0400 (Thu, 03 Sep 2009) | 1 line
  
  branches/zip: ut_align(): Make ptr const, like in ut_align_down().
  ------------------------------------------------------------------------
  r5743 | marko | 2009-09-03 02:36:12 -0400 (Thu, 03 Sep 2009) | 3 lines
  
  branches/zip: log_reserve_and_write_fast(): Remove the redundant
  output parameter "success".
  Success is also indicated by a nonzero return value.
  ------------------------------------------------------------------------
  r5736 | marko | 2009-09-02 03:53:19 -0400 (Wed, 02 Sep 2009) | 1 line
  
  branches/zip: Enclose some timestamp functions in #ifndef UNIV_HOTBACKUP.
  ------------------------------------------------------------------------
  r5735 | marko | 2009-09-02 03:43:09 -0400 (Wed, 02 Sep 2009) | 2 lines
  
  branches/zip: univ.i: Do not undefine PACKAGE or VERSION.
  InnoDB source code does not refer to these macros.
  ------------------------------------------------------------------------
  r5734 | sunny | 2009-09-02 03:08:45 -0400 (Wed, 02 Sep 2009) | 2 lines
  
  branches/zip: Update ChangeLog with r5733 changes.
  
  ------------------------------------------------------------------------
  r5733 | sunny | 2009-09-02 03:05:15 -0400 (Wed, 02 Sep 2009) | 6 lines
  
  branches/zip: Fix a regression introduced by the fix for bug#26316. We check
  whether a transaction holds any AUTOINC locks before we acquire the kernel
  mutex and release those locks.
  
  Fix for rb://153. Approved by Marko.
  
  ------------------------------------------------------------------------
  r5716 | vasil | 2009-08-31 03:47:49 -0400 (Mon, 31 Aug 2009) | 9 lines
  
  branches/zip:
  
  Fix Bug#46718 InnoDB plugin incompatible with gcc 4.1 (at least: on PPC): "Undefined symbol"
  
  by implementing our own check in plug.in instead of using the result from
  the check from MySQL because it is insufficient.
  
  Approved by:	Marko (rb://154)
  
  ------------------------------------------------------------------------
  r5714 | marko | 2009-08-31 02:10:10 -0400 (Mon, 31 Aug 2009) | 5 lines
  
  branches/zip: buf_chunk_not_freed(): Do not acquire block->mutex unless
  block->page.state == BUF_BLOCK_FILE_PAGE.  Check that block->page.state
  makes sense.
  
  Approved by Sunny Bains over the IM.
  ------------------------------------------------------------------------
  r5709 | inaam | 2009-08-28 02:22:46 -0400 (Fri, 28 Aug 2009) | 5 lines
  
  branches/zip rb://152
  
  Disable display of deprecated parameter innodb_file_io_threads in
  'show variables'.
  
  ------------------------------------------------------------------------
  r5708 | inaam | 2009-08-27 18:43:32 -0400 (Thu, 27 Aug 2009) | 4 lines
  
  branches/zip
  
  Remove redundant TRUE : FALSE from the return statement
  
  ------------------------------------------------------------------------
  r5707 | inaam | 2009-08-27 12:20:35 -0400 (Thu, 27 Aug 2009) | 6 lines
  
  branches/zip
  
  Remove unused macros as we erased the random readahead code in r5703.
  Also fixed some comments.
  
  
  ------------------------------------------------------------------------
  r5706 | inaam | 2009-08-27 12:00:27 -0400 (Thu, 27 Aug 2009) | 20 lines
  
  branches/zip rb://147
  
  Done away with following two status variables:
  
  innodb_buffer_pool_read_ahead_rnd
  innodb_buffer_pool_read_ahead_seq
  
  Introduced two new status variables:
  innodb_buffer_pool_read_ahead = number of pages read as part of
  readahead since server startup
  innodb_buffer_pool_read_ahead_evicted = number of pages that are read
  in as readahead but were evicted before ever being accessed since
  server startup i.e.: a measure of how badly our readahead is
  performing
  
  SHOW INNODB STATUS will show two extra numbers in buffer pool section:
  pages read ahead/sec and pages evicted without access/sec
  
  Approved by: Marko
  
  ------------------------------------------------------------------------
  r5705 | marko | 2009-08-27 07:56:24 -0400 (Thu, 27 Aug 2009) | 11 lines
  
  branches/zip: dict_index_find_cols(): On column name lookup failure,
  return DB_CORRUPTION (HA_ERR_CRASHED) instead of abnormally
  terminating the server.  Also, disable the previously added diagnostic
  output to the error log, because mysql-test-run does not like extra
  output in the error log.  (Bug #44571)
  
  dict_index_add_to_cache(): Handle errors from dict_index_find_cols().
  
  mysql-test/innodb_bug44571.test: A test case for triggering the bug.
  
  rb://135 approved by Sunny Bains.
  ------------------------------------------------------------------------
  r5704 | marko | 2009-08-27 04:31:17 -0400 (Thu, 27 Aug 2009) | 32 lines
  
  branches/zip: Fix a critical bug in fast index creation that could
  corrupt the created indexes.
  
  row_merge(): Make "half" an in/out parameter. Determine the offset of
  half the output file. Copy the last blocks record-by-record instead of
  block-by-block, so that the records can be counted. Check that the
  input and output have matching n_rec.
  
  row_merge_sort(): Do not assume that two blocks of size N are merged
  into a block of size 2*N. The output block can be shorter than the
  input if the last page of each input block is almost empty. Use an
  accurate termination condition, based on the "half" computed by
  row_merge().
  
  row_merge_read(), row_merge_write(), row_merge_blocks(): Add debug output.
  
  merge_file_t, row_merge_file_create(): Add n_rec, the number of records
  in the merge file.
  
  row_merge_read_clustered_index(): Update n_rec.
  
  row_merge_blocks(): Update and check n_rec.
  
  row_merge_blocks_copy(): New function, for copying the last blocks in
  row_merge().  Update and check n_rec.
  
  This bug was discovered with a user-supplied test case that creates an
  index where the initial temporary file is 249 one-megabyte blocks and
  the merged files become smaller. In the test, possible merge record
  sizes are 10, 18, and 26 bytes.
  
  rb://150 approved by Sunny Bains.  This addresses Issue #320.
  ------------------------------------------------------------------------
  r5703 | marko | 2009-08-27 03:25:00 -0400 (Thu, 27 Aug 2009) | 41 lines
  
  branches/zip: Replace the constant 3/8 ratio that controls the LRU_old
  size with the settable global variable innodb_old_blocks_pct. The
  minimum and maximum values are 5 and 95 per cent, respectively. The
  default is 100*3/8, in line with the old behavior.
  
  ut_time_ms(): New utility function, to return the current time in
  milliseconds. TODO: Is there a more efficient timestamp function, such
  as rdtsc divided by a power of two?
  
  buf_LRU_old_threshold_ms: New variable, corresponding to
  innodb_old_blocks_time. The value 0 is the default behaviour: no
  timeout before making blocks 'new'.
  
  bpage->accessed, bpage->LRU_position, buf_pool->ulint_clock: Remove.
  
  bpage->access_time: New field, replacing bpage->accessed. Protected by
  buf_pool_mutex instead of bpage->mutex. Updated when a page is created
  or accessed the first time in the buffer pool.
  
  buf_LRU_old_ratio, innobase_old_blocks_pct: New variables,
  corresponding to innodb_old_blocks_pct
  
  buf_LRU_old_ratio_update(), innobase_old_blocks_pct_update(): Update
  functions for buf_LRU_old_ratio, innobase_old_blocks_pct.
  
  buf_page_peek_if_too_old(): Compare ut_time_ms() to bpage->access_time
  if buf_LRU_old_threshold_ms && bpage->old.  Else observe
  buf_LRU_old_ratio and bpage->freed_page_clock.
  
  buf_pool_t: Add n_pages_made_young, n_pages_not_made_young,
  n_pages_made_young_old, n_pages_not_made_young, for statistics.
  
  buf_print(): Display buf_pool->n_pages_made_young,
  buf_pool->n_pages_not_made_young.  This function is only for crash
  diagnostics.
  
  buf_print_io(): Display buf_pool->LRU_old_len and quantities derived
  from buf_pool->n_pages_made_young, buf_pool->n_pages_not_made_young.
  This function is invoked by SHOW ENGINE INNODB STATUS.
  
  rb://129 approved by Heikki Tuuri.  This addresses Bug #45015.
  ------------------------------------------------------------------------
  r5702 | marko | 2009-08-27 03:03:15 -0400 (Thu, 27 Aug 2009) | 1 line
  
  branches/zip: Document also the files affected by r5698 in the ChangeLog.
  ------------------------------------------------------------------------
  r5701 | marko | 2009-08-27 03:01:42 -0400 (Thu, 27 Aug 2009) | 1 line
  
  branches/zip: Document r5698 in the ChangeLog.
  ------------------------------------------------------------------------
  r5698 | inaam | 2009-08-26 10:34:35 -0400 (Wed, 26 Aug 2009) | 13 lines
  
  branches/zip  bug#42885  rb://148
  
  The call to put IO threads to sleep was most probably meant for Windows
  only as the comment in buf0rea.c suggests. However it was enabled on
  all platforms. This patch restricts the sleep call to windows. This
  approach of not putting threads to sleep makes even more sense because
  now we have multiple threads working in the background and it probably
  is not a good idea to put all of them to sleep because a user thread
  wants to post a batch for readahead.
  
  Approved by: Marko
  
  
  ------------------------------------------------------------------------
  r5697 | vasil | 2009-08-26 09:44:40 -0400 (Wed, 26 Aug 2009) | 4 lines
  
  branches/zip:
  
  Fix typo.
  
  ------------------------------------------------------------------------
  r5696 | vasil | 2009-08-26 09:15:59 -0400 (Wed, 26 Aug 2009) | 14 lines
  
  branches/zip:
  
  Merge a change from MySQL:
  
   http://lists.mysql.com/commits/80832
  
   2968 Jonathan Perkin	2009-08-14
        Build fixes for Windows, AIX, HP/UX and Sun Studio11, from Timothy Smith.
        modified:
          CMakeLists.txt
          cmd-line-utils/readline/util.c
          storage/innodb_plugin/handler/i_s.cc
          storage/innodb_plugin/include/univ.i
  
  ------------------------------------------------------------------------
  r5695 | marko | 2009-08-26 09:14:59 -0400 (Wed, 26 Aug 2009) | 1 line
  
  branches/zip: UNIV_DEBUG_LOCK_VALIDATE: Move the definition to univ.i.
  ------------------------------------------------------------------------
  r5694 | marko | 2009-08-26 07:25:26 -0400 (Wed, 26 Aug 2009) | 2 lines
  
  branches/zip: buf_page_t: Clarify that bpage->list may contain garbage.
  This comment was provoked by Inaam.
  ------------------------------------------------------------------------
  r5687 | vasil | 2009-08-20 05:20:22 -0400 (Thu, 20 Aug 2009) | 8 lines
  
  branches/zip:
  
  ChangeLog:
  Follow the convention from the rest of the ChangeLog: for bugfixes from
  bugs.mysql.com only the bug number and title goes in the ChangeLog. Detailed
  explanation on what is the problem and how it was fixed is present in
  the bugs database.
  
  ------------------------------------------------------------------------
  r5686 | vasil | 2009-08-20 05:15:05 -0400 (Thu, 20 Aug 2009) | 4 lines
  
  branches/zip:
  
  White-space fixup.
  
  ------------------------------------------------------------------------
  r5685 | sunny | 2009-08-20 04:18:29 -0400 (Thu, 20 Aug 2009) | 2 lines
  
  branches/zip: Update the ChangeLog with r5684 change.
  
  ------------------------------------------------------------------------
  r5684 | sunny | 2009-08-20 04:05:30 -0400 (Thu, 20 Aug 2009) | 10 lines
  
  branches/zip: Fix bug# 46650:  Innodb assertion autoinc_lock == lock in lock_table_remove_low on INSERT SELECT
  
  We only store the autoinc locks that are granted in the transaction's autoinc
  lock vector. A transacton, that has been rolled back due to a deadlock because
  of an AUTOINC lock attempt, will not have added that lock to the vector. We
  need to check for that when we remove that lock.
  
  rb://145
  Approved by Marko.
  
  ------------------------------------------------------------------------
  r5681 | sunny | 2009-08-14 02:16:24 -0400 (Fri, 14 Aug 2009) | 3 lines
  
  branches/zip: When building HotBackup srv_use_sys_malloc is #ifdef out. We
  move access to the this variable within a !UNIV_HOTBACKUP block.
  
  ------------------------------------------------------------------------
  r5671 | marko | 2009-08-13 04:46:33 -0400 (Thu, 13 Aug 2009) | 5 lines
  
  branches/zip: ha_innobase::add_index(): Fix Bug #46557:
  after a successful operation, read innodb_table->flags from
  the newly created table object, not from the old one that was just freed.
  
  Approved by Sunny.
  ------------------------------------------------------------------------
  r5670 | marko | 2009-08-12 09:16:37 -0400 (Wed, 12 Aug 2009) | 2 lines
  
  branches/zip: trx_undo_rec_copy(): Add const qualifier to undo_rec.
  This is a non-functional change.
  ------------------------------------------------------------------------
  r5663 | marko | 2009-08-11 07:42:37 -0400 (Tue, 11 Aug 2009) | 2 lines
  
  branches/zip: trx_general_rollback_for_mysql(): Remove the redundant
  parameter partial. If savept==NULL, partial==FALSE.
  ------------------------------------------------------------------------
  r5662 | marko | 2009-08-11 05:54:16 -0400 (Tue, 11 Aug 2009) | 1 line
  
  branches/zip: Bump the version number to 1.0.5 after releasing 1.0.4.
  ------------------------------------------------------------------------
  r5642 | calvin | 2009-08-06 19:04:03 -0400 (Thu, 06 Aug 2009) | 2 lines
  
  branches/zip: remove duplicate "the" in comments.
  
  ------------------------------------------------------------------------
  r5639 | marko | 2009-08-06 06:39:34 -0400 (Thu, 06 Aug 2009) | 3 lines
  
  branches/zip: mem_heap_block_free(): If innodb_use_sys_malloc is set,
  do not tell Valgrind that the memory is free, to avoid
  a bogus warning in Valgrind's built-in free() hook.
  ------------------------------------------------------------------------
  r5636 | marko | 2009-08-05 08:27:30 -0400 (Wed, 05 Aug 2009) | 2 lines
  
  branches/zip: lock_rec_validate_page(): Add the parameter zip_size.
  This should help track down Mantis Issue #289.
  ------------------------------------------------------------------------
  r5635 | marko | 2009-08-05 07:06:55 -0400 (Wed, 05 Aug 2009) | 2 lines
  
  branches/zip: Replace <number> with NUMBER in some comments,
  to avoid problems with Doxygen XML output.
  ------------------------------------------------------------------------
  r5629 | marko | 2009-08-04 07:42:44 -0400 (Tue, 04 Aug 2009) | 1 line
  
  branches/zip: mysql-test: Pass MTR's internal checks.
  ------------------------------------------------------------------------
  r5626 | vasil | 2009-08-04 01:53:31 -0400 (Tue, 04 Aug 2009) | 4 lines
  
  branches/zip:
  
  Revert the dummy change from c5625.
  
  ------------------------------------------------------------------------
  r5625 | vasil | 2009-08-04 01:52:48 -0400 (Tue, 04 Aug 2009) | 32 lines
  
  branches/zip: Merge 5518:5622 from branches/5.1, resolving conflict in r5622
  (after resolving the conflict Makefile.am was not changed so I have made
  a dummy change so I can commit and thus record that branches/5.1 has been
  merged in branches/zip up to 5622):
  
    ------------------------------------------------------------------------
    r5622 | vasil | 2009-08-03 15:27:00 +0300 (Mon, 03 Aug 2009) | 20 lines
    Changed paths:
       M /branches/5.1/Makefile.am
    
    branches/5.1:
    
    Merge a change from MySQL:
    
      ------------------------------------------------------------
      revno: 2988
      committer: Satya B <satya.bn@sun.com>
      branch nick: mysql-5.1-bugteam
      timestamp: Wed 2009-07-01 11:06:05 +0530
      message:
        Fix build failure after applying Innodb snapshot 5.1-ss5282
        
        After applying Innodb snapshot 5.1-ss5282, build was broken
        because of missing header file. 
        
        Adding the header file to Makefile.am after informing the 
        innodb developers.
      modified:
        storage/innobase/Makefile.am
    
    ------------------------------------------------------------------------
  
  ------------------------------------------------------------------------
  r5614 | vasil | 2009-07-31 11:09:07 -0400 (Fri, 31 Jul 2009) | 6 lines
  
  branches/zip:
  
  Add fsp0types.h to the list of noinst_HEADERS
  
  Suggested by:	Sergey Vojtovich <svoj@sun.com>
  
  ------------------------------------------------------------------------
  r5539 | vasil | 2009-07-21 06:28:27 -0400 (Tue, 21 Jul 2009) | 4 lines
  
  branches/zip:
  
  Add a test program to check whether the PAUSE instruction is available.
  
  ------------------------------------------------------------------------
  r5537 | vasil | 2009-07-21 05:31:26 -0400 (Tue, 21 Jul 2009) | 5 lines
  
  branches/zip:
  
  Fixups in ChangeLog: sort filenames alphabetically and wrap to 78 chars per
  line.
  
  ------------------------------------------------------------------------
  r5527 | sunny | 2009-07-20 17:56:30 -0400 (Mon, 20 Jul 2009) | 2 lines
  
  branches/zip: For HotBackup builds we don't want to hide the symbols.
  
  ------------------------------------------------------------------------
  r5525 | calvin | 2009-07-20 13:14:30 -0400 (Mon, 20 Jul 2009) | 2 lines
  
  branches/zip: add ChangeLog entry for r5524.
  
  ------------------------------------------------------------------------
2009-09-28 17:34:23 +00:00

4260 lines
118 KiB
C

/*****************************************************************************
Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
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
*****************************************************************************/
/******************************************************************//**
@file fsp/fsp0fsp.c
File space management
Created 11/29/1995 Heikki Tuuri
***********************************************************************/
#include "fsp0fsp.h"
#ifdef UNIV_NONINL
#include "fsp0fsp.ic"
#endif
#include "buf0buf.h"
#include "fil0fil.h"
#include "mtr0log.h"
#include "ut0byte.h"
#include "page0page.h"
#include "page0zip.h"
#ifdef UNIV_HOTBACKUP
# include "fut0lst.h"
#else /* UNIV_HOTBACKUP */
# include "sync0sync.h"
# include "fut0fut.h"
# include "srv0srv.h"
# include "ibuf0ibuf.h"
# include "btr0btr.h"
# include "btr0sea.h"
# include "dict0boot.h"
# include "log0log.h"
#endif /* UNIV_HOTBACKUP */
#include "dict0mem.h"
#define FSP_HEADER_OFFSET FIL_PAGE_DATA /* Offset of the space header
within a file page */
/* The data structures in files are defined just as byte strings in C */
typedef byte fsp_header_t;
typedef byte xdes_t;
/* SPACE HEADER
============
File space header data structure: this data structure is contained in the
first page of a space. The space for this header is reserved in every extent
descriptor page, but used only in the first. */
/*-------------------------------------*/
#define FSP_SPACE_ID 0 /* space id */
#define FSP_NOT_USED 4 /* this field contained a value up to
which we know that the modifications
in the database have been flushed to
the file space; not used now */
#define FSP_SIZE 8 /* Current size of the space in
pages */
#define FSP_FREE_LIMIT 12 /* Minimum page number for which the
free list has not been initialized:
the pages >= this limit are, by
definition, free; note that in a
single-table tablespace where size
< 64 pages, this number is 64, i.e.,
we have initialized the space
about the first extent, but have not
physically allocted those pages to the
file */
#define FSP_SPACE_FLAGS 16 /* table->flags & ~DICT_TF_COMPACT */
#define FSP_FRAG_N_USED 20 /* number of used pages in the
FSP_FREE_FRAG list */
#define FSP_FREE 24 /* list of free extents */
#define FSP_FREE_FRAG (24 + FLST_BASE_NODE_SIZE)
/* list of partially free extents not
belonging to any segment */
#define FSP_FULL_FRAG (24 + 2 * FLST_BASE_NODE_SIZE)
/* list of full extents not belonging
to any segment */
#define FSP_SEG_ID (24 + 3 * FLST_BASE_NODE_SIZE)
/* 8 bytes which give the first unused
segment id */
#define FSP_SEG_INODES_FULL (32 + 3 * FLST_BASE_NODE_SIZE)
/* list of pages containing segment
headers, where all the segment inode
slots are reserved */
#define FSP_SEG_INODES_FREE (32 + 4 * FLST_BASE_NODE_SIZE)
/* list of pages containing segment
headers, where not all the segment
header slots are reserved */
/*-------------------------------------*/
/* File space header size */
#define FSP_HEADER_SIZE (32 + 5 * FLST_BASE_NODE_SIZE)
#define FSP_FREE_ADD 4 /* this many free extents are added
to the free list from above
FSP_FREE_LIMIT at a time */
/* FILE SEGMENT INODE
==================
Segment inode which is created for each segment in a tablespace. NOTE: in
purge we assume that a segment having only one currently used page can be
freed in a few steps, so that the freeing cannot fill the file buffer with
bufferfixed file pages. */
typedef byte fseg_inode_t;
#define FSEG_INODE_PAGE_NODE FSEG_PAGE_DATA
/* the list node for linking
segment inode pages */
#define FSEG_ARR_OFFSET (FSEG_PAGE_DATA + FLST_NODE_SIZE)
/*-------------------------------------*/
#define FSEG_ID 0 /* 8 bytes of segment id: if this is
ut_dulint_zero, it means that the
header is unused */
#define FSEG_NOT_FULL_N_USED 8
/* number of used segment pages in
the FSEG_NOT_FULL list */
#define FSEG_FREE 12
/* list of free extents of this
segment */
#define FSEG_NOT_FULL (12 + FLST_BASE_NODE_SIZE)
/* list of partially free extents */
#define FSEG_FULL (12 + 2 * FLST_BASE_NODE_SIZE)
/* list of full extents */
#define FSEG_MAGIC_N (12 + 3 * FLST_BASE_NODE_SIZE)
/* magic number used in debugging */
#define FSEG_FRAG_ARR (16 + 3 * FLST_BASE_NODE_SIZE)
/* array of individual pages
belonging to this segment in fsp
fragment extent lists */
#define FSEG_FRAG_ARR_N_SLOTS (FSP_EXTENT_SIZE / 2)
/* number of slots in the array for
the fragment pages */
#define FSEG_FRAG_SLOT_SIZE 4 /* a fragment page slot contains its
page number within space, FIL_NULL
means that the slot is not in use */
/*-------------------------------------*/
#define FSEG_INODE_SIZE \
(16 + 3 * FLST_BASE_NODE_SIZE \
+ FSEG_FRAG_ARR_N_SLOTS * FSEG_FRAG_SLOT_SIZE)
#define FSP_SEG_INODES_PER_PAGE(zip_size) \
(((zip_size ? zip_size : UNIV_PAGE_SIZE) \
- FSEG_ARR_OFFSET - 10) / FSEG_INODE_SIZE)
/* Number of segment inodes which fit on a
single page */
#define FSEG_MAGIC_N_VALUE 97937874
#define FSEG_FILLFACTOR 8 /* If this value is x, then if
the number of unused but reserved
pages in a segment is less than
reserved pages * 1/x, and there are
at least FSEG_FRAG_LIMIT used pages,
then we allow a new empty extent to
be added to the segment in
fseg_alloc_free_page. Otherwise, we
use unused pages of the segment. */
#define FSEG_FRAG_LIMIT FSEG_FRAG_ARR_N_SLOTS
/* If the segment has >= this many
used pages, it may be expanded by
allocating extents to the segment;
until that only individual fragment
pages are allocated from the space */
#define FSEG_FREE_LIST_LIMIT 40 /* If the reserved size of a segment
is at least this many extents, we
allow extents to be put to the free
list of the extent: at most
FSEG_FREE_LIST_MAX_LEN many */
#define FSEG_FREE_LIST_MAX_LEN 4
/* EXTENT DESCRIPTOR
=================
File extent descriptor data structure: contains bits to tell which pages in
the extent are free and which contain old tuple version to clean. */
/*-------------------------------------*/
#define XDES_ID 0 /* The identifier of the segment
to which this extent belongs */
#define XDES_FLST_NODE 8 /* The list node data structure
for the descriptors */
#define XDES_STATE (FLST_NODE_SIZE + 8)
/* contains state information
of the extent */
#define XDES_BITMAP (FLST_NODE_SIZE + 12)
/* Descriptor bitmap of the pages
in the extent */
/*-------------------------------------*/
#define XDES_BITS_PER_PAGE 2 /* How many bits are there per page */
#define XDES_FREE_BIT 0 /* Index of the bit which tells if
the page is free */
#define XDES_CLEAN_BIT 1 /* NOTE: currently not used!
Index of the bit which tells if
there are old versions of tuples
on the page */
/* States of a descriptor */
#define XDES_FREE 1 /* extent is in free list of space */
#define XDES_FREE_FRAG 2 /* extent is in free fragment list of
space */
#define XDES_FULL_FRAG 3 /* extent is in full fragment list of
space */
#define XDES_FSEG 4 /* extent belongs to a segment */
/* File extent data structure size in bytes. */
#define XDES_SIZE \
(XDES_BITMAP + UT_BITS_IN_BYTES(FSP_EXTENT_SIZE * XDES_BITS_PER_PAGE))
/* Offset of the descriptor array on a descriptor page */
#define XDES_ARR_OFFSET (FSP_HEADER_OFFSET + FSP_HEADER_SIZE)
/* Flag to indicate if we have printed the tablespace full error. */
static ibool fsp_tbs_full_error_printed = FALSE;
#ifndef UNIV_HOTBACKUP
/**********************************************************************//**
Returns an extent to the free list of a space. */
static
void
fsp_free_extent(
/*============*/
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint page, /*!< in: page offset in the extent */
mtr_t* mtr); /*!< in: mtr */
/**********************************************************************//**
Frees an extent of a segment to the space free list. */
static
void
fseg_free_extent(
/*=============*/
fseg_inode_t* seg_inode, /*!< in: segment inode */
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint page, /*!< in: page offset in the extent */
mtr_t* mtr); /*!< in: mtr handle */
/**********************************************************************//**
Calculates the number of pages reserved by a segment, and how
many pages are currently used.
@return number of reserved pages */
static
ulint
fseg_n_reserved_pages_low(
/*======================*/
fseg_inode_t* header, /*!< in: segment inode */
ulint* used, /*!< out: number of pages used (not
more than reserved) */
mtr_t* mtr); /*!< in: mtr handle */
/********************************************************************//**
Marks a page used. The page must reside within the extents of the given
segment. */
static
void
fseg_mark_page_used(
/*================*/
fseg_inode_t* seg_inode,/*!< in: segment inode */
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint page, /*!< in: page offset */
mtr_t* mtr); /*!< in: mtr */
/**********************************************************************//**
Returns the first extent descriptor for a segment. We think of the extent
lists of the segment catenated in the order FSEG_FULL -> FSEG_NOT_FULL
-> FSEG_FREE.
@return the first extent descriptor, or NULL if none */
static
xdes_t*
fseg_get_first_extent(
/*==================*/
fseg_inode_t* inode, /*!< in: segment inode */
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
mtr_t* mtr); /*!< in: mtr */
/**********************************************************************//**
Puts new extents to the free list if
there are free extents above the free limit. If an extent happens
to contain an extent descriptor page, the extent is put to
the FSP_FREE_FRAG list with the page marked as used. */
static
void
fsp_fill_free_list(
/*===============*/
ibool init_space, /*!< in: TRUE if this is a single-table
tablespace and we are only initing
the tablespace's first extent
descriptor page and ibuf bitmap page;
then we do not allocate more extents */
ulint space, /*!< in: space */
fsp_header_t* header, /*!< in: space header */
mtr_t* mtr); /*!< in: mtr */
/**********************************************************************//**
Allocates a single free page from a segment. This function implements
the intelligent allocation strategy which tries to minimize file space
fragmentation.
@return the allocated page number, FIL_NULL if no page could be allocated */
static
ulint
fseg_alloc_free_page_low(
/*=====================*/
ulint space, /*!< in: space */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
fseg_inode_t* seg_inode, /*!< in: segment inode */
ulint hint, /*!< in: hint of which page would be desirable */
byte direction, /*!< in: if the new page is needed because
of an index page split, and records are
inserted there in order, into which
direction they go alphabetically: FSP_DOWN,
FSP_UP, FSP_NO_DIR */
mtr_t* mtr); /*!< in: mtr handle */
#endif /* !UNIV_HOTBACKUP */
/**********************************************************************//**
Reads the file space size stored in the header page.
@return tablespace size stored in the space header */
UNIV_INTERN
ulint
fsp_get_size_low(
/*=============*/
page_t* page) /*!< in: header page (page 0 in the tablespace) */
{
return(mach_read_from_4(page + FSP_HEADER_OFFSET + FSP_SIZE));
}
#ifndef UNIV_HOTBACKUP
/**********************************************************************//**
Gets a pointer to the space header and x-locks its page.
@return pointer to the space header, page x-locked */
UNIV_INLINE
fsp_header_t*
fsp_get_space_header(
/*=================*/
ulint id, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
mtr_t* mtr) /*!< in: mtr */
{
buf_block_t* block;
fsp_header_t* header;
ut_ad(ut_is_2pow(zip_size));
ut_ad(zip_size <= UNIV_PAGE_SIZE);
ut_ad(!zip_size || zip_size >= PAGE_ZIP_MIN_SIZE);
ut_ad(id || !zip_size);
block = buf_page_get(id, zip_size, 0, RW_X_LATCH, mtr);
header = FSP_HEADER_OFFSET + buf_block_get_frame(block);
buf_block_dbg_add_level(block, SYNC_FSP_PAGE);
ut_ad(id == mach_read_from_4(FSP_SPACE_ID + header));
ut_ad(zip_size == dict_table_flags_to_zip_size(
mach_read_from_4(FSP_SPACE_FLAGS + header)));
return(header);
}
/**********************************************************************//**
Gets a descriptor bit of a page.
@return TRUE if free */
UNIV_INLINE
ibool
xdes_get_bit(
/*=========*/
xdes_t* descr, /*!< in: descriptor */
ulint bit, /*!< in: XDES_FREE_BIT or XDES_CLEAN_BIT */
ulint offset, /*!< in: page offset within extent:
0 ... FSP_EXTENT_SIZE - 1 */
mtr_t* mtr) /*!< in: mtr */
{
ulint index;
ulint byte_index;
ulint bit_index;
ut_ad(mtr_memo_contains_page(mtr, descr, MTR_MEMO_PAGE_X_FIX));
ut_ad((bit == XDES_FREE_BIT) || (bit == XDES_CLEAN_BIT));
ut_ad(offset < FSP_EXTENT_SIZE);
index = bit + XDES_BITS_PER_PAGE * offset;
byte_index = index / 8;
bit_index = index % 8;
return(ut_bit_get_nth(mtr_read_ulint(descr + XDES_BITMAP + byte_index,
MLOG_1BYTE, mtr),
bit_index));
}
/**********************************************************************//**
Sets a descriptor bit of a page. */
UNIV_INLINE
void
xdes_set_bit(
/*=========*/
xdes_t* descr, /*!< in: descriptor */
ulint bit, /*!< in: XDES_FREE_BIT or XDES_CLEAN_BIT */
ulint offset, /*!< in: page offset within extent:
0 ... FSP_EXTENT_SIZE - 1 */
ibool val, /*!< in: bit value */
mtr_t* mtr) /*!< in: mtr */
{
ulint index;
ulint byte_index;
ulint bit_index;
ulint descr_byte;
ut_ad(mtr_memo_contains_page(mtr, descr, MTR_MEMO_PAGE_X_FIX));
ut_ad((bit == XDES_FREE_BIT) || (bit == XDES_CLEAN_BIT));
ut_ad(offset < FSP_EXTENT_SIZE);
index = bit + XDES_BITS_PER_PAGE * offset;
byte_index = index / 8;
bit_index = index % 8;
descr_byte = mtr_read_ulint(descr + XDES_BITMAP + byte_index,
MLOG_1BYTE, mtr);
descr_byte = ut_bit_set_nth(descr_byte, bit_index, val);
mlog_write_ulint(descr + XDES_BITMAP + byte_index, descr_byte,
MLOG_1BYTE, mtr);
}
/**********************************************************************//**
Looks for a descriptor bit having the desired value. Starts from hint
and scans upward; at the end of the extent the search is wrapped to
the start of the extent.
@return bit index of the bit, ULINT_UNDEFINED if not found */
UNIV_INLINE
ulint
xdes_find_bit(
/*==========*/
xdes_t* descr, /*!< in: descriptor */
ulint bit, /*!< in: XDES_FREE_BIT or XDES_CLEAN_BIT */
ibool val, /*!< in: desired bit value */
ulint hint, /*!< in: hint of which bit position would be desirable */
mtr_t* mtr) /*!< in: mtr */
{
ulint i;
ut_ad(descr && mtr);
ut_ad(val <= TRUE);
ut_ad(hint < FSP_EXTENT_SIZE);
ut_ad(mtr_memo_contains_page(mtr, descr, MTR_MEMO_PAGE_X_FIX));
for (i = hint; i < FSP_EXTENT_SIZE; i++) {
if (val == xdes_get_bit(descr, bit, i, mtr)) {
return(i);
}
}
for (i = 0; i < hint; i++) {
if (val == xdes_get_bit(descr, bit, i, mtr)) {
return(i);
}
}
return(ULINT_UNDEFINED);
}
/**********************************************************************//**
Looks for a descriptor bit having the desired value. Scans the extent in
a direction opposite to xdes_find_bit.
@return bit index of the bit, ULINT_UNDEFINED if not found */
UNIV_INLINE
ulint
xdes_find_bit_downward(
/*===================*/
xdes_t* descr, /*!< in: descriptor */
ulint bit, /*!< in: XDES_FREE_BIT or XDES_CLEAN_BIT */
ibool val, /*!< in: desired bit value */
ulint hint, /*!< in: hint of which bit position would be desirable */
mtr_t* mtr) /*!< in: mtr */
{
ulint i;
ut_ad(descr && mtr);
ut_ad(val <= TRUE);
ut_ad(hint < FSP_EXTENT_SIZE);
ut_ad(mtr_memo_contains_page(mtr, descr, MTR_MEMO_PAGE_X_FIX));
for (i = hint + 1; i > 0; i--) {
if (val == xdes_get_bit(descr, bit, i - 1, mtr)) {
return(i - 1);
}
}
for (i = FSP_EXTENT_SIZE - 1; i > hint; i--) {
if (val == xdes_get_bit(descr, bit, i, mtr)) {
return(i);
}
}
return(ULINT_UNDEFINED);
}
/**********************************************************************//**
Returns the number of used pages in a descriptor.
@return number of pages used */
UNIV_INLINE
ulint
xdes_get_n_used(
/*============*/
xdes_t* descr, /*!< in: descriptor */
mtr_t* mtr) /*!< in: mtr */
{
ulint i;
ulint count = 0;
ut_ad(descr && mtr);
ut_ad(mtr_memo_contains_page(mtr, descr, MTR_MEMO_PAGE_X_FIX));
for (i = 0; i < FSP_EXTENT_SIZE; i++) {
if (FALSE == xdes_get_bit(descr, XDES_FREE_BIT, i, mtr)) {
count++;
}
}
return(count);
}
/**********************************************************************//**
Returns true if extent contains no used pages.
@return TRUE if totally free */
UNIV_INLINE
ibool
xdes_is_free(
/*=========*/
xdes_t* descr, /*!< in: descriptor */
mtr_t* mtr) /*!< in: mtr */
{
if (0 == xdes_get_n_used(descr, mtr)) {
return(TRUE);
}
return(FALSE);
}
/**********************************************************************//**
Returns true if extent contains no free pages.
@return TRUE if full */
UNIV_INLINE
ibool
xdes_is_full(
/*=========*/
xdes_t* descr, /*!< in: descriptor */
mtr_t* mtr) /*!< in: mtr */
{
if (FSP_EXTENT_SIZE == xdes_get_n_used(descr, mtr)) {
return(TRUE);
}
return(FALSE);
}
/**********************************************************************//**
Sets the state of an xdes. */
UNIV_INLINE
void
xdes_set_state(
/*===========*/
xdes_t* descr, /*!< in: descriptor */
ulint state, /*!< in: state to set */
mtr_t* mtr) /*!< in: mtr handle */
{
ut_ad(descr && mtr);
ut_ad(state >= XDES_FREE);
ut_ad(state <= XDES_FSEG);
ut_ad(mtr_memo_contains_page(mtr, descr, MTR_MEMO_PAGE_X_FIX));
mlog_write_ulint(descr + XDES_STATE, state, MLOG_4BYTES, mtr);
}
/**********************************************************************//**
Gets the state of an xdes.
@return state */
UNIV_INLINE
ulint
xdes_get_state(
/*===========*/
xdes_t* descr, /*!< in: descriptor */
mtr_t* mtr) /*!< in: mtr handle */
{
ulint state;
ut_ad(descr && mtr);
ut_ad(mtr_memo_contains_page(mtr, descr, MTR_MEMO_PAGE_X_FIX));
state = mtr_read_ulint(descr + XDES_STATE, MLOG_4BYTES, mtr);
ut_ad(state - 1 < XDES_FSEG);
return(state);
}
/**********************************************************************//**
Inits an extent descriptor to the free and clean state. */
UNIV_INLINE
void
xdes_init(
/*======*/
xdes_t* descr, /*!< in: descriptor */
mtr_t* mtr) /*!< in: mtr */
{
ulint i;
ut_ad(descr && mtr);
ut_ad(mtr_memo_contains_page(mtr, descr, MTR_MEMO_PAGE_X_FIX));
ut_ad((XDES_SIZE - XDES_BITMAP) % 4 == 0);
for (i = XDES_BITMAP; i < XDES_SIZE; i += 4) {
mlog_write_ulint(descr + i, 0xFFFFFFFFUL, MLOG_4BYTES, mtr);
}
xdes_set_state(descr, XDES_FREE, mtr);
}
/********************************************************************//**
Calculates the page where the descriptor of a page resides.
@return descriptor page offset */
UNIV_INLINE
ulint
xdes_calc_descriptor_page(
/*======================*/
ulint zip_size, /*!< in: compressed page size in bytes;
0 for uncompressed pages */
ulint offset) /*!< in: page offset */
{
#ifndef DOXYGEN /* Doxygen gets confused of these */
# if UNIV_PAGE_SIZE <= XDES_ARR_OFFSET \
+ (UNIV_PAGE_SIZE / FSP_EXTENT_SIZE) * XDES_SIZE
# error
# endif
# if PAGE_ZIP_MIN_SIZE <= XDES_ARR_OFFSET \
+ (PAGE_ZIP_MIN_SIZE / FSP_EXTENT_SIZE) * XDES_SIZE
# error
# endif
#endif /* !DOXYGEN */
ut_ad(ut_is_2pow(zip_size));
if (!zip_size) {
return(ut_2pow_round(offset, UNIV_PAGE_SIZE));
} else {
ut_ad(zip_size > XDES_ARR_OFFSET
+ (zip_size / FSP_EXTENT_SIZE) * XDES_SIZE);
return(ut_2pow_round(offset, zip_size));
}
}
/********************************************************************//**
Calculates the descriptor index within a descriptor page.
@return descriptor index */
UNIV_INLINE
ulint
xdes_calc_descriptor_index(
/*=======================*/
ulint zip_size, /*!< in: compressed page size in bytes;
0 for uncompressed pages */
ulint offset) /*!< in: page offset */
{
ut_ad(ut_is_2pow(zip_size));
if (!zip_size) {
return(ut_2pow_remainder(offset, UNIV_PAGE_SIZE)
/ FSP_EXTENT_SIZE);
} else {
return(ut_2pow_remainder(offset, zip_size) / FSP_EXTENT_SIZE);
}
}
/********************************************************************//**
Gets pointer to a the extent descriptor of a page. The page where the extent
descriptor resides is x-locked. If the page offset is equal to the free limit
of the space, adds new extents from above the free limit to the space free
list, if not free limit == space size. This adding is necessary to make the
descriptor defined, as they are uninitialized above the free limit.
@return pointer to the extent descriptor, NULL if the page does not
exist in the space or if the offset exceeds the free limit */
UNIV_INLINE
xdes_t*
xdes_get_descriptor_with_space_hdr(
/*===============================*/
fsp_header_t* sp_header,/*!< in: space header, x-latched */
ulint space, /*!< in: space id */
ulint offset, /*!< in: page offset;
if equal to the free limit,
we try to add new extents to
the space free list */
mtr_t* mtr) /*!< in: mtr handle */
{
ulint limit;
ulint size;
ulint zip_size;
ulint descr_page_no;
page_t* descr_page;
ut_ad(mtr);
ut_ad(mtr_memo_contains(mtr, fil_space_get_latch(space, NULL),
MTR_MEMO_X_LOCK));
ut_ad(mtr_memo_contains_page(mtr, sp_header, MTR_MEMO_PAGE_S_FIX)
|| mtr_memo_contains_page(mtr, sp_header, MTR_MEMO_PAGE_X_FIX));
ut_ad(page_offset(sp_header) == FSP_HEADER_OFFSET);
/* Read free limit and space size */
limit = mach_read_from_4(sp_header + FSP_FREE_LIMIT);
size = mach_read_from_4(sp_header + FSP_SIZE);
zip_size = dict_table_flags_to_zip_size(
mach_read_from_4(sp_header + FSP_SPACE_FLAGS));
/* If offset is >= size or > limit, return NULL */
if ((offset >= size) || (offset > limit)) {
return(NULL);
}
/* If offset is == limit, fill free list of the space. */
if (offset == limit) {
fsp_fill_free_list(FALSE, space, sp_header, mtr);
}
descr_page_no = xdes_calc_descriptor_page(zip_size, offset);
if (descr_page_no == 0) {
/* It is on the space header page */
descr_page = page_align(sp_header);
} else {
buf_block_t* block;
block = buf_page_get(space, zip_size, descr_page_no,
RW_X_LATCH, mtr);
buf_block_dbg_add_level(block, SYNC_FSP_PAGE);
descr_page = buf_block_get_frame(block);
}
return(descr_page + XDES_ARR_OFFSET
+ XDES_SIZE * xdes_calc_descriptor_index(zip_size, offset));
}
/********************************************************************//**
Gets pointer to a the extent descriptor of a page. The page where the
extent descriptor resides is x-locked. If the page offset is equal to
the free limit of the space, adds new extents from above the free limit
to the space free list, if not free limit == space size. This adding
is necessary to make the descriptor defined, as they are uninitialized
above the free limit.
@return pointer to the extent descriptor, NULL if the page does not
exist in the space or if the offset exceeds the free limit */
static
xdes_t*
xdes_get_descriptor(
/*================*/
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint offset, /*!< in: page offset; if equal to the free limit,
we try to add new extents to the space free list */
mtr_t* mtr) /*!< in: mtr handle */
{
buf_block_t* block;
fsp_header_t* sp_header;
block = buf_page_get(space, zip_size, 0, RW_X_LATCH, mtr);
buf_block_dbg_add_level(block, SYNC_FSP_PAGE);
sp_header = FSP_HEADER_OFFSET + buf_block_get_frame(block);
return(xdes_get_descriptor_with_space_hdr(sp_header, space, offset,
mtr));
}
/********************************************************************//**
Gets pointer to a the extent descriptor if the file address
of the descriptor list node is known. The page where the
extent descriptor resides is x-locked.
@return pointer to the extent descriptor */
UNIV_INLINE
xdes_t*
xdes_lst_get_descriptor(
/*====================*/
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
fil_addr_t lst_node,/*!< in: file address of the list node
contained in the descriptor */
mtr_t* mtr) /*!< in: mtr handle */
{
xdes_t* descr;
ut_ad(mtr);
ut_ad(mtr_memo_contains(mtr, fil_space_get_latch(space, NULL),
MTR_MEMO_X_LOCK));
descr = fut_get_ptr(space, zip_size, lst_node, RW_X_LATCH, mtr)
- XDES_FLST_NODE;
return(descr);
}
/********************************************************************//**
Returns page offset of the first page in extent described by a descriptor.
@return offset of the first page in extent */
UNIV_INLINE
ulint
xdes_get_offset(
/*============*/
xdes_t* descr) /*!< in: extent descriptor */
{
ut_ad(descr);
return(page_get_page_no(page_align(descr))
+ ((page_offset(descr) - XDES_ARR_OFFSET) / XDES_SIZE)
* FSP_EXTENT_SIZE);
}
#endif /* !UNIV_HOTBACKUP */
/***********************************************************//**
Inits a file page whose prior contents should be ignored. */
static
void
fsp_init_file_page_low(
/*===================*/
buf_block_t* block) /*!< in: pointer to a page */
{
page_t* page = buf_block_get_frame(block);
page_zip_des_t* page_zip= buf_block_get_page_zip(block);
#ifndef UNIV_HOTBACKUP
block->check_index_page_at_flush = FALSE;
#endif /* !UNIV_HOTBACKUP */
if (UNIV_LIKELY_NULL(page_zip)) {
memset(page, 0, UNIV_PAGE_SIZE);
memset(page_zip->data, 0, page_zip_get_size(page_zip));
mach_write_to_4(page + FIL_PAGE_OFFSET,
buf_block_get_page_no(block));
mach_write_to_4(page
+ FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID,
buf_block_get_space(block));
memcpy(page_zip->data + FIL_PAGE_OFFSET,
page + FIL_PAGE_OFFSET, 4);
memcpy(page_zip->data + FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID,
page + FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID, 4);
return;
}
#ifdef UNIV_BASIC_LOG_DEBUG
memset(page, 0xff, UNIV_PAGE_SIZE);
#endif
mach_write_to_4(page + FIL_PAGE_OFFSET, buf_block_get_page_no(block));
memset(page + FIL_PAGE_LSN, 0, 8);
mach_write_to_4(page + FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID,
buf_block_get_space(block));
memset(page + UNIV_PAGE_SIZE - FIL_PAGE_END_LSN_OLD_CHKSUM, 0, 8);
}
#ifndef UNIV_HOTBACKUP
/***********************************************************//**
Inits a file page whose prior contents should be ignored. */
static
void
fsp_init_file_page(
/*===============*/
buf_block_t* block, /*!< in: pointer to a page */
mtr_t* mtr) /*!< in: mtr */
{
fsp_init_file_page_low(block);
mlog_write_initial_log_record(buf_block_get_frame(block),
MLOG_INIT_FILE_PAGE, mtr);
}
#endif /* !UNIV_HOTBACKUP */
/***********************************************************//**
Parses a redo log record of a file page init.
@return end of log record or NULL */
UNIV_INTERN
byte*
fsp_parse_init_file_page(
/*=====================*/
byte* ptr, /*!< in: buffer */
byte* end_ptr __attribute__((unused)), /*!< in: buffer end */
buf_block_t* block) /*!< in: block or NULL */
{
ut_ad(ptr && end_ptr);
if (block) {
fsp_init_file_page_low(block);
}
return(ptr);
}
/**********************************************************************//**
Initializes the fsp system. */
UNIV_INTERN
void
fsp_init(void)
/*==========*/
{
/* Does nothing at the moment */
}
/**********************************************************************//**
Writes the space id and compressed page size to a tablespace header.
This function is used past the buffer pool when we in fil0fil.c create
a new single-table tablespace. */
UNIV_INTERN
void
fsp_header_init_fields(
/*===================*/
page_t* page, /*!< in/out: first page in the space */
ulint space_id, /*!< in: space id */
ulint flags) /*!< in: tablespace flags (FSP_SPACE_FLAGS):
0, or table->flags if newer than COMPACT */
{
/* The tablespace flags (FSP_SPACE_FLAGS) should be 0 for
ROW_FORMAT=COMPACT (table->flags == DICT_TF_COMPACT) and
ROW_FORMAT=REDUNDANT (table->flags == 0). For any other
format, the tablespace flags should equal table->flags. */
ut_a(flags != DICT_TF_COMPACT);
mach_write_to_4(FSP_HEADER_OFFSET + FSP_SPACE_ID + page,
space_id);
mach_write_to_4(FSP_HEADER_OFFSET + FSP_SPACE_FLAGS + page,
flags);
}
#ifndef UNIV_HOTBACKUP
/**********************************************************************//**
Initializes the space header of a new created space and creates also the
insert buffer tree root if space == 0. */
UNIV_INTERN
void
fsp_header_init(
/*============*/
ulint space, /*!< in: space id */
ulint size, /*!< in: current size in blocks */
mtr_t* mtr) /*!< in: mini-transaction handle */
{
fsp_header_t* header;
buf_block_t* block;
page_t* page;
ulint flags;
ulint zip_size;
ut_ad(mtr);
mtr_x_lock(fil_space_get_latch(space, &flags), mtr);
zip_size = dict_table_flags_to_zip_size(flags);
block = buf_page_create(space, 0, zip_size, mtr);
buf_page_get(space, zip_size, 0, RW_X_LATCH, mtr);
buf_block_dbg_add_level(block, SYNC_FSP_PAGE);
/* The prior contents of the file page should be ignored */
fsp_init_file_page(block, mtr);
page = buf_block_get_frame(block);
mlog_write_ulint(page + FIL_PAGE_TYPE, FIL_PAGE_TYPE_FSP_HDR,
MLOG_2BYTES, mtr);
header = FSP_HEADER_OFFSET + page;
mlog_write_ulint(header + FSP_SPACE_ID, space, MLOG_4BYTES, mtr);
mlog_write_ulint(header + FSP_NOT_USED, 0, MLOG_4BYTES, mtr);
mlog_write_ulint(header + FSP_SIZE, size, MLOG_4BYTES, mtr);
mlog_write_ulint(header + FSP_FREE_LIMIT, 0, MLOG_4BYTES, mtr);
mlog_write_ulint(header + FSP_SPACE_FLAGS, flags,
MLOG_4BYTES, mtr);
mlog_write_ulint(header + FSP_FRAG_N_USED, 0, MLOG_4BYTES, mtr);
flst_init(header + FSP_FREE, mtr);
flst_init(header + FSP_FREE_FRAG, mtr);
flst_init(header + FSP_FULL_FRAG, mtr);
flst_init(header + FSP_SEG_INODES_FULL, mtr);
flst_init(header + FSP_SEG_INODES_FREE, mtr);
mlog_write_dulint(header + FSP_SEG_ID, ut_dulint_create(0, 1), mtr);
if (space == 0) {
fsp_fill_free_list(FALSE, space, header, mtr);
btr_create(DICT_CLUSTERED | DICT_UNIVERSAL | DICT_IBUF,
0, 0, ut_dulint_add(DICT_IBUF_ID_MIN, space),
dict_ind_redundant, mtr);
} else {
fsp_fill_free_list(TRUE, space, header, mtr);
}
}
#endif /* !UNIV_HOTBACKUP */
/**********************************************************************//**
Reads the space id from the first page of a tablespace.
@return space id, ULINT UNDEFINED if error */
UNIV_INTERN
ulint
fsp_header_get_space_id(
/*====================*/
const page_t* page) /*!< in: first page of a tablespace */
{
ulint fsp_id;
ulint id;
fsp_id = mach_read_from_4(FSP_HEADER_OFFSET + page + FSP_SPACE_ID);
id = mach_read_from_4(page + FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID);
if (id != fsp_id) {
fprintf(stderr,
"InnoDB: Error: space id in fsp header %lu,"
" but in the page header %lu\n",
(ulong) fsp_id, (ulong) id);
return(ULINT_UNDEFINED);
}
return(id);
}
/**********************************************************************//**
Reads the space flags from the first page of a tablespace.
@return flags */
UNIV_INTERN
ulint
fsp_header_get_flags(
/*=================*/
const page_t* page) /*!< in: first page of a tablespace */
{
ut_ad(!page_offset(page));
return(mach_read_from_4(FSP_HEADER_OFFSET + FSP_SPACE_FLAGS + page));
}
/**********************************************************************//**
Reads the compressed page size from the first page of a tablespace.
@return compressed page size in bytes, or 0 if uncompressed */
UNIV_INTERN
ulint
fsp_header_get_zip_size(
/*====================*/
const page_t* page) /*!< in: first page of a tablespace */
{
ulint flags = fsp_header_get_flags(page);
return(dict_table_flags_to_zip_size(flags));
}
#ifndef UNIV_HOTBACKUP
/**********************************************************************//**
Increases the space size field of a space. */
UNIV_INTERN
void
fsp_header_inc_size(
/*================*/
ulint space, /*!< in: space id */
ulint size_inc,/*!< in: size increment in pages */
mtr_t* mtr) /*!< in: mini-transaction handle */
{
fsp_header_t* header;
ulint size;
ulint flags;
ut_ad(mtr);
mtr_x_lock(fil_space_get_latch(space, &flags), mtr);
header = fsp_get_space_header(space,
dict_table_flags_to_zip_size(flags),
mtr);
size = mtr_read_ulint(header + FSP_SIZE, MLOG_4BYTES, mtr);
mlog_write_ulint(header + FSP_SIZE, size + size_inc, MLOG_4BYTES,
mtr);
}
/**********************************************************************//**
Gets the current free limit of the system tablespace. The free limit
means the place of the first page which has never been put to the
free list for allocation. The space above that address is initialized
to zero. Sets also the global variable log_fsp_current_free_limit.
@return free limit in megabytes */
UNIV_INTERN
ulint
fsp_header_get_free_limit(void)
/*===========================*/
{
fsp_header_t* header;
ulint limit;
mtr_t mtr;
mtr_start(&mtr);
mtr_x_lock(fil_space_get_latch(0, NULL), &mtr);
header = fsp_get_space_header(0, 0, &mtr);
limit = mtr_read_ulint(header + FSP_FREE_LIMIT, MLOG_4BYTES, &mtr);
limit /= ((1024 * 1024) / UNIV_PAGE_SIZE);
log_fsp_current_free_limit_set_and_checkpoint(limit);
mtr_commit(&mtr);
return(limit);
}
/**********************************************************************//**
Gets the size of the system tablespace from the tablespace header. If
we do not have an auto-extending data file, this should be equal to
the size of the data files. If there is an auto-extending data file,
this can be smaller.
@return size in pages */
UNIV_INTERN
ulint
fsp_header_get_tablespace_size(void)
/*================================*/
{
fsp_header_t* header;
ulint size;
mtr_t mtr;
mtr_start(&mtr);
mtr_x_lock(fil_space_get_latch(0, NULL), &mtr);
header = fsp_get_space_header(0, 0, &mtr);
size = mtr_read_ulint(header + FSP_SIZE, MLOG_4BYTES, &mtr);
mtr_commit(&mtr);
return(size);
}
/***********************************************************************//**
Tries to extend a single-table tablespace so that a page would fit in the
data file.
@return TRUE if success */
static
ibool
fsp_try_extend_data_file_with_pages(
/*================================*/
ulint space, /*!< in: space */
ulint page_no, /*!< in: page number */
fsp_header_t* header, /*!< in: space header */
mtr_t* mtr) /*!< in: mtr */
{
ibool success;
ulint actual_size;
ulint size;
ut_a(space != 0);
size = mtr_read_ulint(header + FSP_SIZE, MLOG_4BYTES, mtr);
ut_a(page_no >= size);
success = fil_extend_space_to_desired_size(&actual_size, space,
page_no + 1);
/* actual_size now has the space size in pages; it may be less than
we wanted if we ran out of disk space */
mlog_write_ulint(header + FSP_SIZE, actual_size, MLOG_4BYTES, mtr);
return(success);
}
/***********************************************************************//**
Tries to extend the last data file of a tablespace if it is auto-extending.
@return FALSE if not auto-extending */
static
ibool
fsp_try_extend_data_file(
/*=====================*/
ulint* actual_increase,/*!< out: actual increase in pages, where
we measure the tablespace size from
what the header field says; it may be
the actual file size rounded down to
megabyte */
ulint space, /*!< in: space */
fsp_header_t* header, /*!< in: space header */
mtr_t* mtr) /*!< in: mtr */
{
ulint size;
ulint zip_size;
ulint new_size;
ulint old_size;
ulint size_increase;
ulint actual_size;
ibool success;
*actual_increase = 0;
if (space == 0 && !srv_auto_extend_last_data_file) {
/* We print the error message only once to avoid
spamming the error log. Note that we don't need
to reset the flag to FALSE as dealing with this
error requires server restart. */
if (fsp_tbs_full_error_printed == FALSE) {
fprintf(stderr,
"InnoDB: Error: Data file(s) ran"
" out of space.\n"
"Please add another data file or"
" use \'autoextend\' for the last"
" data file.\n");
fsp_tbs_full_error_printed = TRUE;
}
return(FALSE);
}
size = mtr_read_ulint(header + FSP_SIZE, MLOG_4BYTES, mtr);
zip_size = dict_table_flags_to_zip_size(
mach_read_from_4(header + FSP_SPACE_FLAGS));
old_size = size;
if (space == 0) {
if (!srv_last_file_size_max) {
size_increase = SRV_AUTO_EXTEND_INCREMENT;
} else {
if (srv_last_file_size_max
< srv_data_file_sizes[srv_n_data_files - 1]) {
fprintf(stderr,
"InnoDB: Error: Last data file size"
" is %lu, max size allowed %lu\n",
(ulong) srv_data_file_sizes[
srv_n_data_files - 1],
(ulong) srv_last_file_size_max);
}
size_increase = srv_last_file_size_max
- srv_data_file_sizes[srv_n_data_files - 1];
if (size_increase > SRV_AUTO_EXTEND_INCREMENT) {
size_increase = SRV_AUTO_EXTEND_INCREMENT;
}
}
} else {
/* We extend single-table tablespaces first one extent
at a time, but for bigger tablespaces more. It is not
enough to extend always by one extent, because some
extents are frag page extents. */
ulint extent_size; /*!< one megabyte, in pages */
if (!zip_size) {
extent_size = FSP_EXTENT_SIZE;
} else {
extent_size = FSP_EXTENT_SIZE
* UNIV_PAGE_SIZE / zip_size;
}
if (size < extent_size) {
/* Let us first extend the file to extent_size */
success = fsp_try_extend_data_file_with_pages(
space, extent_size - 1, header, mtr);
if (!success) {
new_size = mtr_read_ulint(header + FSP_SIZE,
MLOG_4BYTES, mtr);
*actual_increase = new_size - old_size;
return(FALSE);
}
size = extent_size;
}
if (size < 32 * extent_size) {
size_increase = extent_size;
} else {
/* Below in fsp_fill_free_list() we assume
that we add at most FSP_FREE_ADD extents at
a time */
size_increase = FSP_FREE_ADD * extent_size;
}
}
if (size_increase == 0) {
return(TRUE);
}
success = fil_extend_space_to_desired_size(&actual_size, space,
size + size_increase);
/* We ignore any fragments of a full megabyte when storing the size
to the space header */
if (!zip_size) {
new_size = ut_calc_align_down(actual_size,
(1024 * 1024) / UNIV_PAGE_SIZE);
} else {
new_size = ut_calc_align_down(actual_size,
(1024 * 1024) / zip_size);
}
mlog_write_ulint(header + FSP_SIZE, new_size, MLOG_4BYTES, mtr);
*actual_increase = new_size - old_size;
return(TRUE);
}
/**********************************************************************//**
Puts new extents to the free list if there are free extents above the free
limit. If an extent happens to contain an extent descriptor page, the extent
is put to the FSP_FREE_FRAG list with the page marked as used. */
static
void
fsp_fill_free_list(
/*===============*/
ibool init_space, /*!< in: TRUE if this is a single-table
tablespace and we are only initing
the tablespace's first extent
descriptor page and ibuf bitmap page;
then we do not allocate more extents */
ulint space, /*!< in: space */
fsp_header_t* header, /*!< in: space header */
mtr_t* mtr) /*!< in: mtr */
{
ulint limit;
ulint size;
ulint zip_size;
xdes_t* descr;
ulint count = 0;
ulint frag_n_used;
ulint actual_increase;
ulint i;
mtr_t ibuf_mtr;
ut_ad(header && mtr);
ut_ad(page_offset(header) == FSP_HEADER_OFFSET);
/* Check if we can fill free list from above the free list limit */
size = mtr_read_ulint(header + FSP_SIZE, MLOG_4BYTES, mtr);
limit = mtr_read_ulint(header + FSP_FREE_LIMIT, MLOG_4BYTES, mtr);
zip_size = dict_table_flags_to_zip_size(
mach_read_from_4(FSP_SPACE_FLAGS + header));
ut_a(ut_is_2pow(zip_size));
ut_a(zip_size <= UNIV_PAGE_SIZE);
ut_a(!zip_size || zip_size >= PAGE_ZIP_MIN_SIZE);
if (space == 0 && srv_auto_extend_last_data_file
&& size < limit + FSP_EXTENT_SIZE * FSP_FREE_ADD) {
/* Try to increase the last data file size */
fsp_try_extend_data_file(&actual_increase, space, header, mtr);
size = mtr_read_ulint(header + FSP_SIZE, MLOG_4BYTES, mtr);
}
if (space != 0 && !init_space
&& size < limit + FSP_EXTENT_SIZE * FSP_FREE_ADD) {
/* Try to increase the .ibd file size */
fsp_try_extend_data_file(&actual_increase, space, header, mtr);
size = mtr_read_ulint(header + FSP_SIZE, MLOG_4BYTES, mtr);
}
i = limit;
while ((init_space && i < 1)
|| ((i + FSP_EXTENT_SIZE <= size) && (count < FSP_FREE_ADD))) {
ibool init_xdes;
if (zip_size) {
init_xdes = ut_2pow_remainder(i, zip_size) == 0;
} else {
init_xdes = ut_2pow_remainder(i, UNIV_PAGE_SIZE) == 0;
}
mlog_write_ulint(header + FSP_FREE_LIMIT, i + FSP_EXTENT_SIZE,
MLOG_4BYTES, mtr);
/* Update the free limit info in the log system and make
a checkpoint */
if (space == 0) {
ut_a(!zip_size);
log_fsp_current_free_limit_set_and_checkpoint(
(i + FSP_EXTENT_SIZE)
/ ((1024 * 1024) / UNIV_PAGE_SIZE));
}
if (UNIV_UNLIKELY(init_xdes)) {
buf_block_t* block;
/* We are going to initialize a new descriptor page
and a new ibuf bitmap page: the prior contents of the
pages should be ignored. */
if (i > 0) {
block = buf_page_create(
space, i, zip_size, mtr);
buf_page_get(space, zip_size, i,
RW_X_LATCH, mtr);
buf_block_dbg_add_level(block,
SYNC_FSP_PAGE);
fsp_init_file_page(block, mtr);
mlog_write_ulint(buf_block_get_frame(block)
+ FIL_PAGE_TYPE,
FIL_PAGE_TYPE_XDES,
MLOG_2BYTES, mtr);
}
/* Initialize the ibuf bitmap page in a separate
mini-transaction because it is low in the latching
order, and we must be able to release its latch
before returning from the fsp routine */
mtr_start(&ibuf_mtr);
block = buf_page_create(space,
i + FSP_IBUF_BITMAP_OFFSET,
zip_size, &ibuf_mtr);
buf_page_get(space, zip_size,
i + FSP_IBUF_BITMAP_OFFSET,
RW_X_LATCH, &ibuf_mtr);
buf_block_dbg_add_level(block, SYNC_FSP_PAGE);
fsp_init_file_page(block, &ibuf_mtr);
ibuf_bitmap_page_init(block, &ibuf_mtr);
mtr_commit(&ibuf_mtr);
}
descr = xdes_get_descriptor_with_space_hdr(header, space, i,
mtr);
xdes_init(descr, mtr);
#if UNIV_PAGE_SIZE % FSP_EXTENT_SIZE
# error "UNIV_PAGE_SIZE % FSP_EXTENT_SIZE != 0"
#endif
#if PAGE_ZIP_MIN_SIZE % FSP_EXTENT_SIZE
# error "PAGE_ZIP_MIN_SIZE % FSP_EXTENT_SIZE != 0"
#endif
if (UNIV_UNLIKELY(init_xdes)) {
/* The first page in the extent is a descriptor page
and the second is an ibuf bitmap page: mark them
used */
xdes_set_bit(descr, XDES_FREE_BIT, 0, FALSE, mtr);
xdes_set_bit(descr, XDES_FREE_BIT,
FSP_IBUF_BITMAP_OFFSET, FALSE, mtr);
xdes_set_state(descr, XDES_FREE_FRAG, mtr);
flst_add_last(header + FSP_FREE_FRAG,
descr + XDES_FLST_NODE, mtr);
frag_n_used = mtr_read_ulint(header + FSP_FRAG_N_USED,
MLOG_4BYTES, mtr);
mlog_write_ulint(header + FSP_FRAG_N_USED,
frag_n_used + 2, MLOG_4BYTES, mtr);
} else {
flst_add_last(header + FSP_FREE,
descr + XDES_FLST_NODE, mtr);
count++;
}
i += FSP_EXTENT_SIZE;
}
}
/**********************************************************************//**
Allocates a new free extent.
@return extent descriptor, NULL if cannot be allocated */
static
xdes_t*
fsp_alloc_free_extent(
/*==================*/
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint hint, /*!< in: hint of which extent would be desirable: any
page offset in the extent goes; the hint must not
be > FSP_FREE_LIMIT */
mtr_t* mtr) /*!< in: mtr */
{
fsp_header_t* header;
fil_addr_t first;
xdes_t* descr;
ut_ad(mtr);
header = fsp_get_space_header(space, zip_size, mtr);
descr = xdes_get_descriptor_with_space_hdr(header, space, hint, mtr);
if (descr && (xdes_get_state(descr, mtr) == XDES_FREE)) {
/* Ok, we can take this extent */
} else {
/* Take the first extent in the free list */
first = flst_get_first(header + FSP_FREE, mtr);
if (fil_addr_is_null(first)) {
fsp_fill_free_list(FALSE, space, header, mtr);
first = flst_get_first(header + FSP_FREE, mtr);
}
if (fil_addr_is_null(first)) {
return(NULL); /* No free extents left */
}
descr = xdes_lst_get_descriptor(space, zip_size, first, mtr);
}
flst_remove(header + FSP_FREE, descr + XDES_FLST_NODE, mtr);
return(descr);
}
/**********************************************************************//**
Allocates a single free page from a space. The page is marked as used.
@return the page offset, FIL_NULL if no page could be allocated */
static
ulint
fsp_alloc_free_page(
/*================*/
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint hint, /*!< in: hint of which page would be desirable */
mtr_t* mtr) /*!< in: mtr handle */
{
fsp_header_t* header;
fil_addr_t first;
xdes_t* descr;
buf_block_t* block;
ulint free;
ulint frag_n_used;
ulint page_no;
ulint space_size;
ibool success;
ut_ad(mtr);
header = fsp_get_space_header(space, zip_size, mtr);
/* Get the hinted descriptor */
descr = xdes_get_descriptor_with_space_hdr(header, space, hint, mtr);
if (descr && (xdes_get_state(descr, mtr) == XDES_FREE_FRAG)) {
/* Ok, we can take this extent */
} else {
/* Else take the first extent in free_frag list */
first = flst_get_first(header + FSP_FREE_FRAG, mtr);
if (fil_addr_is_null(first)) {
/* There are no partially full fragments: allocate
a free extent and add it to the FREE_FRAG list. NOTE
that the allocation may have as a side-effect that an
extent containing a descriptor page is added to the
FREE_FRAG list. But we will allocate our page from the
the free extent anyway. */
descr = fsp_alloc_free_extent(space, zip_size,
hint, mtr);
if (descr == NULL) {
/* No free space left */
return(FIL_NULL);
}
xdes_set_state(descr, XDES_FREE_FRAG, mtr);
flst_add_last(header + FSP_FREE_FRAG,
descr + XDES_FLST_NODE, mtr);
} else {
descr = xdes_lst_get_descriptor(space, zip_size,
first, mtr);
}
/* Reset the hint */
hint = 0;
}
/* Now we have in descr an extent with at least one free page. Look
for a free page in the extent. */
free = xdes_find_bit(descr, XDES_FREE_BIT, TRUE,
hint % FSP_EXTENT_SIZE, mtr);
if (free == ULINT_UNDEFINED) {
ut_print_buf(stderr, ((byte*)descr) - 500, 1000);
putc('\n', stderr);
ut_error;
}
page_no = xdes_get_offset(descr) + free;
space_size = mtr_read_ulint(header + FSP_SIZE, MLOG_4BYTES, mtr);
if (space_size <= page_no) {
/* It must be that we are extending a single-table tablespace
whose size is still < 64 pages */
ut_a(space != 0);
if (page_no >= FSP_EXTENT_SIZE) {
fprintf(stderr,
"InnoDB: Error: trying to extend a"
" single-table tablespace %lu\n"
"InnoDB: by single page(s) though the"
" space size %lu. Page no %lu.\n",
(ulong) space, (ulong) space_size,
(ulong) page_no);
return(FIL_NULL);
}
success = fsp_try_extend_data_file_with_pages(space, page_no,
header, mtr);
if (!success) {
/* No disk space left */
return(FIL_NULL);
}
}
xdes_set_bit(descr, XDES_FREE_BIT, free, FALSE, mtr);
/* Update the FRAG_N_USED field */
frag_n_used = mtr_read_ulint(header + FSP_FRAG_N_USED, MLOG_4BYTES,
mtr);
frag_n_used++;
mlog_write_ulint(header + FSP_FRAG_N_USED, frag_n_used, MLOG_4BYTES,
mtr);
if (xdes_is_full(descr, mtr)) {
/* The fragment is full: move it to another list */
flst_remove(header + FSP_FREE_FRAG, descr + XDES_FLST_NODE,
mtr);
xdes_set_state(descr, XDES_FULL_FRAG, mtr);
flst_add_last(header + FSP_FULL_FRAG, descr + XDES_FLST_NODE,
mtr);
mlog_write_ulint(header + FSP_FRAG_N_USED,
frag_n_used - FSP_EXTENT_SIZE, MLOG_4BYTES,
mtr);
}
/* Initialize the allocated page to the buffer pool, so that it can
be obtained immediately with buf_page_get without need for a disk
read. */
buf_page_create(space, page_no, zip_size, mtr);
block = buf_page_get(space, zip_size, page_no, RW_X_LATCH, mtr);
buf_block_dbg_add_level(block, SYNC_FSP_PAGE);
/* Prior contents of the page should be ignored */
fsp_init_file_page(block, mtr);
return(page_no);
}
/**********************************************************************//**
Frees a single page of a space. The page is marked as free and clean. */
static
void
fsp_free_page(
/*==========*/
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint page, /*!< in: page offset */
mtr_t* mtr) /*!< in: mtr handle */
{
fsp_header_t* header;
xdes_t* descr;
ulint state;
ulint frag_n_used;
ut_ad(mtr);
/* fprintf(stderr, "Freeing page %lu in space %lu\n", page, space); */
header = fsp_get_space_header(space, zip_size, mtr);
descr = xdes_get_descriptor_with_space_hdr(header, space, page, mtr);
state = xdes_get_state(descr, mtr);
if (state != XDES_FREE_FRAG && state != XDES_FULL_FRAG) {
fprintf(stderr,
"InnoDB: Error: File space extent descriptor"
" of page %lu has state %lu\n",
(ulong) page,
(ulong) state);
fputs("InnoDB: Dump of descriptor: ", stderr);
ut_print_buf(stderr, ((byte*)descr) - 50, 200);
putc('\n', stderr);
if (state == XDES_FREE) {
/* We put here some fault tolerance: if the page
is already free, return without doing anything! */
return;
}
ut_error;
}
if (xdes_get_bit(descr, XDES_FREE_BIT, page % FSP_EXTENT_SIZE, mtr)) {
fprintf(stderr,
"InnoDB: Error: File space extent descriptor"
" of page %lu says it is free\n"
"InnoDB: Dump of descriptor: ", (ulong) page);
ut_print_buf(stderr, ((byte*)descr) - 50, 200);
putc('\n', stderr);
/* We put here some fault tolerance: if the page
is already free, return without doing anything! */
return;
}
xdes_set_bit(descr, XDES_FREE_BIT, page % FSP_EXTENT_SIZE, TRUE, mtr);
xdes_set_bit(descr, XDES_CLEAN_BIT, page % FSP_EXTENT_SIZE, TRUE, mtr);
frag_n_used = mtr_read_ulint(header + FSP_FRAG_N_USED, MLOG_4BYTES,
mtr);
if (state == XDES_FULL_FRAG) {
/* The fragment was full: move it to another list */
flst_remove(header + FSP_FULL_FRAG, descr + XDES_FLST_NODE,
mtr);
xdes_set_state(descr, XDES_FREE_FRAG, mtr);
flst_add_last(header + FSP_FREE_FRAG, descr + XDES_FLST_NODE,
mtr);
mlog_write_ulint(header + FSP_FRAG_N_USED,
frag_n_used + FSP_EXTENT_SIZE - 1,
MLOG_4BYTES, mtr);
} else {
ut_a(frag_n_used > 0);
mlog_write_ulint(header + FSP_FRAG_N_USED, frag_n_used - 1,
MLOG_4BYTES, mtr);
}
if (xdes_is_free(descr, mtr)) {
/* The extent has become free: move it to another list */
flst_remove(header + FSP_FREE_FRAG, descr + XDES_FLST_NODE,
mtr);
fsp_free_extent(space, zip_size, page, mtr);
}
}
/**********************************************************************//**
Returns an extent to the free list of a space. */
static
void
fsp_free_extent(
/*============*/
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint page, /*!< in: page offset in the extent */
mtr_t* mtr) /*!< in: mtr */
{
fsp_header_t* header;
xdes_t* descr;
ut_ad(mtr);
header = fsp_get_space_header(space, zip_size, mtr);
descr = xdes_get_descriptor_with_space_hdr(header, space, page, mtr);
if (xdes_get_state(descr, mtr) == XDES_FREE) {
ut_print_buf(stderr, (byte*)descr - 500, 1000);
putc('\n', stderr);
ut_error;
}
xdes_init(descr, mtr);
flst_add_last(header + FSP_FREE, descr + XDES_FLST_NODE, mtr);
}
/**********************************************************************//**
Returns the nth inode slot on an inode page.
@return segment inode */
UNIV_INLINE
fseg_inode_t*
fsp_seg_inode_page_get_nth_inode(
/*=============================*/
page_t* page, /*!< in: segment inode page */
ulint i, /*!< in: inode index on page */
ulint zip_size __attribute__((unused)),
/*!< in: compressed page size, or 0 */
mtr_t* mtr __attribute__((unused)))
/*!< in: mini-transaction handle */
{
ut_ad(i < FSP_SEG_INODES_PER_PAGE(zip_size));
ut_ad(mtr_memo_contains_page(mtr, page, MTR_MEMO_PAGE_X_FIX));
return(page + FSEG_ARR_OFFSET + FSEG_INODE_SIZE * i);
}
/**********************************************************************//**
Looks for a used segment inode on a segment inode page.
@return segment inode index, or ULINT_UNDEFINED if not found */
static
ulint
fsp_seg_inode_page_find_used(
/*=========================*/
page_t* page, /*!< in: segment inode page */
ulint zip_size,/*!< in: compressed page size, or 0 */
mtr_t* mtr) /*!< in: mini-transaction handle */
{
ulint i;
fseg_inode_t* inode;
for (i = 0; i < FSP_SEG_INODES_PER_PAGE(zip_size); i++) {
inode = fsp_seg_inode_page_get_nth_inode(
page, i, zip_size, mtr);
if (!ut_dulint_is_zero(mach_read_from_8(inode + FSEG_ID))) {
/* This is used */
return(i);
}
}
return(ULINT_UNDEFINED);
}
/**********************************************************************//**
Looks for an unused segment inode on a segment inode page.
@return segment inode index, or ULINT_UNDEFINED if not found */
static
ulint
fsp_seg_inode_page_find_free(
/*=========================*/
page_t* page, /*!< in: segment inode page */
ulint i, /*!< in: search forward starting from this index */
ulint zip_size,/*!< in: compressed page size, or 0 */
mtr_t* mtr) /*!< in: mini-transaction handle */
{
fseg_inode_t* inode;
for (; i < FSP_SEG_INODES_PER_PAGE(zip_size); i++) {
inode = fsp_seg_inode_page_get_nth_inode(
page, i, zip_size, mtr);
if (ut_dulint_is_zero(mach_read_from_8(inode + FSEG_ID))) {
/* This is unused */
return(i);
}
}
return(ULINT_UNDEFINED);
}
/**********************************************************************//**
Allocates a new file segment inode page.
@return TRUE if could be allocated */
static
ibool
fsp_alloc_seg_inode_page(
/*=====================*/
fsp_header_t* space_header, /*!< in: space header */
mtr_t* mtr) /*!< in: mini-transaction handle */
{
fseg_inode_t* inode;
buf_block_t* block;
page_t* page;
ulint page_no;
ulint space;
ulint zip_size;
ulint i;
ut_ad(page_offset(space_header) == FSP_HEADER_OFFSET);
space = page_get_space_id(page_align(space_header));
zip_size = dict_table_flags_to_zip_size(
mach_read_from_4(FSP_SPACE_FLAGS + space_header));
page_no = fsp_alloc_free_page(space, zip_size, 0, mtr);
if (page_no == FIL_NULL) {
return(FALSE);
}
block = buf_page_get(space, zip_size, page_no, RW_X_LATCH, mtr);
buf_block_dbg_add_level(block, SYNC_FSP_PAGE);
block->check_index_page_at_flush = FALSE;
page = buf_block_get_frame(block);
mlog_write_ulint(page + FIL_PAGE_TYPE, FIL_PAGE_INODE,
MLOG_2BYTES, mtr);
for (i = 0; i < FSP_SEG_INODES_PER_PAGE(zip_size); i++) {
inode = fsp_seg_inode_page_get_nth_inode(page, i,
zip_size, mtr);
mlog_write_dulint(inode + FSEG_ID, ut_dulint_zero, mtr);
}
flst_add_last(space_header + FSP_SEG_INODES_FREE,
page + FSEG_INODE_PAGE_NODE, mtr);
return(TRUE);
}
/**********************************************************************//**
Allocates a new file segment inode.
@return segment inode, or NULL if not enough space */
static
fseg_inode_t*
fsp_alloc_seg_inode(
/*================*/
fsp_header_t* space_header, /*!< in: space header */
mtr_t* mtr) /*!< in: mini-transaction handle */
{
ulint page_no;
buf_block_t* block;
page_t* page;
fseg_inode_t* inode;
ibool success;
ulint zip_size;
ulint n;
ut_ad(page_offset(space_header) == FSP_HEADER_OFFSET);
if (flst_get_len(space_header + FSP_SEG_INODES_FREE, mtr) == 0) {
/* Allocate a new segment inode page */
success = fsp_alloc_seg_inode_page(space_header, mtr);
if (!success) {
return(NULL);
}
}
page_no = flst_get_first(space_header + FSP_SEG_INODES_FREE, mtr).page;
zip_size = dict_table_flags_to_zip_size(
mach_read_from_4(FSP_SPACE_FLAGS + space_header));
block = buf_page_get(page_get_space_id(page_align(space_header)),
zip_size, page_no, RW_X_LATCH, mtr);
buf_block_dbg_add_level(block, SYNC_FSP_PAGE);
page = buf_block_get_frame(block);
n = fsp_seg_inode_page_find_free(page, 0, zip_size, mtr);
ut_a(n != ULINT_UNDEFINED);
inode = fsp_seg_inode_page_get_nth_inode(page, n, zip_size, mtr);
if (ULINT_UNDEFINED == fsp_seg_inode_page_find_free(page, n + 1,
zip_size, mtr)) {
/* There are no other unused headers left on the page: move it
to another list */
flst_remove(space_header + FSP_SEG_INODES_FREE,
page + FSEG_INODE_PAGE_NODE, mtr);
flst_add_last(space_header + FSP_SEG_INODES_FULL,
page + FSEG_INODE_PAGE_NODE, mtr);
}
return(inode);
}
/**********************************************************************//**
Frees a file segment inode. */
static
void
fsp_free_seg_inode(
/*===============*/
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
fseg_inode_t* inode, /*!< in: segment inode */
mtr_t* mtr) /*!< in: mini-transaction handle */
{
page_t* page;
fsp_header_t* space_header;
page = page_align(inode);
space_header = fsp_get_space_header(space, zip_size, mtr);
ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) == FSEG_MAGIC_N_VALUE);
if (ULINT_UNDEFINED
== fsp_seg_inode_page_find_free(page, 0, zip_size, mtr)) {
/* Move the page to another list */
flst_remove(space_header + FSP_SEG_INODES_FULL,
page + FSEG_INODE_PAGE_NODE, mtr);
flst_add_last(space_header + FSP_SEG_INODES_FREE,
page + FSEG_INODE_PAGE_NODE, mtr);
}
mlog_write_dulint(inode + FSEG_ID, ut_dulint_zero, mtr);
mlog_write_ulint(inode + FSEG_MAGIC_N, 0, MLOG_4BYTES, mtr);
if (ULINT_UNDEFINED
== fsp_seg_inode_page_find_used(page, zip_size, mtr)) {
/* There are no other used headers left on the page: free it */
flst_remove(space_header + FSP_SEG_INODES_FREE,
page + FSEG_INODE_PAGE_NODE, mtr);
fsp_free_page(space, zip_size, page_get_page_no(page), mtr);
}
}
/**********************************************************************//**
Returns the file segment inode, page x-latched.
@return segment inode, page x-latched */
static
fseg_inode_t*
fseg_inode_get(
/*===========*/
fseg_header_t* header, /*!< in: segment header */
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
mtr_t* mtr) /*!< in: mtr handle */
{
fil_addr_t inode_addr;
fseg_inode_t* inode;
inode_addr.page = mach_read_from_4(header + FSEG_HDR_PAGE_NO);
inode_addr.boffset = mach_read_from_2(header + FSEG_HDR_OFFSET);
ut_ad(space == mach_read_from_4(header + FSEG_HDR_SPACE));
inode = fut_get_ptr(space, zip_size, inode_addr, RW_X_LATCH, mtr);
ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) == FSEG_MAGIC_N_VALUE);
return(inode);
}
/**********************************************************************//**
Gets the page number from the nth fragment page slot.
@return page number, FIL_NULL if not in use */
UNIV_INLINE
ulint
fseg_get_nth_frag_page_no(
/*======================*/
fseg_inode_t* inode, /*!< in: segment inode */
ulint n, /*!< in: slot index */
mtr_t* mtr __attribute__((unused))) /*!< in: mtr handle */
{
ut_ad(inode && mtr);
ut_ad(n < FSEG_FRAG_ARR_N_SLOTS);
ut_ad(mtr_memo_contains_page(mtr, inode, MTR_MEMO_PAGE_X_FIX));
return(mach_read_from_4(inode + FSEG_FRAG_ARR
+ n * FSEG_FRAG_SLOT_SIZE));
}
/**********************************************************************//**
Sets the page number in the nth fragment page slot. */
UNIV_INLINE
void
fseg_set_nth_frag_page_no(
/*======================*/
fseg_inode_t* inode, /*!< in: segment inode */
ulint n, /*!< in: slot index */
ulint page_no,/*!< in: page number to set */
mtr_t* mtr) /*!< in: mtr handle */
{
ut_ad(inode && mtr);
ut_ad(n < FSEG_FRAG_ARR_N_SLOTS);
ut_ad(mtr_memo_contains_page(mtr, inode, MTR_MEMO_PAGE_X_FIX));
mlog_write_ulint(inode + FSEG_FRAG_ARR + n * FSEG_FRAG_SLOT_SIZE,
page_no, MLOG_4BYTES, mtr);
}
/**********************************************************************//**
Finds a fragment page slot which is free.
@return slot index; ULINT_UNDEFINED if none found */
static
ulint
fseg_find_free_frag_page_slot(
/*==========================*/
fseg_inode_t* inode, /*!< in: segment inode */
mtr_t* mtr) /*!< in: mtr handle */
{
ulint i;
ulint page_no;
ut_ad(inode && mtr);
for (i = 0; i < FSEG_FRAG_ARR_N_SLOTS; i++) {
page_no = fseg_get_nth_frag_page_no(inode, i, mtr);
if (page_no == FIL_NULL) {
return(i);
}
}
return(ULINT_UNDEFINED);
}
/**********************************************************************//**
Finds a fragment page slot which is used and last in the array.
@return slot index; ULINT_UNDEFINED if none found */
static
ulint
fseg_find_last_used_frag_page_slot(
/*===============================*/
fseg_inode_t* inode, /*!< in: segment inode */
mtr_t* mtr) /*!< in: mtr handle */
{
ulint i;
ulint page_no;
ut_ad(inode && mtr);
for (i = 0; i < FSEG_FRAG_ARR_N_SLOTS; i++) {
page_no = fseg_get_nth_frag_page_no(
inode, FSEG_FRAG_ARR_N_SLOTS - i - 1, mtr);
if (page_no != FIL_NULL) {
return(FSEG_FRAG_ARR_N_SLOTS - i - 1);
}
}
return(ULINT_UNDEFINED);
}
/**********************************************************************//**
Calculates reserved fragment page slots.
@return number of fragment pages */
static
ulint
fseg_get_n_frag_pages(
/*==================*/
fseg_inode_t* inode, /*!< in: segment inode */
mtr_t* mtr) /*!< in: mtr handle */
{
ulint i;
ulint count = 0;
ut_ad(inode && mtr);
for (i = 0; i < FSEG_FRAG_ARR_N_SLOTS; i++) {
if (FIL_NULL != fseg_get_nth_frag_page_no(inode, i, mtr)) {
count++;
}
}
return(count);
}
/**********************************************************************//**
Creates a new segment.
@return the block where the segment header is placed, x-latched, NULL
if could not create segment because of lack of space */
UNIV_INTERN
buf_block_t*
fseg_create_general(
/*================*/
ulint space, /*!< in: space id */
ulint page, /*!< in: page where the segment header is placed: if
this is != 0, the page must belong to another segment,
if this is 0, a new page will be allocated and it
will belong to the created segment */
ulint byte_offset, /*!< in: byte offset of the created segment header
on the page */
ibool has_done_reservation, /*!< in: TRUE if the caller has already
done the reservation for the pages with
fsp_reserve_free_extents (at least 2 extents: one for
the inode and the other for the segment) then there is
no need to do the check for this individual
operation */
mtr_t* mtr) /*!< in: mtr */
{
ulint flags;
ulint zip_size;
fsp_header_t* space_header;
fseg_inode_t* inode;
dulint seg_id;
buf_block_t* block = 0; /* remove warning */
fseg_header_t* header = 0; /* remove warning */
rw_lock_t* latch;
ibool success;
ulint n_reserved;
ulint i;
ut_ad(mtr);
ut_ad(byte_offset + FSEG_HEADER_SIZE
<= UNIV_PAGE_SIZE - FIL_PAGE_DATA_END);
latch = fil_space_get_latch(space, &flags);
zip_size = dict_table_flags_to_zip_size(flags);
if (page != 0) {
block = buf_page_get(space, zip_size, page, RW_X_LATCH, mtr);
header = byte_offset + buf_block_get_frame(block);
}
ut_ad(!mutex_own(&kernel_mutex)
|| mtr_memo_contains(mtr, latch, MTR_MEMO_X_LOCK));
mtr_x_lock(latch, mtr);
if (rw_lock_get_x_lock_count(latch) == 1) {
/* This thread did not own the latch before this call: free
excess pages from the insert buffer free list */
if (space == IBUF_SPACE_ID) {
ibuf_free_excess_pages();
}
}
if (!has_done_reservation) {
success = fsp_reserve_free_extents(&n_reserved, space, 2,
FSP_NORMAL, mtr);
if (!success) {
return(NULL);
}
}
space_header = fsp_get_space_header(space, zip_size, mtr);
inode = fsp_alloc_seg_inode(space_header, mtr);
if (inode == NULL) {
goto funct_exit;
}
/* Read the next segment id from space header and increment the
value in space header */
seg_id = mtr_read_dulint(space_header + FSP_SEG_ID, mtr);
mlog_write_dulint(space_header + FSP_SEG_ID, ut_dulint_add(seg_id, 1),
mtr);
mlog_write_dulint(inode + FSEG_ID, seg_id, mtr);
mlog_write_ulint(inode + FSEG_NOT_FULL_N_USED, 0, MLOG_4BYTES, mtr);
flst_init(inode + FSEG_FREE, mtr);
flst_init(inode + FSEG_NOT_FULL, mtr);
flst_init(inode + FSEG_FULL, mtr);
mlog_write_ulint(inode + FSEG_MAGIC_N, FSEG_MAGIC_N_VALUE,
MLOG_4BYTES, mtr);
for (i = 0; i < FSEG_FRAG_ARR_N_SLOTS; i++) {
fseg_set_nth_frag_page_no(inode, i, FIL_NULL, mtr);
}
if (page == 0) {
page = fseg_alloc_free_page_low(space, zip_size,
inode, 0, FSP_UP, mtr);
if (page == FIL_NULL) {
fsp_free_seg_inode(space, zip_size, inode, mtr);
goto funct_exit;
}
block = buf_page_get(space, zip_size, page, RW_X_LATCH, mtr);
header = byte_offset + buf_block_get_frame(block);
mlog_write_ulint(header - byte_offset + FIL_PAGE_TYPE,
FIL_PAGE_TYPE_SYS, MLOG_2BYTES, mtr);
}
mlog_write_ulint(header + FSEG_HDR_OFFSET,
page_offset(inode), MLOG_2BYTES, mtr);
mlog_write_ulint(header + FSEG_HDR_PAGE_NO,
page_get_page_no(page_align(inode)),
MLOG_4BYTES, mtr);
mlog_write_ulint(header + FSEG_HDR_SPACE, space, MLOG_4BYTES, mtr);
funct_exit:
if (!has_done_reservation) {
fil_space_release_free_extents(space, n_reserved);
}
return(block);
}
/**********************************************************************//**
Creates a new segment.
@return the block where the segment header is placed, x-latched, NULL
if could not create segment because of lack of space */
UNIV_INTERN
buf_block_t*
fseg_create(
/*========*/
ulint space, /*!< in: space id */
ulint page, /*!< in: page where the segment header is placed: if
this is != 0, the page must belong to another segment,
if this is 0, a new page will be allocated and it
will belong to the created segment */
ulint byte_offset, /*!< in: byte offset of the created segment header
on the page */
mtr_t* mtr) /*!< in: mtr */
{
return(fseg_create_general(space, page, byte_offset, FALSE, mtr));
}
/**********************************************************************//**
Calculates the number of pages reserved by a segment, and how many pages are
currently used.
@return number of reserved pages */
static
ulint
fseg_n_reserved_pages_low(
/*======================*/
fseg_inode_t* inode, /*!< in: segment inode */
ulint* used, /*!< out: number of pages used (not
more than reserved) */
mtr_t* mtr) /*!< in: mtr handle */
{
ulint ret;
ut_ad(inode && used && mtr);
ut_ad(mtr_memo_contains_page(mtr, inode, MTR_MEMO_PAGE_X_FIX));
*used = mtr_read_ulint(inode + FSEG_NOT_FULL_N_USED, MLOG_4BYTES, mtr)
+ FSP_EXTENT_SIZE * flst_get_len(inode + FSEG_FULL, mtr)
+ fseg_get_n_frag_pages(inode, mtr);
ret = fseg_get_n_frag_pages(inode, mtr)
+ FSP_EXTENT_SIZE * flst_get_len(inode + FSEG_FREE, mtr)
+ FSP_EXTENT_SIZE * flst_get_len(inode + FSEG_NOT_FULL, mtr)
+ FSP_EXTENT_SIZE * flst_get_len(inode + FSEG_FULL, mtr);
return(ret);
}
/**********************************************************************//**
Calculates the number of pages reserved by a segment, and how many pages are
currently used.
@return number of reserved pages */
UNIV_INTERN
ulint
fseg_n_reserved_pages(
/*==================*/
fseg_header_t* header, /*!< in: segment header */
ulint* used, /*!< out: number of pages used (<= reserved) */
mtr_t* mtr) /*!< in: mtr handle */
{
ulint ret;
fseg_inode_t* inode;
ulint space;
ulint flags;
ulint zip_size;
rw_lock_t* latch;
space = page_get_space_id(page_align(header));
latch = fil_space_get_latch(space, &flags);
zip_size = dict_table_flags_to_zip_size(flags);
ut_ad(!mutex_own(&kernel_mutex)
|| mtr_memo_contains(mtr, latch, MTR_MEMO_X_LOCK));
mtr_x_lock(latch, mtr);
inode = fseg_inode_get(header, space, zip_size, mtr);
ret = fseg_n_reserved_pages_low(inode, used, mtr);
return(ret);
}
/*********************************************************************//**
Tries to fill the free list of a segment with consecutive free extents.
This happens if the segment is big enough to allow extents in the free list,
the free list is empty, and the extents can be allocated consecutively from
the hint onward. */
static
void
fseg_fill_free_list(
/*================*/
fseg_inode_t* inode, /*!< in: segment inode */
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint hint, /*!< in: hint which extent would be good as
the first extent */
mtr_t* mtr) /*!< in: mtr */
{
xdes_t* descr;
ulint i;
dulint seg_id;
ulint reserved;
ulint used;
ut_ad(inode && mtr);
ut_ad(!((page_offset(inode) - FSEG_ARR_OFFSET) % FSEG_INODE_SIZE));
reserved = fseg_n_reserved_pages_low(inode, &used, mtr);
if (reserved < FSEG_FREE_LIST_LIMIT * FSP_EXTENT_SIZE) {
/* The segment is too small to allow extents in free list */
return;
}
if (flst_get_len(inode + FSEG_FREE, mtr) > 0) {
/* Free list is not empty */
return;
}
for (i = 0; i < FSEG_FREE_LIST_MAX_LEN; i++) {
descr = xdes_get_descriptor(space, zip_size, hint, mtr);
if ((descr == NULL)
|| (XDES_FREE != xdes_get_state(descr, mtr))) {
/* We cannot allocate the desired extent: stop */
return;
}
descr = fsp_alloc_free_extent(space, zip_size, hint, mtr);
xdes_set_state(descr, XDES_FSEG, mtr);
seg_id = mtr_read_dulint(inode + FSEG_ID, mtr);
mlog_write_dulint(descr + XDES_ID, seg_id, mtr);
flst_add_last(inode + FSEG_FREE, descr + XDES_FLST_NODE, mtr);
hint += FSP_EXTENT_SIZE;
}
}
/*********************************************************************//**
Allocates a free extent for the segment: looks first in the free list of the
segment, then tries to allocate from the space free list. NOTE that the extent
returned still resides in the segment free list, it is not yet taken off it!
@return allocated extent, still placed in the segment free list, NULL
if could not be allocated */
static
xdes_t*
fseg_alloc_free_extent(
/*===================*/
fseg_inode_t* inode, /*!< in: segment inode */
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
mtr_t* mtr) /*!< in: mtr */
{
xdes_t* descr;
dulint seg_id;
fil_addr_t first;
ut_ad(!((page_offset(inode) - FSEG_ARR_OFFSET) % FSEG_INODE_SIZE));
if (flst_get_len(inode + FSEG_FREE, mtr) > 0) {
/* Segment free list is not empty, allocate from it */
first = flst_get_first(inode + FSEG_FREE, mtr);
descr = xdes_lst_get_descriptor(space, zip_size, first, mtr);
} else {
/* Segment free list was empty, allocate from space */
descr = fsp_alloc_free_extent(space, zip_size, 0, mtr);
if (descr == NULL) {
return(NULL);
}
seg_id = mtr_read_dulint(inode + FSEG_ID, mtr);
xdes_set_state(descr, XDES_FSEG, mtr);
mlog_write_dulint(descr + XDES_ID, seg_id, mtr);
flst_add_last(inode + FSEG_FREE, descr + XDES_FLST_NODE, mtr);
/* Try to fill the segment free list */
fseg_fill_free_list(inode, space, zip_size,
xdes_get_offset(descr) + FSP_EXTENT_SIZE,
mtr);
}
return(descr);
}
/**********************************************************************//**
Allocates a single free page from a segment. This function implements
the intelligent allocation strategy which tries to minimize file space
fragmentation.
@return the allocated page number, FIL_NULL if no page could be allocated */
static
ulint
fseg_alloc_free_page_low(
/*=====================*/
ulint space, /*!< in: space */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
fseg_inode_t* seg_inode, /*!< in: segment inode */
ulint hint, /*!< in: hint of which page would be desirable */
byte direction, /*!< in: if the new page is needed because
of an index page split, and records are
inserted there in order, into which
direction they go alphabetically: FSP_DOWN,
FSP_UP, FSP_NO_DIR */
mtr_t* mtr) /*!< in: mtr handle */
{
fsp_header_t* space_header;
ulint space_size;
dulint seg_id;
ulint used;
ulint reserved;
xdes_t* descr; /*!< extent of the hinted page */
ulint ret_page; /*!< the allocated page offset, FIL_NULL
if could not be allocated */
xdes_t* ret_descr; /*!< the extent of the allocated page */
ibool frag_page_allocated = FALSE;
ibool success;
ulint n;
ut_ad(mtr);
ut_ad((direction >= FSP_UP) && (direction <= FSP_NO_DIR));
ut_ad(mach_read_from_4(seg_inode + FSEG_MAGIC_N)
== FSEG_MAGIC_N_VALUE);
ut_ad(!((page_offset(seg_inode) - FSEG_ARR_OFFSET) % FSEG_INODE_SIZE));
seg_id = mtr_read_dulint(seg_inode + FSEG_ID, mtr);
ut_ad(!ut_dulint_is_zero(seg_id));
reserved = fseg_n_reserved_pages_low(seg_inode, &used, mtr);
space_header = fsp_get_space_header(space, zip_size, mtr);
descr = xdes_get_descriptor_with_space_hdr(space_header, space,
hint, mtr);
if (descr == NULL) {
/* Hint outside space or too high above free limit: reset
hint */
hint = 0;
descr = xdes_get_descriptor(space, zip_size, hint, mtr);
}
/* In the big if-else below we look for ret_page and ret_descr */
/*-------------------------------------------------------------*/
if ((xdes_get_state(descr, mtr) == XDES_FSEG)
&& (0 == ut_dulint_cmp(mtr_read_dulint(descr + XDES_ID,
mtr), seg_id))
&& (xdes_get_bit(descr, XDES_FREE_BIT,
hint % FSP_EXTENT_SIZE, mtr) == TRUE)) {
/* 1. We can take the hinted page
=================================*/
ret_descr = descr;
ret_page = hint;
/*-----------------------------------------------------------*/
} else if ((xdes_get_state(descr, mtr) == XDES_FREE)
&& ((reserved - used) < reserved / FSEG_FILLFACTOR)
&& (used >= FSEG_FRAG_LIMIT)) {
/* 2. We allocate the free extent from space and can take
=========================================================
the hinted page
===============*/
ret_descr = fsp_alloc_free_extent(space, zip_size, hint, mtr);
ut_a(ret_descr == descr);
xdes_set_state(ret_descr, XDES_FSEG, mtr);
mlog_write_dulint(ret_descr + XDES_ID, seg_id, mtr);
flst_add_last(seg_inode + FSEG_FREE,
ret_descr + XDES_FLST_NODE, mtr);
/* Try to fill the segment free list */
fseg_fill_free_list(seg_inode, space, zip_size,
hint + FSP_EXTENT_SIZE, mtr);
ret_page = hint;
/*-----------------------------------------------------------*/
} else if ((direction != FSP_NO_DIR)
&& ((reserved - used) < reserved / FSEG_FILLFACTOR)
&& (used >= FSEG_FRAG_LIMIT)
&& (!!(ret_descr
= fseg_alloc_free_extent(seg_inode,
space, zip_size, mtr)))) {
/* 3. We take any free extent (which was already assigned above
===============================================================
in the if-condition to ret_descr) and take the lowest or
========================================================
highest page in it, depending on the direction
==============================================*/
ret_page = xdes_get_offset(ret_descr);
if (direction == FSP_DOWN) {
ret_page += FSP_EXTENT_SIZE - 1;
}
/*-----------------------------------------------------------*/
} else if ((xdes_get_state(descr, mtr) == XDES_FSEG)
&& (0 == ut_dulint_cmp(mtr_read_dulint(descr + XDES_ID,
mtr), seg_id))
&& (!xdes_is_full(descr, mtr))) {
/* 4. We can take the page from the same extent as the
======================================================
hinted page (and the extent already belongs to the
==================================================
segment)
========*/
ret_descr = descr;
ret_page = xdes_get_offset(ret_descr)
+ xdes_find_bit(ret_descr, XDES_FREE_BIT, TRUE,
hint % FSP_EXTENT_SIZE, mtr);
/*-----------------------------------------------------------*/
} else if (reserved - used > 0) {
/* 5. We take any unused page from the segment
==============================================*/
fil_addr_t first;
if (flst_get_len(seg_inode + FSEG_NOT_FULL, mtr) > 0) {
first = flst_get_first(seg_inode + FSEG_NOT_FULL,
mtr);
} else if (flst_get_len(seg_inode + FSEG_FREE, mtr) > 0) {
first = flst_get_first(seg_inode + FSEG_FREE, mtr);
} else {
ut_error;
return(FIL_NULL);
}
ret_descr = xdes_lst_get_descriptor(space, zip_size,
first, mtr);
ret_page = xdes_get_offset(ret_descr)
+ xdes_find_bit(ret_descr, XDES_FREE_BIT, TRUE,
0, mtr);
/*-----------------------------------------------------------*/
} else if (used < FSEG_FRAG_LIMIT) {
/* 6. We allocate an individual page from the space
===================================================*/
ret_page = fsp_alloc_free_page(space, zip_size, hint, mtr);
ret_descr = NULL;
frag_page_allocated = TRUE;
if (ret_page != FIL_NULL) {
/* Put the page in the fragment page array of the
segment */
n = fseg_find_free_frag_page_slot(seg_inode, mtr);
ut_a(n != FIL_NULL);
fseg_set_nth_frag_page_no(seg_inode, n, ret_page,
mtr);
}
/*-----------------------------------------------------------*/
} else {
/* 7. We allocate a new extent and take its first page
======================================================*/
ret_descr = fseg_alloc_free_extent(seg_inode,
space, zip_size, mtr);
if (ret_descr == NULL) {
ret_page = FIL_NULL;
} else {
ret_page = xdes_get_offset(ret_descr);
}
}
if (ret_page == FIL_NULL) {
/* Page could not be allocated */
return(FIL_NULL);
}
if (space != 0) {
space_size = fil_space_get_size(space);
if (space_size <= ret_page) {
/* It must be that we are extending a single-table
tablespace whose size is still < 64 pages */
if (ret_page >= FSP_EXTENT_SIZE) {
fprintf(stderr,
"InnoDB: Error (2): trying to extend"
" a single-table tablespace %lu\n"
"InnoDB: by single page(s) though"
" the space size %lu. Page no %lu.\n",
(ulong) space, (ulong) space_size,
(ulong) ret_page);
return(FIL_NULL);
}
success = fsp_try_extend_data_file_with_pages(
space, ret_page, space_header, mtr);
if (!success) {
/* No disk space left */
return(FIL_NULL);
}
}
}
if (!frag_page_allocated) {
/* Initialize the allocated page to buffer pool, so that it
can be obtained immediately with buf_page_get without need
for a disk read */
buf_block_t* block;
ulint zip_size = dict_table_flags_to_zip_size(
mach_read_from_4(FSP_SPACE_FLAGS + space_header));
block = buf_page_create(space, ret_page, zip_size, mtr);
buf_block_dbg_add_level(block, SYNC_FSP_PAGE);
if (UNIV_UNLIKELY(block != buf_page_get(space, zip_size,
ret_page, RW_X_LATCH,
mtr))) {
ut_error;
}
/* The prior contents of the page should be ignored */
fsp_init_file_page(block, mtr);
/* At this point we know the extent and the page offset.
The extent is still in the appropriate list (FSEG_NOT_FULL
or FSEG_FREE), and the page is not yet marked as used. */
ut_ad(xdes_get_descriptor(space, zip_size, ret_page, mtr)
== ret_descr);
ut_ad(xdes_get_bit(ret_descr, XDES_FREE_BIT,
ret_page % FSP_EXTENT_SIZE, mtr) == TRUE);
fseg_mark_page_used(seg_inode, space, zip_size, ret_page, mtr);
}
buf_reset_check_index_page_at_flush(space, ret_page);
return(ret_page);
}
/**********************************************************************//**
Allocates a single free page from a segment. This function implements
the intelligent allocation strategy which tries to minimize file space
fragmentation.
@return allocated page offset, FIL_NULL if no page could be allocated */
UNIV_INTERN
ulint
fseg_alloc_free_page_general(
/*=========================*/
fseg_header_t* seg_header,/*!< in: segment header */
ulint hint, /*!< in: hint of which page would be desirable */
byte direction,/*!< in: if the new page is needed because
of an index page split, and records are
inserted there in order, into which
direction they go alphabetically: FSP_DOWN,
FSP_UP, FSP_NO_DIR */
ibool has_done_reservation, /*!< in: TRUE if the caller has
already done the reservation for the page
with fsp_reserve_free_extents, then there
is no need to do the check for this individual
page */
mtr_t* mtr) /*!< in: mtr handle */
{
fseg_inode_t* inode;
ulint space;
ulint flags;
ulint zip_size;
rw_lock_t* latch;
ibool success;
ulint page_no;
ulint n_reserved;
space = page_get_space_id(page_align(seg_header));
latch = fil_space_get_latch(space, &flags);
zip_size = dict_table_flags_to_zip_size(flags);
ut_ad(!mutex_own(&kernel_mutex)
|| mtr_memo_contains(mtr, latch, MTR_MEMO_X_LOCK));
mtr_x_lock(latch, mtr);
if (rw_lock_get_x_lock_count(latch) == 1) {
/* This thread did not own the latch before this call: free
excess pages from the insert buffer free list */
if (space == IBUF_SPACE_ID) {
ibuf_free_excess_pages();
}
}
inode = fseg_inode_get(seg_header, space, zip_size, mtr);
if (!has_done_reservation) {
success = fsp_reserve_free_extents(&n_reserved, space, 2,
FSP_NORMAL, mtr);
if (!success) {
return(FIL_NULL);
}
}
page_no = fseg_alloc_free_page_low(space, zip_size,
inode, hint, direction, mtr);
if (!has_done_reservation) {
fil_space_release_free_extents(space, n_reserved);
}
return(page_no);
}
/**********************************************************************//**
Allocates a single free page from a segment. This function implements
the intelligent allocation strategy which tries to minimize file space
fragmentation.
@return allocated page offset, FIL_NULL if no page could be allocated */
UNIV_INTERN
ulint
fseg_alloc_free_page(
/*=================*/
fseg_header_t* seg_header,/*!< in: segment header */
ulint hint, /*!< in: hint of which page would be desirable */
byte direction,/*!< in: if the new page is needed because
of an index page split, and records are
inserted there in order, into which
direction they go alphabetically: FSP_DOWN,
FSP_UP, FSP_NO_DIR */
mtr_t* mtr) /*!< in: mtr handle */
{
return(fseg_alloc_free_page_general(seg_header, hint, direction,
FALSE, mtr));
}
/**********************************************************************//**
Checks that we have at least 2 frag pages free in the first extent of a
single-table tablespace, and they are also physically initialized to the data
file. That is we have already extended the data file so that those pages are
inside the data file. If not, this function extends the tablespace with
pages.
@return TRUE if there were >= 3 free pages, or we were able to extend */
static
ibool
fsp_reserve_free_pages(
/*===================*/
ulint space, /*!< in: space id, must be != 0 */
fsp_header_t* space_header, /*!< in: header of that space,
x-latched */
ulint size, /*!< in: size of the tablespace in pages,
must be < FSP_EXTENT_SIZE / 2 */
mtr_t* mtr) /*!< in: mtr */
{
xdes_t* descr;
ulint n_used;
ut_a(space != 0);
ut_a(size < FSP_EXTENT_SIZE / 2);
descr = xdes_get_descriptor_with_space_hdr(space_header, space, 0,
mtr);
n_used = xdes_get_n_used(descr, mtr);
ut_a(n_used <= size);
if (size >= n_used + 2) {
return(TRUE);
}
return(fsp_try_extend_data_file_with_pages(space, n_used + 1,
space_header, mtr));
}
/**********************************************************************//**
Reserves free pages from a tablespace. All mini-transactions which may
use several pages from the tablespace should call this function beforehand
and reserve enough free extents so that they certainly will be able
to do their operation, like a B-tree page split, fully. Reservations
must be released with function fil_space_release_free_extents!
The alloc_type below has the following meaning: FSP_NORMAL means an
operation which will probably result in more space usage, like an
insert in a B-tree; FSP_UNDO means allocation to undo logs: if we are
deleting rows, then this allocation will in the long run result in
less space usage (after a purge); FSP_CLEANING means allocation done
in a physical record delete (like in a purge) or other cleaning operation
which will result in less space usage in the long run. We prefer the latter
two types of allocation: when space is scarce, FSP_NORMAL allocations
will not succeed, but the latter two allocations will succeed, if possible.
The purpose is to avoid dead end where the database is full but the
user cannot free any space because these freeing operations temporarily
reserve some space.
Single-table tablespaces whose size is < 32 pages are a special case. In this
function we would liberally reserve several 64 page extents for every page
split or merge in a B-tree. But we do not want to waste disk space if the table
only occupies < 32 pages. That is why we apply different rules in that special
case, just ensuring that there are 3 free pages available.
@return TRUE if we were able to make the reservation */
UNIV_INTERN
ibool
fsp_reserve_free_extents(
/*=====================*/
ulint* n_reserved,/*!< out: number of extents actually reserved; if we
return TRUE and the tablespace size is < 64 pages,
then this can be 0, otherwise it is n_ext */
ulint space, /*!< in: space id */
ulint n_ext, /*!< in: number of extents to reserve */
ulint alloc_type,/*!< in: FSP_NORMAL, FSP_UNDO, or FSP_CLEANING */
mtr_t* mtr) /*!< in: mtr */
{
fsp_header_t* space_header;
rw_lock_t* latch;
ulint n_free_list_ext;
ulint free_limit;
ulint size;
ulint flags;
ulint zip_size;
ulint n_free;
ulint n_free_up;
ulint reserve;
ibool success;
ulint n_pages_added;
ut_ad(mtr);
*n_reserved = n_ext;
latch = fil_space_get_latch(space, &flags);
zip_size = dict_table_flags_to_zip_size(flags);
ut_ad(!mutex_own(&kernel_mutex)
|| mtr_memo_contains(mtr, latch, MTR_MEMO_X_LOCK));
mtr_x_lock(latch, mtr);
space_header = fsp_get_space_header(space, zip_size, mtr);
try_again:
size = mtr_read_ulint(space_header + FSP_SIZE, MLOG_4BYTES, mtr);
if (size < FSP_EXTENT_SIZE / 2) {
/* Use different rules for small single-table tablespaces */
*n_reserved = 0;
return(fsp_reserve_free_pages(space, space_header, size, mtr));
}
n_free_list_ext = flst_get_len(space_header + FSP_FREE, mtr);
free_limit = mtr_read_ulint(space_header + FSP_FREE_LIMIT,
MLOG_4BYTES, mtr);
/* Below we play safe when counting free extents above the free limit:
some of them will contain extent descriptor pages, and therefore
will not be free extents */
n_free_up = (size - free_limit) / FSP_EXTENT_SIZE;
if (n_free_up > 0) {
n_free_up--;
if (!zip_size) {
n_free_up -= n_free_up
/ (UNIV_PAGE_SIZE / FSP_EXTENT_SIZE);
} else {
n_free_up -= n_free_up
/ (zip_size / FSP_EXTENT_SIZE);
}
}
n_free = n_free_list_ext + n_free_up;
if (alloc_type == FSP_NORMAL) {
/* We reserve 1 extent + 0.5 % of the space size to undo logs
and 1 extent + 0.5 % to cleaning operations; NOTE: this source
code is duplicated in the function below! */
reserve = 2 + ((size / FSP_EXTENT_SIZE) * 2) / 200;
if (n_free <= reserve + n_ext) {
goto try_to_extend;
}
} else if (alloc_type == FSP_UNDO) {
/* We reserve 0.5 % of the space size to cleaning operations */
reserve = 1 + ((size / FSP_EXTENT_SIZE) * 1) / 200;
if (n_free <= reserve + n_ext) {
goto try_to_extend;
}
} else {
ut_a(alloc_type == FSP_CLEANING);
}
success = fil_space_reserve_free_extents(space, n_free, n_ext);
if (success) {
return(TRUE);
}
try_to_extend:
success = fsp_try_extend_data_file(&n_pages_added, space,
space_header, mtr);
if (success && n_pages_added > 0) {
goto try_again;
}
return(FALSE);
}
/**********************************************************************//**
This function should be used to get information on how much we still
will be able to insert new data to the database without running out the
tablespace. Only free extents are taken into account and we also subtract
the safety margin required by the above function fsp_reserve_free_extents.
@return available space in kB */
UNIV_INTERN
ullint
fsp_get_available_space_in_free_extents(
/*====================================*/
ulint space) /*!< in: space id */
{
fsp_header_t* space_header;
ulint n_free_list_ext;
ulint free_limit;
ulint size;
ulint flags;
ulint zip_size;
ulint n_free;
ulint n_free_up;
ulint reserve;
rw_lock_t* latch;
mtr_t mtr;
ut_ad(!mutex_own(&kernel_mutex));
mtr_start(&mtr);
latch = fil_space_get_latch(space, &flags);
zip_size = dict_table_flags_to_zip_size(flags);
mtr_x_lock(latch, &mtr);
space_header = fsp_get_space_header(space, zip_size, &mtr);
size = mtr_read_ulint(space_header + FSP_SIZE, MLOG_4BYTES, &mtr);
n_free_list_ext = flst_get_len(space_header + FSP_FREE, &mtr);
free_limit = mtr_read_ulint(space_header + FSP_FREE_LIMIT,
MLOG_4BYTES, &mtr);
mtr_commit(&mtr);
if (size < FSP_EXTENT_SIZE) {
ut_a(space != 0); /* This must be a single-table
tablespace */
return(0); /* TODO: count free frag pages and
return a value based on that */
}
/* Below we play safe when counting free extents above the free limit:
some of them will contain extent descriptor pages, and therefore
will not be free extents */
n_free_up = (size - free_limit) / FSP_EXTENT_SIZE;
if (n_free_up > 0) {
n_free_up--;
if (!zip_size) {
n_free_up -= n_free_up
/ (UNIV_PAGE_SIZE / FSP_EXTENT_SIZE);
} else {
n_free_up -= n_free_up
/ (zip_size / FSP_EXTENT_SIZE);
}
}
n_free = n_free_list_ext + n_free_up;
/* We reserve 1 extent + 0.5 % of the space size to undo logs
and 1 extent + 0.5 % to cleaning operations; NOTE: this source
code is duplicated in the function above! */
reserve = 2 + ((size / FSP_EXTENT_SIZE) * 2) / 200;
if (reserve > n_free) {
return(0);
}
if (!zip_size) {
return((ullint) (n_free - reserve)
* FSP_EXTENT_SIZE
* (UNIV_PAGE_SIZE / 1024));
} else {
return((ullint) (n_free - reserve)
* FSP_EXTENT_SIZE
* (zip_size / 1024));
}
}
/********************************************************************//**
Marks a page used. The page must reside within the extents of the given
segment. */
static
void
fseg_mark_page_used(
/*================*/
fseg_inode_t* seg_inode,/*!< in: segment inode */
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint page, /*!< in: page offset */
mtr_t* mtr) /*!< in: mtr */
{
xdes_t* descr;
ulint not_full_n_used;
ut_ad(seg_inode && mtr);
ut_ad(!((page_offset(seg_inode) - FSEG_ARR_OFFSET) % FSEG_INODE_SIZE));
descr = xdes_get_descriptor(space, zip_size, page, mtr);
ut_ad(mtr_read_ulint(seg_inode + FSEG_ID, MLOG_4BYTES, mtr)
== mtr_read_ulint(descr + XDES_ID, MLOG_4BYTES, mtr));
if (xdes_is_free(descr, mtr)) {
/* We move the extent from the free list to the
NOT_FULL list */
flst_remove(seg_inode + FSEG_FREE, descr + XDES_FLST_NODE,
mtr);
flst_add_last(seg_inode + FSEG_NOT_FULL,
descr + XDES_FLST_NODE, mtr);
}
ut_ad(xdes_get_bit(descr, XDES_FREE_BIT, page % FSP_EXTENT_SIZE, mtr)
== TRUE);
/* We mark the page as used */
xdes_set_bit(descr, XDES_FREE_BIT, page % FSP_EXTENT_SIZE, FALSE, mtr);
not_full_n_used = mtr_read_ulint(seg_inode + FSEG_NOT_FULL_N_USED,
MLOG_4BYTES, mtr);
not_full_n_used++;
mlog_write_ulint(seg_inode + FSEG_NOT_FULL_N_USED, not_full_n_used,
MLOG_4BYTES, mtr);
if (xdes_is_full(descr, mtr)) {
/* We move the extent from the NOT_FULL list to the
FULL list */
flst_remove(seg_inode + FSEG_NOT_FULL,
descr + XDES_FLST_NODE, mtr);
flst_add_last(seg_inode + FSEG_FULL,
descr + XDES_FLST_NODE, mtr);
mlog_write_ulint(seg_inode + FSEG_NOT_FULL_N_USED,
not_full_n_used - FSP_EXTENT_SIZE,
MLOG_4BYTES, mtr);
}
}
/**********************************************************************//**
Frees a single page of a segment. */
static
void
fseg_free_page_low(
/*===============*/
fseg_inode_t* seg_inode, /*!< in: segment inode */
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint page, /*!< in: page offset */
mtr_t* mtr) /*!< in: mtr handle */
{
xdes_t* descr;
ulint not_full_n_used;
ulint state;
dulint descr_id;
dulint seg_id;
ulint i;
ut_ad(seg_inode && mtr);
ut_ad(mach_read_from_4(seg_inode + FSEG_MAGIC_N)
== FSEG_MAGIC_N_VALUE);
ut_ad(!((page_offset(seg_inode) - FSEG_ARR_OFFSET) % FSEG_INODE_SIZE));
/* Drop search system page hash index if the page is found in
the pool and is hashed */
btr_search_drop_page_hash_when_freed(space, zip_size, page);
descr = xdes_get_descriptor(space, zip_size, page, mtr);
ut_a(descr);
if (xdes_get_bit(descr, XDES_FREE_BIT, page % FSP_EXTENT_SIZE, mtr)) {
fputs("InnoDB: Dump of the tablespace extent descriptor: ",
stderr);
ut_print_buf(stderr, descr, 40);
fprintf(stderr, "\n"
"InnoDB: Serious error! InnoDB is trying to"
" free page %lu\n"
"InnoDB: though it is already marked as free"
" in the tablespace!\n"
"InnoDB: The tablespace free space info is corrupt.\n"
"InnoDB: You may need to dump your"
" InnoDB tables and recreate the whole\n"
"InnoDB: database!\n", (ulong) page);
crash:
fputs("InnoDB: Please refer to\n"
"InnoDB: " REFMAN "forcing-recovery.html\n"
"InnoDB: about forcing recovery.\n", stderr);
ut_error;
}
state = xdes_get_state(descr, mtr);
if (state != XDES_FSEG) {
/* The page is in the fragment pages of the segment */
for (i = 0;; i++) {
if (fseg_get_nth_frag_page_no(seg_inode, i, mtr)
== page) {
fseg_set_nth_frag_page_no(seg_inode, i,
FIL_NULL, mtr);
break;
}
}
fsp_free_page(space, zip_size, page, mtr);
return;
}
/* If we get here, the page is in some extent of the segment */
descr_id = mtr_read_dulint(descr + XDES_ID, mtr);
seg_id = mtr_read_dulint(seg_inode + FSEG_ID, mtr);
#if 0
fprintf(stderr,
"InnoDB: InnoDB is freeing space %lu page %lu,\n"
"InnoDB: which belongs to descr seg %lu %lu\n"
"InnoDB: segment %lu %lu.\n",
(ulong) space, (ulong) page,
(ulong) ut_dulint_get_high(descr_id),
(ulong) ut_dulint_get_low(descr_id),
(ulong) ut_dulint_get_high(seg_id),
(ulong) ut_dulint_get_low(seg_id));
#endif /* 0 */
if (0 != ut_dulint_cmp(descr_id, seg_id)) {
fputs("InnoDB: Dump of the tablespace extent descriptor: ",
stderr);
ut_print_buf(stderr, descr, 40);
fputs("\nInnoDB: Dump of the segment inode: ", stderr);
ut_print_buf(stderr, seg_inode, 40);
putc('\n', stderr);
fprintf(stderr,
"InnoDB: Serious error: InnoDB is trying to"
" free space %lu page %lu,\n"
"InnoDB: which does not belong to"
" segment %lu %lu but belongs\n"
"InnoDB: to segment %lu %lu.\n",
(ulong) space, (ulong) page,
(ulong) ut_dulint_get_high(descr_id),
(ulong) ut_dulint_get_low(descr_id),
(ulong) ut_dulint_get_high(seg_id),
(ulong) ut_dulint_get_low(seg_id));
goto crash;
}
not_full_n_used = mtr_read_ulint(seg_inode + FSEG_NOT_FULL_N_USED,
MLOG_4BYTES, mtr);
if (xdes_is_full(descr, mtr)) {
/* The fragment is full: move it to another list */
flst_remove(seg_inode + FSEG_FULL,
descr + XDES_FLST_NODE, mtr);
flst_add_last(seg_inode + FSEG_NOT_FULL,
descr + XDES_FLST_NODE, mtr);
mlog_write_ulint(seg_inode + FSEG_NOT_FULL_N_USED,
not_full_n_used + FSP_EXTENT_SIZE - 1,
MLOG_4BYTES, mtr);
} else {
ut_a(not_full_n_used > 0);
mlog_write_ulint(seg_inode + FSEG_NOT_FULL_N_USED,
not_full_n_used - 1, MLOG_4BYTES, mtr);
}
xdes_set_bit(descr, XDES_FREE_BIT, page % FSP_EXTENT_SIZE, TRUE, mtr);
xdes_set_bit(descr, XDES_CLEAN_BIT, page % FSP_EXTENT_SIZE, TRUE, mtr);
if (xdes_is_free(descr, mtr)) {
/* The extent has become free: free it to space */
flst_remove(seg_inode + FSEG_NOT_FULL,
descr + XDES_FLST_NODE, mtr);
fsp_free_extent(space, zip_size, page, mtr);
}
}
/**********************************************************************//**
Frees a single page of a segment. */
UNIV_INTERN
void
fseg_free_page(
/*===========*/
fseg_header_t* seg_header, /*!< in: segment header */
ulint space, /*!< in: space id */
ulint page, /*!< in: page offset */
mtr_t* mtr) /*!< in: mtr handle */
{
ulint flags;
ulint zip_size;
fseg_inode_t* seg_inode;
rw_lock_t* latch;
latch = fil_space_get_latch(space, &flags);
zip_size = dict_table_flags_to_zip_size(flags);
ut_ad(!mutex_own(&kernel_mutex)
|| mtr_memo_contains(mtr, latch, MTR_MEMO_X_LOCK));
mtr_x_lock(latch, mtr);
seg_inode = fseg_inode_get(seg_header, space, zip_size, mtr);
fseg_free_page_low(seg_inode, space, zip_size, page, mtr);
#ifdef UNIV_DEBUG_FILE_ACCESSES
buf_page_set_file_page_was_freed(space, page);
#endif
}
/**********************************************************************//**
Frees an extent of a segment to the space free list. */
static
void
fseg_free_extent(
/*=============*/
fseg_inode_t* seg_inode, /*!< in: segment inode */
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint page, /*!< in: a page in the extent */
mtr_t* mtr) /*!< in: mtr handle */
{
ulint first_page_in_extent;
xdes_t* descr;
ulint not_full_n_used;
ulint descr_n_used;
ulint i;
ut_ad(seg_inode && mtr);
descr = xdes_get_descriptor(space, zip_size, page, mtr);
ut_a(xdes_get_state(descr, mtr) == XDES_FSEG);
ut_a(0 == ut_dulint_cmp(mtr_read_dulint(descr + XDES_ID, mtr),
mtr_read_dulint(seg_inode + FSEG_ID, mtr)));
first_page_in_extent = page - (page % FSP_EXTENT_SIZE);
for (i = 0; i < FSP_EXTENT_SIZE; i++) {
if (FALSE == xdes_get_bit(descr, XDES_FREE_BIT, i, mtr)) {
/* Drop search system page hash index if the page is
found in the pool and is hashed */
btr_search_drop_page_hash_when_freed(
space, zip_size, first_page_in_extent + i);
}
}
if (xdes_is_full(descr, mtr)) {
flst_remove(seg_inode + FSEG_FULL,
descr + XDES_FLST_NODE, mtr);
} else if (xdes_is_free(descr, mtr)) {
flst_remove(seg_inode + FSEG_FREE,
descr + XDES_FLST_NODE, mtr);
} else {
flst_remove(seg_inode + FSEG_NOT_FULL,
descr + XDES_FLST_NODE, mtr);
not_full_n_used = mtr_read_ulint(
seg_inode + FSEG_NOT_FULL_N_USED, MLOG_4BYTES, mtr);
descr_n_used = xdes_get_n_used(descr, mtr);
ut_a(not_full_n_used >= descr_n_used);
mlog_write_ulint(seg_inode + FSEG_NOT_FULL_N_USED,
not_full_n_used - descr_n_used,
MLOG_4BYTES, mtr);
}
fsp_free_extent(space, zip_size, page, mtr);
#ifdef UNIV_DEBUG_FILE_ACCESSES
for (i = 0; i < FSP_EXTENT_SIZE; i++) {
buf_page_set_file_page_was_freed(space,
first_page_in_extent + i);
}
#endif
}
/**********************************************************************//**
Frees part of a segment. This function can be used to free a segment by
repeatedly calling this function in different mini-transactions. Doing
the freeing in a single mini-transaction might result in too big a
mini-transaction.
@return TRUE if freeing completed */
UNIV_INTERN
ibool
fseg_free_step(
/*===========*/
fseg_header_t* header, /*!< in, own: segment header; NOTE: if the header
resides on the first page of the frag list
of the segment, this pointer becomes obsolete
after the last freeing step */
mtr_t* mtr) /*!< in: mtr */
{
ulint n;
ulint page;
xdes_t* descr;
fseg_inode_t* inode;
ulint space;
ulint flags;
ulint zip_size;
ulint header_page;
rw_lock_t* latch;
space = page_get_space_id(page_align(header));
header_page = page_get_page_no(page_align(header));
latch = fil_space_get_latch(space, &flags);
zip_size = dict_table_flags_to_zip_size(flags);
ut_ad(!mutex_own(&kernel_mutex)
|| mtr_memo_contains(mtr, latch, MTR_MEMO_X_LOCK));
mtr_x_lock(latch, mtr);
descr = xdes_get_descriptor(space, zip_size, header_page, mtr);
/* Check that the header resides on a page which has not been
freed yet */
ut_a(descr);
ut_a(xdes_get_bit(descr, XDES_FREE_BIT,
header_page % FSP_EXTENT_SIZE, mtr) == FALSE);
inode = fseg_inode_get(header, space, zip_size, mtr);
descr = fseg_get_first_extent(inode, space, zip_size, mtr);
if (descr != NULL) {
/* Free the extent held by the segment */
page = xdes_get_offset(descr);
fseg_free_extent(inode, space, zip_size, page, mtr);
return(FALSE);
}
/* Free a frag page */
n = fseg_find_last_used_frag_page_slot(inode, mtr);
if (n == ULINT_UNDEFINED) {
/* Freeing completed: free the segment inode */
fsp_free_seg_inode(space, zip_size, inode, mtr);
return(TRUE);
}
fseg_free_page_low(inode, space, zip_size,
fseg_get_nth_frag_page_no(inode, n, mtr), mtr);
n = fseg_find_last_used_frag_page_slot(inode, mtr);
if (n == ULINT_UNDEFINED) {
/* Freeing completed: free the segment inode */
fsp_free_seg_inode(space, zip_size, inode, mtr);
return(TRUE);
}
return(FALSE);
}
/**********************************************************************//**
Frees part of a segment. Differs from fseg_free_step because this function
leaves the header page unfreed.
@return TRUE if freeing completed, except the header page */
UNIV_INTERN
ibool
fseg_free_step_not_header(
/*======================*/
fseg_header_t* header, /*!< in: segment header which must reside on
the first fragment page of the segment */
mtr_t* mtr) /*!< in: mtr */
{
ulint n;
ulint page;
xdes_t* descr;
fseg_inode_t* inode;
ulint space;
ulint flags;
ulint zip_size;
ulint page_no;
rw_lock_t* latch;
space = page_get_space_id(page_align(header));
latch = fil_space_get_latch(space, &flags);
zip_size = dict_table_flags_to_zip_size(flags);
ut_ad(!mutex_own(&kernel_mutex)
|| mtr_memo_contains(mtr, latch, MTR_MEMO_X_LOCK));
mtr_x_lock(latch, mtr);
inode = fseg_inode_get(header, space, zip_size, mtr);
descr = fseg_get_first_extent(inode, space, zip_size, mtr);
if (descr != NULL) {
/* Free the extent held by the segment */
page = xdes_get_offset(descr);
fseg_free_extent(inode, space, zip_size, page, mtr);
return(FALSE);
}
/* Free a frag page */
n = fseg_find_last_used_frag_page_slot(inode, mtr);
if (n == ULINT_UNDEFINED) {
ut_error;
}
page_no = fseg_get_nth_frag_page_no(inode, n, mtr);
if (page_no == page_get_page_no(page_align(header))) {
return(TRUE);
}
fseg_free_page_low(inode, space, zip_size, page_no, mtr);
return(FALSE);
}
/**********************************************************************//**
Returns the first extent descriptor for a segment. We think of the extent
lists of the segment catenated in the order FSEG_FULL -> FSEG_NOT_FULL
-> FSEG_FREE.
@return the first extent descriptor, or NULL if none */
static
xdes_t*
fseg_get_first_extent(
/*==================*/
fseg_inode_t* inode, /*!< in: segment inode */
ulint space, /*!< in: space id */
ulint zip_size,/*!< in: compressed page size in bytes
or 0 for uncompressed pages */
mtr_t* mtr) /*!< in: mtr */
{
fil_addr_t first;
xdes_t* descr;
ut_ad(inode && mtr);
ut_ad(space == page_get_space_id(page_align(inode)));
first = fil_addr_null;
if (flst_get_len(inode + FSEG_FULL, mtr) > 0) {
first = flst_get_first(inode + FSEG_FULL, mtr);
} else if (flst_get_len(inode + FSEG_NOT_FULL, mtr) > 0) {
first = flst_get_first(inode + FSEG_NOT_FULL, mtr);
} else if (flst_get_len(inode + FSEG_FREE, mtr) > 0) {
first = flst_get_first(inode + FSEG_FREE, mtr);
}
if (first.page == FIL_NULL) {
return(NULL);
}
descr = xdes_lst_get_descriptor(space, zip_size, first, mtr);
return(descr);
}
/*******************************************************************//**
Validates a segment.
@return TRUE if ok */
static
ibool
fseg_validate_low(
/*==============*/
fseg_inode_t* inode, /*!< in: segment inode */
mtr_t* mtr2) /*!< in: mtr */
{
ulint space;
dulint seg_id;
mtr_t mtr;
xdes_t* descr;
fil_addr_t node_addr;
ulint n_used = 0;
ulint n_used2 = 0;
ut_ad(mtr_memo_contains_page(mtr2, inode, MTR_MEMO_PAGE_X_FIX));
ut_ad(mach_read_from_4(inode + FSEG_MAGIC_N) == FSEG_MAGIC_N_VALUE);
space = page_get_space_id(page_align(inode));
seg_id = mtr_read_dulint(inode + FSEG_ID, mtr2);
n_used = mtr_read_ulint(inode + FSEG_NOT_FULL_N_USED,
MLOG_4BYTES, mtr2);
flst_validate(inode + FSEG_FREE, mtr2);
flst_validate(inode + FSEG_NOT_FULL, mtr2);
flst_validate(inode + FSEG_FULL, mtr2);
/* Validate FSEG_FREE list */
node_addr = flst_get_first(inode + FSEG_FREE, mtr2);
while (!fil_addr_is_null(node_addr)) {
ulint flags;
ulint zip_size;
mtr_start(&mtr);
mtr_x_lock(fil_space_get_latch(space, &flags), &mtr);
zip_size = dict_table_flags_to_zip_size(flags);
descr = xdes_lst_get_descriptor(space, zip_size,
node_addr, &mtr);
ut_a(xdes_get_n_used(descr, &mtr) == 0);
ut_a(xdes_get_state(descr, &mtr) == XDES_FSEG);
ut_a(!ut_dulint_cmp(mtr_read_dulint(descr + XDES_ID, &mtr),
seg_id));
node_addr = flst_get_next_addr(descr + XDES_FLST_NODE, &mtr);
mtr_commit(&mtr);
}
/* Validate FSEG_NOT_FULL list */
node_addr = flst_get_first(inode + FSEG_NOT_FULL, mtr2);
while (!fil_addr_is_null(node_addr)) {
ulint flags;
ulint zip_size;
mtr_start(&mtr);
mtr_x_lock(fil_space_get_latch(space, &flags), &mtr);
zip_size = dict_table_flags_to_zip_size(flags);
descr = xdes_lst_get_descriptor(space, zip_size,
node_addr, &mtr);
ut_a(xdes_get_n_used(descr, &mtr) > 0);
ut_a(xdes_get_n_used(descr, &mtr) < FSP_EXTENT_SIZE);
ut_a(xdes_get_state(descr, &mtr) == XDES_FSEG);
ut_a(!ut_dulint_cmp(mtr_read_dulint(descr + XDES_ID, &mtr),
seg_id));
n_used2 += xdes_get_n_used(descr, &mtr);
node_addr = flst_get_next_addr(descr + XDES_FLST_NODE, &mtr);
mtr_commit(&mtr);
}
/* Validate FSEG_FULL list */
node_addr = flst_get_first(inode + FSEG_FULL, mtr2);
while (!fil_addr_is_null(node_addr)) {
ulint flags;
ulint zip_size;
mtr_start(&mtr);
mtr_x_lock(fil_space_get_latch(space, &flags), &mtr);
zip_size = dict_table_flags_to_zip_size(flags);
descr = xdes_lst_get_descriptor(space, zip_size,
node_addr, &mtr);
ut_a(xdes_get_n_used(descr, &mtr) == FSP_EXTENT_SIZE);
ut_a(xdes_get_state(descr, &mtr) == XDES_FSEG);
ut_a(!ut_dulint_cmp(mtr_read_dulint(descr + XDES_ID, &mtr),
seg_id));
node_addr = flst_get_next_addr(descr + XDES_FLST_NODE, &mtr);
mtr_commit(&mtr);
}
ut_a(n_used == n_used2);
return(TRUE);
}
#ifdef UNIV_DEBUG
/*******************************************************************//**
Validates a segment.
@return TRUE if ok */
UNIV_INTERN
ibool
fseg_validate(
/*==========*/
fseg_header_t* header, /*!< in: segment header */
mtr_t* mtr) /*!< in: mtr */
{
fseg_inode_t* inode;
ibool ret;
ulint space;
ulint flags;
ulint zip_size;
space = page_get_space_id(page_align(header));
mtr_x_lock(fil_space_get_latch(space, &flags), mtr);
zip_size = dict_table_flags_to_zip_size(flags);
inode = fseg_inode_get(header, space, zip_size, mtr);
ret = fseg_validate_low(inode, mtr);
return(ret);
}
#endif /* UNIV_DEBUG */
/*******************************************************************//**
Writes info of a segment. */
static
void
fseg_print_low(
/*===========*/
fseg_inode_t* inode, /*!< in: segment inode */
mtr_t* mtr) /*!< in: mtr */
{
ulint space;
ulint seg_id_low;
ulint seg_id_high;
ulint n_used;
ulint n_frag;
ulint n_free;
ulint n_not_full;
ulint n_full;
ulint reserved;
ulint used;
ulint page_no;
dulint d_var;
ut_ad(mtr_memo_contains_page(mtr, inode, MTR_MEMO_PAGE_X_FIX));
space = page_get_space_id(page_align(inode));
page_no = page_get_page_no(page_align(inode));
reserved = fseg_n_reserved_pages_low(inode, &used, mtr);
d_var = mtr_read_dulint(inode + FSEG_ID, mtr);
seg_id_low = ut_dulint_get_low(d_var);
seg_id_high = ut_dulint_get_high(d_var);
n_used = mtr_read_ulint(inode + FSEG_NOT_FULL_N_USED,
MLOG_4BYTES, mtr);
n_frag = fseg_get_n_frag_pages(inode, mtr);
n_free = flst_get_len(inode + FSEG_FREE, mtr);
n_not_full = flst_get_len(inode + FSEG_NOT_FULL, mtr);
n_full = flst_get_len(inode + FSEG_FULL, mtr);
fprintf(stderr,
"SEGMENT id %lu %lu space %lu; page %lu;"
" res %lu used %lu; full ext %lu\n"
"fragm pages %lu; free extents %lu;"
" not full extents %lu: pages %lu\n",
(ulong) seg_id_high, (ulong) seg_id_low,
(ulong) space, (ulong) page_no,
(ulong) reserved, (ulong) used, (ulong) n_full,
(ulong) n_frag, (ulong) n_free, (ulong) n_not_full,
(ulong) n_used);
}
#ifdef UNIV_BTR_PRINT
/*******************************************************************//**
Writes info of a segment. */
UNIV_INTERN
void
fseg_print(
/*=======*/
fseg_header_t* header, /*!< in: segment header */
mtr_t* mtr) /*!< in: mtr */
{
fseg_inode_t* inode;
ulint space;
ulint flags;
ulint zip_size;
space = page_get_space_id(page_align(header));
mtr_x_lock(fil_space_get_latch(space, &flags), mtr);
zip_size = dict_table_flags_to_zip_size(flags);
inode = fseg_inode_get(header, space, zip_size, mtr);
fseg_print_low(inode, mtr);
}
#endif /* UNIV_BTR_PRINT */
/*******************************************************************//**
Validates the file space system and its segments.
@return TRUE if ok */
UNIV_INTERN
ibool
fsp_validate(
/*=========*/
ulint space) /*!< in: space id */
{
fsp_header_t* header;
fseg_inode_t* seg_inode;
page_t* seg_inode_page;
rw_lock_t* latch;
ulint size;
ulint flags;
ulint zip_size;
ulint free_limit;
ulint frag_n_used;
mtr_t mtr;
mtr_t mtr2;
xdes_t* descr;
fil_addr_t node_addr;
fil_addr_t next_node_addr;
ulint descr_count = 0;
ulint n_used = 0;
ulint n_used2 = 0;
ulint n_full_frag_pages;
ulint n;
ulint seg_inode_len_free;
ulint seg_inode_len_full;
latch = fil_space_get_latch(space, &flags);
zip_size = dict_table_flags_to_zip_size(flags);
ut_a(ut_is_2pow(zip_size));
ut_a(zip_size <= UNIV_PAGE_SIZE);
ut_a(!zip_size || zip_size >= PAGE_ZIP_MIN_SIZE);
/* Start first a mini-transaction mtr2 to lock out all other threads
from the fsp system */
mtr_start(&mtr2);
mtr_x_lock(latch, &mtr2);
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
header = fsp_get_space_header(space, zip_size, &mtr);
size = mtr_read_ulint(header + FSP_SIZE, MLOG_4BYTES, &mtr);
free_limit = mtr_read_ulint(header + FSP_FREE_LIMIT,
MLOG_4BYTES, &mtr);
frag_n_used = mtr_read_ulint(header + FSP_FRAG_N_USED,
MLOG_4BYTES, &mtr);
n_full_frag_pages = FSP_EXTENT_SIZE
* flst_get_len(header + FSP_FULL_FRAG, &mtr);
if (UNIV_UNLIKELY(free_limit > size)) {
ut_a(space != 0);
ut_a(size < FSP_EXTENT_SIZE);
}
flst_validate(header + FSP_FREE, &mtr);
flst_validate(header + FSP_FREE_FRAG, &mtr);
flst_validate(header + FSP_FULL_FRAG, &mtr);
mtr_commit(&mtr);
/* Validate FSP_FREE list */
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
header = fsp_get_space_header(space, zip_size, &mtr);
node_addr = flst_get_first(header + FSP_FREE, &mtr);
mtr_commit(&mtr);
while (!fil_addr_is_null(node_addr)) {
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
descr_count++;
descr = xdes_lst_get_descriptor(space, zip_size,
node_addr, &mtr);
ut_a(xdes_get_n_used(descr, &mtr) == 0);
ut_a(xdes_get_state(descr, &mtr) == XDES_FREE);
node_addr = flst_get_next_addr(descr + XDES_FLST_NODE, &mtr);
mtr_commit(&mtr);
}
/* Validate FSP_FREE_FRAG list */
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
header = fsp_get_space_header(space, zip_size, &mtr);
node_addr = flst_get_first(header + FSP_FREE_FRAG, &mtr);
mtr_commit(&mtr);
while (!fil_addr_is_null(node_addr)) {
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
descr_count++;
descr = xdes_lst_get_descriptor(space, zip_size,
node_addr, &mtr);
ut_a(xdes_get_n_used(descr, &mtr) > 0);
ut_a(xdes_get_n_used(descr, &mtr) < FSP_EXTENT_SIZE);
ut_a(xdes_get_state(descr, &mtr) == XDES_FREE_FRAG);
n_used += xdes_get_n_used(descr, &mtr);
node_addr = flst_get_next_addr(descr + XDES_FLST_NODE, &mtr);
mtr_commit(&mtr);
}
/* Validate FSP_FULL_FRAG list */
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
header = fsp_get_space_header(space, zip_size, &mtr);
node_addr = flst_get_first(header + FSP_FULL_FRAG, &mtr);
mtr_commit(&mtr);
while (!fil_addr_is_null(node_addr)) {
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
descr_count++;
descr = xdes_lst_get_descriptor(space, zip_size,
node_addr, &mtr);
ut_a(xdes_get_n_used(descr, &mtr) == FSP_EXTENT_SIZE);
ut_a(xdes_get_state(descr, &mtr) == XDES_FULL_FRAG);
node_addr = flst_get_next_addr(descr + XDES_FLST_NODE, &mtr);
mtr_commit(&mtr);
}
/* Validate segments */
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
header = fsp_get_space_header(space, zip_size, &mtr);
node_addr = flst_get_first(header + FSP_SEG_INODES_FULL, &mtr);
seg_inode_len_full = flst_get_len(header + FSP_SEG_INODES_FULL, &mtr);
mtr_commit(&mtr);
while (!fil_addr_is_null(node_addr)) {
n = 0;
do {
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
seg_inode_page = fut_get_ptr(
space, zip_size, node_addr, RW_X_LATCH, &mtr)
- FSEG_INODE_PAGE_NODE;
seg_inode = fsp_seg_inode_page_get_nth_inode(
seg_inode_page, n, zip_size, &mtr);
ut_a(!ut_dulint_is_zero(
mach_read_from_8(seg_inode + FSEG_ID)));
fseg_validate_low(seg_inode, &mtr);
descr_count += flst_get_len(seg_inode + FSEG_FREE,
&mtr);
descr_count += flst_get_len(seg_inode + FSEG_FULL,
&mtr);
descr_count += flst_get_len(seg_inode + FSEG_NOT_FULL,
&mtr);
n_used2 += fseg_get_n_frag_pages(seg_inode, &mtr);
next_node_addr = flst_get_next_addr(
seg_inode_page + FSEG_INODE_PAGE_NODE, &mtr);
mtr_commit(&mtr);
} while (++n < FSP_SEG_INODES_PER_PAGE(zip_size));
node_addr = next_node_addr;
}
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
header = fsp_get_space_header(space, zip_size, &mtr);
node_addr = flst_get_first(header + FSP_SEG_INODES_FREE, &mtr);
seg_inode_len_free = flst_get_len(header + FSP_SEG_INODES_FREE, &mtr);
mtr_commit(&mtr);
while (!fil_addr_is_null(node_addr)) {
n = 0;
do {
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
seg_inode_page = fut_get_ptr(
space, zip_size, node_addr, RW_X_LATCH, &mtr)
- FSEG_INODE_PAGE_NODE;
seg_inode = fsp_seg_inode_page_get_nth_inode(
seg_inode_page, n, zip_size, &mtr);
if (!ut_dulint_is_zero(
mach_read_from_8(seg_inode + FSEG_ID))) {
fseg_validate_low(seg_inode, &mtr);
descr_count += flst_get_len(
seg_inode + FSEG_FREE, &mtr);
descr_count += flst_get_len(
seg_inode + FSEG_FULL, &mtr);
descr_count += flst_get_len(
seg_inode + FSEG_NOT_FULL, &mtr);
n_used2 += fseg_get_n_frag_pages(
seg_inode, &mtr);
}
next_node_addr = flst_get_next_addr(
seg_inode_page + FSEG_INODE_PAGE_NODE, &mtr);
mtr_commit(&mtr);
} while (++n < FSP_SEG_INODES_PER_PAGE(zip_size));
node_addr = next_node_addr;
}
ut_a(descr_count * FSP_EXTENT_SIZE == free_limit);
if (!zip_size) {
ut_a(n_used + n_full_frag_pages
== n_used2 + 2 * ((free_limit + (UNIV_PAGE_SIZE - 1))
/ UNIV_PAGE_SIZE)
+ seg_inode_len_full + seg_inode_len_free);
} else {
ut_a(n_used + n_full_frag_pages
== n_used2 + 2 * ((free_limit + (zip_size - 1))
/ zip_size)
+ seg_inode_len_full + seg_inode_len_free);
}
ut_a(frag_n_used == n_used);
mtr_commit(&mtr2);
return(TRUE);
}
/*******************************************************************//**
Prints info of a file space. */
UNIV_INTERN
void
fsp_print(
/*======*/
ulint space) /*!< in: space id */
{
fsp_header_t* header;
fseg_inode_t* seg_inode;
page_t* seg_inode_page;
rw_lock_t* latch;
ulint flags;
ulint zip_size;
ulint size;
ulint free_limit;
ulint frag_n_used;
fil_addr_t node_addr;
fil_addr_t next_node_addr;
ulint n_free;
ulint n_free_frag;
ulint n_full_frag;
ulint seg_id_low;
ulint seg_id_high;
ulint n;
ulint n_segs = 0;
dulint d_var;
mtr_t mtr;
mtr_t mtr2;
latch = fil_space_get_latch(space, &flags);
zip_size = dict_table_flags_to_zip_size(flags);
/* Start first a mini-transaction mtr2 to lock out all other threads
from the fsp system */
mtr_start(&mtr2);
mtr_x_lock(latch, &mtr2);
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
header = fsp_get_space_header(space, zip_size, &mtr);
size = mtr_read_ulint(header + FSP_SIZE, MLOG_4BYTES, &mtr);
free_limit = mtr_read_ulint(header + FSP_FREE_LIMIT, MLOG_4BYTES,
&mtr);
frag_n_used = mtr_read_ulint(header + FSP_FRAG_N_USED, MLOG_4BYTES,
&mtr);
n_free = flst_get_len(header + FSP_FREE, &mtr);
n_free_frag = flst_get_len(header + FSP_FREE_FRAG, &mtr);
n_full_frag = flst_get_len(header + FSP_FULL_FRAG, &mtr);
d_var = mtr_read_dulint(header + FSP_SEG_ID, &mtr);
seg_id_low = ut_dulint_get_low(d_var);
seg_id_high = ut_dulint_get_high(d_var);
fprintf(stderr,
"FILE SPACE INFO: id %lu\n"
"size %lu, free limit %lu, free extents %lu\n"
"not full frag extents %lu: used pages %lu,"
" full frag extents %lu\n"
"first seg id not used %lu %lu\n",
(ulong) space,
(ulong) size, (ulong) free_limit, (ulong) n_free,
(ulong) n_free_frag, (ulong) frag_n_used, (ulong) n_full_frag,
(ulong) seg_id_high, (ulong) seg_id_low);
mtr_commit(&mtr);
/* Print segments */
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
header = fsp_get_space_header(space, zip_size, &mtr);
node_addr = flst_get_first(header + FSP_SEG_INODES_FULL, &mtr);
mtr_commit(&mtr);
while (!fil_addr_is_null(node_addr)) {
n = 0;
do {
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
seg_inode_page = fut_get_ptr(
space, zip_size, node_addr, RW_X_LATCH, &mtr)
- FSEG_INODE_PAGE_NODE;
seg_inode = fsp_seg_inode_page_get_nth_inode(
seg_inode_page, n, zip_size, &mtr);
ut_a(!ut_dulint_is_zero(
mach_read_from_8(seg_inode + FSEG_ID)));
fseg_print_low(seg_inode, &mtr);
n_segs++;
next_node_addr = flst_get_next_addr(
seg_inode_page + FSEG_INODE_PAGE_NODE, &mtr);
mtr_commit(&mtr);
} while (++n < FSP_SEG_INODES_PER_PAGE(zip_size));
node_addr = next_node_addr;
}
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
header = fsp_get_space_header(space, zip_size, &mtr);
node_addr = flst_get_first(header + FSP_SEG_INODES_FREE, &mtr);
mtr_commit(&mtr);
while (!fil_addr_is_null(node_addr)) {
n = 0;
do {
mtr_start(&mtr);
mtr_x_lock(latch, &mtr);
seg_inode_page = fut_get_ptr(
space, zip_size, node_addr, RW_X_LATCH, &mtr)
- FSEG_INODE_PAGE_NODE;
seg_inode = fsp_seg_inode_page_get_nth_inode(
seg_inode_page, n, zip_size, &mtr);
if (!ut_dulint_is_zero(
mach_read_from_8(seg_inode + FSEG_ID))) {
fseg_print_low(seg_inode, &mtr);
n_segs++;
}
next_node_addr = flst_get_next_addr(
seg_inode_page + FSEG_INODE_PAGE_NODE, &mtr);
mtr_commit(&mtr);
} while (++n < FSP_SEG_INODES_PER_PAGE(zip_size));
node_addr = next_node_addr;
}
mtr_commit(&mtr2);
fprintf(stderr, "NUMBER of file segments: %lu\n", (ulong) n_segs);
}
#endif /* !UNIV_HOTBACKUP */