summaryrefslogtreecommitdiff
path: root/storage/innobase/include
diff options
context:
space:
mode:
authorSergei Golubchik <sergii@pisem.net>2011-11-03 19:17:05 +0100
committerSergei Golubchik <sergii@pisem.net>2011-11-03 19:17:05 +0100
commit0e007344eae972b9be7d88ca43373cb33662ac1c (patch)
tree7b3561cb3ea2ad24d013e847680ec165f406387e /storage/innobase/include
parent3794110f0215f0631107c2694dc0f1675a4bb520 (diff)
parent681476255147dacac7e3674b6cd2ae770fee2208 (diff)
downloadmariadb-git-0e007344eae972b9be7d88ca43373cb33662ac1c.tar.gz
mysql-5.5.18 merge
Diffstat (limited to 'storage/innobase/include')
-rw-r--r--storage/innobase/include/btr0btr.h32
-rw-r--r--storage/innobase/include/btr0btr.ic10
-rw-r--r--storage/innobase/include/btr0pcur.h21
-rw-r--r--storage/innobase/include/btr0pcur.ic9
-rw-r--r--storage/innobase/include/buf0buddy.h46
-rw-r--r--storage/innobase/include/buf0buddy.ic68
-rw-r--r--storage/innobase/include/buf0buf.h47
-rw-r--r--storage/innobase/include/buf0buf.ic59
-rw-r--r--storage/innobase/include/buf0lru.h51
-rw-r--r--storage/innobase/include/buf0rea.h25
-rw-r--r--storage/innobase/include/buf0types.h17
-rw-r--r--storage/innobase/include/db0err.h6
-rw-r--r--storage/innobase/include/dict0boot.h6
-rw-r--r--storage/innobase/include/dict0dict.h82
-rw-r--r--storage/innobase/include/dict0dict.ic46
-rw-r--r--storage/innobase/include/dict0mem.h14
-rw-r--r--storage/innobase/include/dict0types.h3
-rw-r--r--storage/innobase/include/mtr0mtr.h10
-rw-r--r--storage/innobase/include/page0page.h5
-rw-r--r--storage/innobase/include/row0row.h28
-rw-r--r--storage/innobase/include/row0row.ic25
-rw-r--r--storage/innobase/include/row0upd.ic9
-rw-r--r--storage/innobase/include/srv0srv.h6
-rw-r--r--storage/innobase/include/sync0rw.h5
-rw-r--r--storage/innobase/include/sync0rw.ic1
-rw-r--r--storage/innobase/include/sync0sync.h10
-rw-r--r--storage/innobase/include/sync0sync.ic5
-rw-r--r--storage/innobase/include/trx0sys.h13
-rw-r--r--storage/innobase/include/trx0sys.ic24
-rw-r--r--storage/innobase/include/trx0undo.h46
-rw-r--r--storage/innobase/include/ut0mem.h39
-rw-r--r--storage/innobase/include/ut0mem.ic23
32 files changed, 502 insertions, 289 deletions
diff --git a/storage/innobase/include/btr0btr.h b/storage/innobase/include/btr0btr.h
index 5b3e166371d..71e772388a0 100644
--- a/storage/innobase/include/btr0btr.h
+++ b/storage/innobase/include/btr0btr.h
@@ -199,26 +199,45 @@ btr_block_get_func(
ulint mode, /*!< in: latch mode */
const char* file, /*!< in: file name */
ulint line, /*!< in: line where called */
- mtr_t* mtr) /*!< in/out: mtr */
- __attribute__((nonnull));
+# ifdef UNIV_SYNC_DEBUG
+ const dict_index_t* index, /*!< in: index tree, may be NULL
+ if it is not an insert buffer tree */
+# endif /* UNIV_SYNC_DEBUG */
+ mtr_t* mtr); /*!< in/out: mini-transaction */
+# ifdef UNIV_SYNC_DEBUG
+/** Gets a buffer page and declares its latching order level.
+@param space tablespace identifier
+@param zip_size compressed page size in bytes or 0 for uncompressed pages
+@param page_no page number
+@param mode latch mode
+@param index index tree, may be NULL if not the insert buffer tree
+@param mtr mini-transaction handle
+@return the block descriptor */
+# define btr_block_get(space,zip_size,page_no,mode,index,mtr) \
+ btr_block_get_func(space,zip_size,page_no,mode, \
+ __FILE__,__LINE__,index,mtr)
+# else /* UNIV_SYNC_DEBUG */
/** Gets a buffer page and declares its latching order level.
@param space tablespace identifier
@param zip_size compressed page size in bytes or 0 for uncompressed pages
@param page_no page number
@param mode latch mode
+@param idx index tree, may be NULL if not the insert buffer tree
@param mtr mini-transaction handle
@return the block descriptor */
-# define btr_block_get(space,zip_size,page_no,mode,mtr) \
+# define btr_block_get(space,zip_size,page_no,mode,idx,mtr) \
btr_block_get_func(space,zip_size,page_no,mode,__FILE__,__LINE__,mtr)
+# endif /* UNIV_SYNC_DEBUG */
/** Gets a buffer page and declares its latching order level.
@param space tablespace identifier
@param zip_size compressed page size in bytes or 0 for uncompressed pages
@param page_no page number
@param mode latch mode
+@param idx index tree, may be NULL if not the insert buffer tree
@param mtr mini-transaction handle
@return the uncompressed page frame */
-# define btr_page_get(space,zip_size,page_no,mode,mtr) \
- buf_block_get_frame(btr_block_get(space,zip_size,page_no,mode,mtr))
+# define btr_page_get(space,zip_size,page_no,mode,idx,mtr) \
+ buf_block_get_frame(btr_block_get(space,zip_size,page_no,mode,idx,mtr))
#endif /* !UNIV_HOTBACKUP */
/**************************************************************//**
Gets the index id field of a page.
@@ -344,8 +363,7 @@ btr_free_root(
ulint zip_size, /*!< in: compressed page size in bytes
or 0 for uncompressed pages */
ulint root_page_no, /*!< in: root page number */
- mtr_t* mtr); /*!< in: a mini-transaction which has already
- been started */
+ mtr_t* mtr); /*!< in/out: mini-transaction */
/*************************************************************//**
Makes tree one level higher by splitting the root, and inserts
the tuple. It is assumed that mtr contains an x-latch on the tree.
diff --git a/storage/innobase/include/btr0btr.ic b/storage/innobase/include/btr0btr.ic
index ccf41904fd0..1bd22a0ebc6 100644
--- a/storage/innobase/include/btr0btr.ic
+++ b/storage/innobase/include/btr0btr.ic
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2011, Oracle and/or its affiliates. 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
@@ -48,6 +48,10 @@ btr_block_get_func(
ulint mode, /*!< in: latch mode */
const char* file, /*!< in: file name */
ulint line, /*!< in: line where called */
+#ifdef UNIV_SYNC_DEBUG
+ const dict_index_t* index, /*!< in: index tree, may be NULL
+ if it is not an insert buffer tree */
+#endif /* UNIV_SYNC_DEBUG */
mtr_t* mtr) /*!< in/out: mtr */
{
buf_block_t* block;
@@ -57,7 +61,9 @@ btr_block_get_func(
if (mode != RW_NO_LATCH) {
- buf_block_dbg_add_level(block, SYNC_TREE_NODE);
+ buf_block_dbg_add_level(
+ block, index != NULL && dict_index_is_ibuf(index)
+ ? SYNC_IBUF_TREE_NODE : SYNC_TREE_NODE);
}
return(block);
diff --git a/storage/innobase/include/btr0pcur.h b/storage/innobase/include/btr0pcur.h
index 6c11c973fc9..f605c476844 100644
--- a/storage/innobase/include/btr0pcur.h
+++ b/storage/innobase/include/btr0pcur.h
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2011, Oracle and/or its affiliates. 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
@@ -244,18 +244,6 @@ btr_pcur_restore_position_func(
mtr_t* mtr); /*!< in: mtr */
#define btr_pcur_restore_position(l,cur,mtr) \
btr_pcur_restore_position_func(l,cur,__FILE__,__LINE__,mtr)
-/**************************************************************//**
-If the latch mode of the cursor is BTR_LEAF_SEARCH or BTR_LEAF_MODIFY,
-releases the page latch and bufferfix reserved by the cursor.
-NOTE! In the case of BTR_LEAF_MODIFY, there should not exist changes
-made by the current mini-transaction to the data protected by the
-cursor latch, as then the latch must not be released until mtr_commit. */
-UNIV_INTERN
-void
-btr_pcur_release_leaf(
-/*==================*/
- btr_pcur_t* cursor, /*!< in: persistent cursor */
- mtr_t* mtr); /*!< in: mtr */
/*********************************************************//**
Gets the rel_pos field for a cursor whose position has been stored.
@return BTR_PCUR_ON, ... */
@@ -266,10 +254,9 @@ btr_pcur_get_rel_pos(
const btr_pcur_t* cursor);/*!< in: persistent cursor */
/**************************************************************//**
Commits the mtr and sets the pcur latch mode to BTR_NO_LATCHES,
-that is, the cursor becomes detached. If there have been modifications
-to the page where pcur is positioned, this can be used instead of
-btr_pcur_release_leaf. Function btr_pcur_store_position should be used
-before calling this, if restoration of cursor is wanted later. */
+that is, the cursor becomes detached.
+Function btr_pcur_store_position should be used before calling this,
+if restoration of cursor is wanted later. */
UNIV_INLINE
void
btr_pcur_commit_specify_mtr(
diff --git a/storage/innobase/include/btr0pcur.ic b/storage/innobase/include/btr0pcur.ic
index 59fdb21824b..d86601e5a32 100644
--- a/storage/innobase/include/btr0pcur.ic
+++ b/storage/innobase/include/btr0pcur.ic
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2011, Oracle and/or its affiliates. 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
@@ -369,10 +369,9 @@ btr_pcur_move_to_next(
/**************************************************************//**
Commits the mtr and sets the pcur latch mode to BTR_NO_LATCHES,
-that is, the cursor becomes detached. If there have been modifications
-to the page where pcur is positioned, this can be used instead of
-btr_pcur_release_leaf. Function btr_pcur_store_position should be used
-before calling this, if restoration of cursor is wanted later. */
+that is, the cursor becomes detached.
+Function btr_pcur_store_position should be used before calling this,
+if restoration of cursor is wanted later. */
UNIV_INLINE
void
btr_pcur_commit_specify_mtr(
diff --git a/storage/innobase/include/buf0buddy.h b/storage/innobase/include/buf0buddy.h
index b255d8c9351..2d7d6146092 100644
--- a/storage/innobase/include/buf0buddy.h
+++ b/storage/innobase/include/buf0buddy.h
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 2006, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 2006, 2011, Oracle and/or its affiliates. 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
@@ -37,37 +37,37 @@ Created December 2006 by Marko Makela
/**********************************************************************//**
Allocate a block. The thread calling this function must hold
buf_pool->mutex and must not hold buf_pool->zip_mutex or any
-block->mutex. The buf_pool->mutex may only be released and reacquired
-if lru != NULL. This function should only be used for allocating
-compressed page frames or control blocks (buf_page_t). Allocated
-control blocks must be properly initialized immediately after
-buf_buddy_alloc() has returned the memory, before releasing
-buf_pool->mutex.
-@return allocated block, possibly NULL if lru == NULL */
+block->mutex. The buf_pool->mutex may be released and reacquired.
+This function should only be used for allocating compressed page frames.
+@return allocated block, never NULL */
UNIV_INLINE
-void*
+byte*
buf_buddy_alloc(
/*============*/
- buf_pool_t* buf_pool,
- /*!< buffer pool in which the block resides */
- ulint size, /*!< in: block size, up to UNIV_PAGE_SIZE */
- ibool* lru) /*!< in: pointer to a variable that will be assigned
- TRUE if storage was allocated from the LRU list
- and buf_pool->mutex was temporarily released,
- or NULL if the LRU list should not be used */
- __attribute__((malloc));
+ buf_pool_t* buf_pool, /*!< in/out: buffer pool in which
+ the page resides */
+ ulint size, /*!< in: compressed page size
+ (between PAGE_ZIP_MIN_SIZE and
+ UNIV_PAGE_SIZE) */
+ ibool* lru) /*!< in: pointer to a variable
+ that will be assigned TRUE if
+ storage was allocated from the
+ LRU list and buf_pool->mutex was
+ temporarily released */
+ __attribute__((malloc, nonnull));
/**********************************************************************//**
-Release a block. */
+Deallocate a block. */
UNIV_INLINE
void
buf_buddy_free(
/*===========*/
- buf_pool_t* buf_pool,
- /*!< buffer pool in which the block resides */
- void* buf, /*!< in: block to be freed, must not be
- pointed to by the buffer pool */
- ulint size) /*!< in: block size, up to UNIV_PAGE_SIZE */
+ buf_pool_t* buf_pool, /*!< in/out: buffer pool in which
+ the block resides */
+ void* buf, /*!< in: block to be freed, must not
+ be pointed to by the buffer pool */
+ ulint size) /*!< in: block size,
+ up to UNIV_PAGE_SIZE */
__attribute__((nonnull));
#ifndef UNIV_NONINL
diff --git a/storage/innobase/include/buf0buddy.ic b/storage/innobase/include/buf0buddy.ic
index e50c33ea15a..b8281f7341a 100644
--- a/storage/innobase/include/buf0buddy.ic
+++ b/storage/innobase/include/buf0buddy.ic
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 2006, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 2006, 2011, Oracle and/or its affiliates. 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
@@ -36,21 +36,21 @@ Created December 2006 by Marko Makela
/**********************************************************************//**
Allocate a block. The thread calling this function must hold
buf_pool->mutex and must not hold buf_pool->zip_mutex or any block->mutex.
-The buf_pool->mutex may only be released and reacquired if lru != NULL.
-@return allocated block, possibly NULL if lru==NULL */
+The buf_pool_mutex may be released and reacquired.
+@return allocated block, never NULL */
UNIV_INTERN
void*
buf_buddy_alloc_low(
/*================*/
- buf_pool_t* buf_pool,
- /*!< in: buffer pool in which the page resides */
- ulint i, /*!< in: index of buf_pool->zip_free[],
- or BUF_BUDDY_SIZES */
- ibool* lru) /*!< in: pointer to a variable that will be assigned
- TRUE if storage was allocated from the LRU list
- and buf_pool->mutex was temporarily released,
- or NULL if the LRU list should not be used */
- __attribute__((malloc));
+ buf_pool_t* buf_pool, /*!< in/out: buffer pool instance */
+ ulint i, /*!< in: index of buf_pool->zip_free[],
+ or BUF_BUDDY_SIZES */
+ ibool* lru) /*!< in: pointer to a variable that
+ will be assigned TRUE if storage was
+ allocated from the LRU list and
+ buf_pool->mutex was temporarily
+ released */
+ __attribute__((malloc, nonnull));
/**********************************************************************//**
Deallocate a block. */
@@ -77,6 +77,8 @@ buf_buddy_get_slot(
ulint i;
ulint s;
+ ut_ad(size >= PAGE_ZIP_MIN_SIZE);
+
for (i = 0, s = BUF_BUDDY_LOW; s < size; i++, s <<= 1) {
}
@@ -87,31 +89,31 @@ buf_buddy_get_slot(
/**********************************************************************//**
Allocate a block. The thread calling this function must hold
buf_pool->mutex and must not hold buf_pool->zip_mutex or any
-block->mutex. The buf_pool->mutex may only be released and reacquired
-if lru != NULL. This function should only be used for allocating
-compressed page frames or control blocks (buf_page_t). Allocated
-control blocks must be properly initialized immediately after
-buf_buddy_alloc() has returned the memory, before releasing
-buf_pool->mutex.
-@return allocated block, possibly NULL if lru == NULL */
+block->mutex. The buf_pool->mutex may be released and reacquired.
+This function should only be used for allocating compressed page frames.
+@return allocated block, never NULL */
UNIV_INLINE
-void*
+byte*
buf_buddy_alloc(
/*============*/
- buf_pool_t* buf_pool, /*!< in: buffer pool in which
+ buf_pool_t* buf_pool, /*!< in/out: buffer pool in which
the page resides */
- ulint size, /*!< in: block size, up to
- UNIV_PAGE_SIZE */
+ ulint size, /*!< in: compressed page size
+ (between PAGE_ZIP_MIN_SIZE and
+ UNIV_PAGE_SIZE) */
ibool* lru) /*!< in: pointer to a variable
that will be assigned TRUE if
storage was allocated from the
LRU list and buf_pool->mutex was
- temporarily released, or NULL if
- the LRU list should not be used */
+ temporarily released */
{
ut_ad(buf_pool_mutex_own(buf_pool));
+ ut_ad(ut_is_2pow(size));
+ ut_ad(size >= PAGE_ZIP_MIN_SIZE);
+ ut_ad(size <= UNIV_PAGE_SIZE);
- return(buf_buddy_alloc_low(buf_pool, buf_buddy_get_slot(size), lru));
+ return((byte*) buf_buddy_alloc_low(buf_pool, buf_buddy_get_slot(size),
+ lru));
}
/**********************************************************************//**
@@ -120,13 +122,17 @@ UNIV_INLINE
void
buf_buddy_free(
/*===========*/
- buf_pool_t* buf_pool, /*!< in: buffer pool instance */
- void* buf, /*!< in: block to be freed, must not be
- pointed to by the buffer pool */
- ulint size) /*!< in: block size, up to
- UNIV_PAGE_SIZE */
+ buf_pool_t* buf_pool, /*!< in/out: buffer pool in which
+ the block resides */
+ void* buf, /*!< in: block to be freed, must not
+ be pointed to by the buffer pool */
+ ulint size) /*!< in: block size,
+ up to UNIV_PAGE_SIZE */
{
ut_ad(buf_pool_mutex_own(buf_pool));
+ ut_ad(ut_is_2pow(size));
+ ut_ad(size >= PAGE_ZIP_MIN_SIZE);
+ ut_ad(size <= UNIV_PAGE_SIZE);
buf_buddy_free_low(buf_pool, buf, buf_buddy_get_slot(size));
}
diff --git a/storage/innobase/include/buf0buf.h b/storage/innobase/include/buf0buf.h
index d83d484bace..ccebb69a4fe 100644
--- a/storage/innobase/include/buf0buf.h
+++ b/storage/innobase/include/buf0buf.h
@@ -147,6 +147,8 @@ struct buf_pool_info_struct{
ulint n_pages_created; /*!< buf_pool->n_pages_created */
ulint n_pages_written; /*!< buf_pool->n_pages_written */
ulint n_page_gets; /*!< buf_pool->n_page_gets */
+ ulint n_ra_pages_read_rnd; /*!< buf_pool->n_ra_pages_read_rnd,
+ number of pages readahead */
ulint n_ra_pages_read; /*!< buf_pool->n_ra_pages_read, number
of pages readahead */
ulint n_ra_pages_evicted; /*!< buf_pool->n_ra_pages_evicted,
@@ -171,6 +173,8 @@ struct buf_pool_info_struct{
last printout */
/* Statistics about read ahead algorithm. */
+ double pages_readahead_rnd_rate;/*!< random readahead rate in pages per
+ second */
double pages_readahead_rate; /*!< readahead rate in pages per
second */
double pages_evicted_rate; /*!< rate of readahead page evicted
@@ -246,12 +250,6 @@ buf_relocate(
BUF_BLOCK_ZIP_DIRTY or BUF_BLOCK_ZIP_PAGE */
buf_page_t* dpage) /*!< in/out: destination control block */
__attribute__((nonnull));
-/********************************************************************//**
-Resizes the buffer pool. */
-UNIV_INTERN
-void
-buf_pool_resize(void);
-/*=================*/
/*********************************************************************//**
Gets the current size of buffer buf_pool in bytes.
@return size in bytes */
@@ -275,6 +273,23 @@ ib_uint64_t
buf_pool_get_oldest_modification(void);
/*==================================*/
/********************************************************************//**
+Allocates a buf_page_t descriptor. This function must succeed. In case
+of failure we assert in this function. */
+UNIV_INLINE
+buf_page_t*
+buf_page_alloc_descriptor(void)
+/*===========================*/
+ __attribute__((malloc));
+/********************************************************************//**
+Free a buf_page_t descriptor. */
+UNIV_INLINE
+void
+buf_page_free_descriptor(
+/*=====================*/
+ buf_page_t* bpage) /*!< in: bpage descriptor to free. */
+ __attribute__((nonnull));
+
+/********************************************************************//**
Allocates a buffer block.
@return own: the allocated block, in state BUF_BLOCK_MEMORY */
UNIV_INTERN
@@ -531,6 +546,18 @@ buf_block_get_freed_page_clock(
__attribute__((pure));
/********************************************************************//**
+Tells if a block is still close enough to the MRU end of the LRU list
+meaning that it is not in danger of getting evicted and also implying
+that it has been accessed recently.
+Note that this is for heuristics only and does not reserve buffer pool
+mutex.
+@return TRUE if block is close to MRU end of LRU */
+UNIV_INLINE
+ibool
+buf_page_peek_if_young(
+/*===================*/
+ const buf_page_t* bpage); /*!< in: block */
+/********************************************************************//**
Recommends a move of a block to the start of the LRU list if there is danger
of dropping from the buffer pool. NOTE: does not reserve the buffer pool
mutex.
@@ -1196,7 +1223,7 @@ ulint
buf_get_free_list_len(void);
/*=======================*/
-/********************************************************************
+/********************************************************************//**
Determine if a block is a sentinel for a buffer pool watch.
@return TRUE if a sentinel for a buffer pool watch, FALSE if not */
UNIV_INTERN
@@ -1569,6 +1596,8 @@ struct buf_pool_stat_struct{
ulint n_pages_written;/*!< number write operations */
ulint n_pages_created;/*!< number of pages created
in the pool with no read */
+ ulint n_ra_pages_read_rnd;/*!< number of pages read in
+ as part of random read ahead */
ulint n_ra_pages_read;/*!< number of pages read in
as part of read ahead */
ulint n_ra_pages_evicted;/*!< number of read ahead
@@ -1708,7 +1737,7 @@ struct buf_pool_struct{
UT_LIST_BASE_NODE_T(buf_page_t) LRU;
/*!< base node of the LRU list */
buf_page_t* LRU_old; /*!< pointer to the about
- buf_LRU_old_ratio/BUF_LRU_OLD_RATIO_DIV
+ LRU_old_ratio/BUF_LRU_OLD_RATIO_DIV
oldest blocks in the LRU list;
NULL if LRU length less than
BUF_LRU_OLD_MIN_LEN;
@@ -1732,8 +1761,10 @@ struct buf_pool_struct{
frames and buf_page_t descriptors of blocks that exist
in the buffer pool only in compressed form. */
/* @{ */
+#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
UT_LIST_BASE_NODE_T(buf_page_t) zip_clean;
/*!< unmodified compressed pages */
+#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
UT_LIST_BASE_NODE_T(buf_page_t) zip_free[BUF_BUDDY_SIZES];
/*!< buddy free lists */
diff --git a/storage/innobase/include/buf0buf.ic b/storage/innobase/include/buf0buf.ic
index e2e83de0a78..b65b5133c15 100644
--- a/storage/innobase/include/buf0buf.ic
+++ b/storage/innobase/include/buf0buf.ic
@@ -125,6 +125,29 @@ buf_block_get_freed_page_clock(
}
/********************************************************************//**
+Tells if a block is still close enough to the MRU end of the LRU list
+meaning that it is not in danger of getting evicted and also implying
+that it has been accessed recently.
+Note that this is for heuristics only and does not reserve buffer pool
+mutex.
+@return TRUE if block is close to MRU end of LRU */
+UNIV_INLINE
+ibool
+buf_page_peek_if_young(
+/*===================*/
+ const buf_page_t* bpage) /*!< in: block */
+{
+ buf_pool_t* buf_pool = buf_pool_from_bpage(bpage);
+
+ /* FIXME: bpage->freed_page_clock is 31 bits */
+ return((buf_pool->freed_page_clock & ((1UL << 31) - 1))
+ < ((ulint) bpage->freed_page_clock
+ + (buf_pool->curr_size
+ * (BUF_LRU_OLD_RATIO_DIV - buf_pool->LRU_old_ratio)
+ / (BUF_LRU_OLD_RATIO_DIV * 4))));
+}
+
+/********************************************************************//**
Recommends a move of a block to the start of the LRU list if there is danger
of dropping from the buffer pool. NOTE: does not reserve the buffer pool
mutex.
@@ -154,12 +177,7 @@ buf_page_peek_if_too_old(
buf_pool->stat.n_pages_not_made_young++;
return(FALSE);
} else {
- /* FIXME: bpage->freed_page_clock is 31 bits */
- return((buf_pool->freed_page_clock & ((1UL << 31) - 1))
- > ((ulint) bpage->freed_page_clock
- + (buf_pool->curr_size
- * (BUF_LRU_OLD_RATIO_DIV - buf_pool->LRU_old_ratio)
- / (BUF_LRU_OLD_RATIO_DIV * 4))));
+ return(!buf_page_peek_if_young(bpage));
}
}
@@ -754,6 +772,35 @@ buf_block_get_lock_hash_val(
}
/********************************************************************//**
+Allocates a buf_page_t descriptor. This function must succeed. In case
+of failure we assert in this function.
+@return: the allocated descriptor. */
+UNIV_INLINE
+buf_page_t*
+buf_page_alloc_descriptor(void)
+/*===========================*/
+{
+ buf_page_t* bpage;
+
+ bpage = (buf_page_t*) ut_malloc(sizeof *bpage);
+ ut_d(memset(bpage, 0, sizeof *bpage));
+ UNIV_MEM_ALLOC(bpage, sizeof *bpage);
+
+ return(bpage);
+}
+
+/********************************************************************//**
+Free a buf_page_t descriptor. */
+UNIV_INLINE
+void
+buf_page_free_descriptor(
+/*=====================*/
+ buf_page_t* bpage) /*!< in: bpage descriptor to free. */
+{
+ ut_free(bpage);
+}
+
+/********************************************************************//**
Frees a buffer block which does not contain a file page. */
UNIV_INLINE
void
diff --git a/storage/innobase/include/buf0lru.h b/storage/innobase/include/buf0lru.h
index 9b150188b03..eb40621abbe 100644
--- a/storage/innobase/include/buf0lru.h
+++ b/storage/innobase/include/buf0lru.h
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2011, Oracle and/or its affiliates. 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
@@ -30,18 +30,6 @@ Created 11/5/1995 Heikki Tuuri
#include "ut0byte.h"
#include "buf0types.h"
-/** The return type of buf_LRU_free_block() */
-enum buf_lru_free_block_status {
- /** freed */
- BUF_LRU_FREED = 0,
- /** not freed because the caller asked to remove the
- uncompressed frame but the control block cannot be
- relocated */
- BUF_LRU_CANNOT_RELOCATE,
- /** not freed because of some other reason */
- BUF_LRU_NOT_FREED
-};
-
/******************************************************************//**
Tries to remove LRU flushed blocks from the end of the LRU list and put them
to the free list. This is beneficial for the efficiency of the insert buffer
@@ -85,6 +73,7 @@ void
buf_LRU_invalidate_tablespace(
/*==========================*/
ulint id); /*!< in: space id */
+#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
/********************************************************************//**
Insert a compressed block into buf_pool->zip_clean in the LRU order. */
UNIV_INTERN
@@ -92,22 +81,22 @@ void
buf_LRU_insert_zip_clean(
/*=====================*/
buf_page_t* bpage); /*!< in: pointer to the block in question */
+#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
/******************************************************************//**
Try to free a block. If bpage is a descriptor of a compressed-only
page, the descriptor object will be freed as well.
-NOTE: If this function returns BUF_LRU_FREED, it will temporarily
+NOTE: If this function returns TRUE, it will temporarily
release buf_pool->mutex. Furthermore, the page frame will no longer be
accessible via bpage.
The caller must hold buf_pool->mutex and buf_page_get_mutex(bpage) and
release these two mutexes after the call. No other
buf_page_get_mutex() may be held when calling this function.
-@return BUF_LRU_FREED if freed, BUF_LRU_CANNOT_RELOCATE or
-BUF_LRU_NOT_FREED otherwise. */
+@return TRUE if freed, FALSE otherwise. */
UNIV_INTERN
-enum buf_lru_free_block_status
+ibool
buf_LRU_free_block(
/*===============*/
buf_page_t* bpage, /*!< in: block to be freed */
@@ -195,7 +184,7 @@ buf_LRU_make_block_old(
/*===================*/
buf_page_t* bpage); /*!< in: control block */
/**********************************************************************//**
-Updates buf_LRU_old_ratio.
+Updates buf_pool->LRU_old_ratio.
@return updated old_pct */
UNIV_INTERN
ulint
@@ -204,7 +193,7 @@ buf_LRU_old_ratio_update(
uint old_pct,/*!< in: Reserve this percentage of
the buffer pool for "old" blocks. */
ibool adjust);/*!< in: TRUE=adjust the LRU list;
- FALSE=just assign buf_LRU_old_ratio
+ FALSE=just assign buf_pool->LRU_old_ratio
during the initialization of InnoDB */
/********************************************************************//**
Update the historical stats that we are collecting for LRU eviction
@@ -214,6 +203,17 @@ void
buf_LRU_stat_update(void);
/*=====================*/
+/******************************************************************//**
+Remove one page from LRU list and put it to free list */
+UNIV_INTERN
+void
+buf_LRU_free_one_page(
+/*==================*/
+ buf_page_t* bpage) /*!< in/out: block, must contain a file page and
+ be in a state where it can be freed; there
+ may or may not be a hash index to the page */
+ __attribute__((nonnull));
+
#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
/**********************************************************************//**
Validates the LRU list.
@@ -233,18 +233,15 @@ buf_LRU_print(void);
#endif /* UNIV_DEBUG_PRINT || UNIV_DEBUG || UNIV_BUF_DEBUG */
/** @name Heuristics for detecting index scan @{ */
-/** Reserve this much/BUF_LRU_OLD_RATIO_DIV of the buffer pool for
-"old" blocks. Protected by buf_pool->mutex. */
-extern uint buf_LRU_old_ratio;
-/** The denominator of buf_LRU_old_ratio. */
+/** The denominator of buf_pool->LRU_old_ratio. */
#define BUF_LRU_OLD_RATIO_DIV 1024
-/** Maximum value of buf_LRU_old_ratio.
+/** Maximum value of buf_pool->LRU_old_ratio.
@see buf_LRU_old_adjust_len
-@see buf_LRU_old_ratio_update */
+@see buf_pool->LRU_old_ratio_update */
#define BUF_LRU_OLD_RATIO_MAX BUF_LRU_OLD_RATIO_DIV
-/** Minimum value of buf_LRU_old_ratio.
+/** Minimum value of buf_pool->LRU_old_ratio.
@see buf_LRU_old_adjust_len
-@see buf_LRU_old_ratio_update
+@see buf_pool->LRU_old_ratio_update
The minimum must exceed
(BUF_LRU_OLD_TOLERANCE + 5) * BUF_LRU_OLD_RATIO_DIV / BUF_LRU_OLD_MIN_LEN. */
#define BUF_LRU_OLD_RATIO_MIN 51
diff --git a/storage/innobase/include/buf0rea.h b/storage/innobase/include/buf0rea.h
index cdf6cdba3d1..cd5eff66ee8 100644
--- a/storage/innobase/include/buf0rea.h
+++ b/storage/innobase/include/buf0rea.h
@@ -43,6 +43,31 @@ buf_read_page(
ulint zip_size,/*!< in: compressed page size in bytes, or 0 */
ulint offset);/*!< in: page number */
/********************************************************************//**
+Applies a random read-ahead in buf_pool if there are at least a threshold
+value of accessed pages from the random read-ahead area. Does not read any
+page, not even the one at the position (space, offset), if the read-ahead
+mechanism is not activated. NOTE 1: the calling thread may own latches on
+pages: to avoid deadlocks this function must be written such that it cannot
+end up waiting for these latches! NOTE 2: the calling thread must want
+access to the page given: this rule is set to prevent unintended read-aheads
+performed by ibuf routines, a situation which could result in a deadlock if
+the OS does not support asynchronous i/o.
+@return number of page read requests issued; NOTE that if we read ibuf
+pages, it may happen that the page at the given page number does not
+get read even if we return a positive value!
+@return number of page read requests issued */
+UNIV_INTERN
+ulint
+buf_read_ahead_random(
+/*==================*/
+ ulint space, /*!< in: space id */
+ ulint zip_size, /*!< in: compressed page size in bytes,
+ or 0 */
+ ulint offset, /*!< in: page number of a page which
+ the current thread wants to access */
+ ibool inside_ibuf); /*!< in: TRUE if we are inside ibuf
+ routine */
+/********************************************************************//**
Applies linear read-ahead if in the buf_pool the page is a border page of
a linear read-ahead area and all the pages in the area have been accessed.
Does not read any page if the read-ahead mechanism is not activated. Note
diff --git a/storage/innobase/include/buf0types.h b/storage/innobase/include/buf0types.h
index a2175098704..0cc2defb3ff 100644
--- a/storage/innobase/include/buf0types.h
+++ b/storage/innobase/include/buf0types.h
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2011, Oracle and/or its affiliates. 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
@@ -26,6 +26,8 @@ Created 11/17/1995 Heikki Tuuri
#ifndef buf0types_h
#define buf0types_h
+#include "page0types.h"
+
/** Buffer page (uncompressed or compressed) */
typedef struct buf_page_struct buf_page_t;
/** Buffer block for which an uncompressed page exists */
@@ -60,17 +62,10 @@ enum buf_io_fix {
/** Parameters of binary buddy system for compressed pages (buf0buddy.h) */
/* @{ */
-#if UNIV_WORD_SIZE <= 4 /* 32-bit system */
-/** Base-2 logarithm of the smallest buddy block size */
-# define BUF_BUDDY_LOW_SHIFT 6
-#else /* 64-bit system */
-/** Base-2 logarithm of the smallest buddy block size */
-# define BUF_BUDDY_LOW_SHIFT 7
-#endif
+#define BUF_BUDDY_LOW_SHIFT PAGE_ZIP_MIN_SIZE_SHIFT
+
#define BUF_BUDDY_LOW (1 << BUF_BUDDY_LOW_SHIFT)
- /*!< minimum block size in the binary
- buddy system; must be at least
- sizeof(buf_page_t) */
+
#define BUF_BUDDY_SIZES (UNIV_PAGE_SIZE_SHIFT - BUF_BUDDY_LOW_SHIFT)
/*!< number of buddy sizes */
diff --git a/storage/innobase/include/db0err.h b/storage/innobase/include/db0err.h
index 74a2354bce3..e0952f0709d 100644
--- a/storage/innobase/include/db0err.h
+++ b/storage/innobase/include/db0err.h
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2011, Oracle and/or its affiliates. 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
@@ -64,8 +64,6 @@ enum db_err {
DB_CANNOT_ADD_CONSTRAINT, /* adding a foreign key constraint
to a table failed */
DB_CORRUPTION, /* data structure corruption noticed */
- DB_COL_APPEARS_TWICE_IN_INDEX, /* InnoDB cannot handle an index
- where same column appears twice */
DB_CANNOT_DROP_CONSTRAINT, /* dropping a foreign key constraint
from a table failed */
DB_NO_SAVEPOINT, /* no savepoint exists with the given
@@ -112,6 +110,8 @@ enum db_err {
foreign keys as its prefix columns */
DB_TOO_BIG_INDEX_COL, /* index column size exceeds maximum
limit */
+ DB_INDEX_CORRUPT, /* we have corrupted index */
+ DB_UNDO_RECORD_TOO_BIG, /* the undo log record is too big */
/* The following are partial failure codes */
DB_FAIL = 1000,
diff --git a/storage/innobase/include/dict0boot.h b/storage/innobase/include/dict0boot.h
index 22df826da65..5d136862bc6 100644
--- a/storage/innobase/include/dict0boot.h
+++ b/storage/innobase/include/dict0boot.h
@@ -137,8 +137,10 @@ dict_create(void);
header is created */
/*-------------------------------------------------------------*/
-/* The field number of the page number field in the sys_indexes table
-clustered index */
+/* The field numbers in the SYS_TABLES clustered index */
+#define DICT_SYS_TABLES_TYPE_FIELD 5
+
+/* The field numbers in the SYS_INDEXES clustered index */
#define DICT_SYS_INDEXES_PAGE_NO_FIELD 8
#define DICT_SYS_INDEXES_SPACE_NO_FIELD 7
#define DICT_SYS_INDEXES_TYPE_FIELD 6
diff --git a/storage/innobase/include/dict0dict.h b/storage/innobase/include/dict0dict.h
index f979d0fcc96..57e51cbb6ba 100644
--- a/storage/innobase/include/dict0dict.h
+++ b/storage/innobase/include/dict0dict.h
@@ -585,6 +585,20 @@ dict_table_get_next_index(
# define dict_table_get_next_index(index) UT_LIST_GET_NEXT(indexes, index)
#endif /* UNIV_DEBUG */
#endif /* !UNIV_HOTBACKUP */
+
+/* Skip corrupted index */
+#define dict_table_skip_corrupt_index(index) \
+ while (index && dict_index_is_corrupted(index)) { \
+ index = dict_table_get_next_index(index); \
+ }
+
+/* Get the next non-corrupt index */
+#define dict_table_next_uncorrupted_index(index) \
+do { \
+ index = dict_table_get_next_index(index); \
+ dict_table_skip_corrupt_index(index); \
+} while (0)
+
/********************************************************************//**
Check whether the index is the clustered index.
@return nonzero for clustered index, zero for other indexes */
@@ -593,7 +607,7 @@ ulint
dict_index_is_clust(
/*================*/
const dict_index_t* index) /*!< in: index */
- __attribute__((pure));
+ __attribute__((nonnull, pure, warn_unused_result));
/********************************************************************//**
Check whether the index is unique.
@return nonzero for unique index, zero for other indexes */
@@ -602,7 +616,7 @@ ulint
dict_index_is_unique(
/*=================*/
const dict_index_t* index) /*!< in: index */
- __attribute__((pure));
+ __attribute__((nonnull, pure, warn_unused_result));
/********************************************************************//**
Check whether the index is the insert buffer tree.
@return nonzero for insert buffer, zero for other indexes */
@@ -611,7 +625,7 @@ ulint
dict_index_is_ibuf(
/*===============*/
const dict_index_t* index) /*!< in: index */
- __attribute__((pure));
+ __attribute__((nonnull, pure, warn_unused_result));
/********************************************************************//**
Check whether the index is a secondary index or the insert buffer tree.
@return nonzero for insert buffer, zero for other indexes */
@@ -620,7 +634,7 @@ ulint
dict_index_is_sec_or_ibuf(
/*======================*/
const dict_index_t* index) /*!< in: index */
- __attribute__((pure));
+ __attribute__((nonnull, pure, warn_unused_result));
/********************************************************************//**
Gets the number of user-defined columns in a table in the dictionary
@@ -630,7 +644,8 @@ UNIV_INLINE
ulint
dict_table_get_n_user_cols(
/*=======================*/
- const dict_table_t* table); /*!< in: table */
+ const dict_table_t* table) /*!< in: table */
+ __attribute__((nonnull, pure, warn_unused_result));
/********************************************************************//**
Gets the number of system columns in a table in the dictionary cache.
@return number of system (e.g., ROW_ID) columns of a table */
@@ -638,7 +653,8 @@ UNIV_INLINE
ulint
dict_table_get_n_sys_cols(
/*======================*/
- const dict_table_t* table); /*!< in: table */
+ const dict_table_t* table) /*!< in: table */
+ __attribute__((nonnull, pure, warn_unused_result));
/********************************************************************//**
Gets the number of all columns (also system) in a table in the dictionary
cache.
@@ -647,7 +663,8 @@ UNIV_INLINE
ulint
dict_table_get_n_cols(
/*==================*/
- const dict_table_t* table); /*!< in: table */
+ const dict_table_t* table) /*!< in: table */
+ __attribute__((nonnull, pure, warn_unused_result));
#ifdef UNIV_DEBUG
/********************************************************************//**
Gets the nth column of a table.
@@ -1243,6 +1260,57 @@ void
dict_close(void);
/*============*/
+/**********************************************************************//**
+Check whether the table is corrupted.
+@return nonzero for corrupted table, zero for valid tables */
+UNIV_INLINE
+ulint
+dict_table_is_corrupted(
+/*====================*/
+ const dict_table_t* table) /*!< in: table */
+ __attribute__((nonnull, pure, warn_unused_result));
+
+/**********************************************************************//**
+Check whether the index is corrupted.
+@return nonzero for corrupted index, zero for valid indexes */
+UNIV_INLINE
+ulint
+dict_index_is_corrupted(
+/*====================*/
+ const dict_index_t* index) /*!< in: index */
+ __attribute__((nonnull, pure, warn_unused_result));
+
+/**********************************************************************//**
+Flags an index and table corrupted both in the data dictionary cache
+and in the system table SYS_INDEXES. */
+UNIV_INTERN
+void
+dict_set_corrupted(
+/*===============*/
+ dict_index_t* index) /*!< in/out: index */
+ UNIV_COLD __attribute__((nonnull));
+
+/**********************************************************************//**
+Flags an index corrupted in the data dictionary cache only. This
+is used mostly to mark a corrupted index when index's own dictionary
+is corrupted, and we force to load such index for repair purpose */
+UNIV_INTERN
+void
+dict_set_corrupted_index_cache_only(
+/*================================*/
+ dict_index_t* index, /*!< in/out: index */
+ dict_table_t* table); /*!< in/out: table */
+
+/**********************************************************************//**
+Flags a table with specified space_id corrupted in the table dictionary
+cache.
+@return TRUE if successful */
+UNIV_INTERN
+ibool
+dict_set_corrupted_by_space(
+/*========================*/
+ ulint space_id); /*!< in: space ID */
+
#ifndef UNIV_NONINL
#include "dict0dict.ic"
#endif
diff --git a/storage/innobase/include/dict0dict.ic b/storage/innobase/include/dict0dict.ic
index 59811568556..ade9e627e29 100644
--- a/storage/innobase/include/dict0dict.ic
+++ b/storage/innobase/include/dict0dict.ic
@@ -27,6 +27,7 @@ Created 1/8/1996 Heikki Tuuri
#ifndef UNIV_HOTBACKUP
#include "dict0load.h"
#include "rem0types.h"
+#include "srv0srv.h"
/*********************************************************************//**
Gets the minimum number of bytes per character.
@@ -828,7 +829,7 @@ dict_table_check_if_in_cache_low(
}
/**********************************************************************//**
-load a table into dictionary cache, ignore any error specified during load;
+load a table into dictionary cache, ignore any error specified during load;
@return table, NULL if not found */
UNIV_INLINE
dict_table_t*
@@ -872,6 +873,18 @@ dict_table_get_low(
table = dict_table_check_if_in_cache_low(table_name);
+ if (table && table->corrupted) {
+ fprintf(stderr, "InnoDB: table");
+ ut_print_name(stderr, NULL, TRUE, table->name);
+ if (srv_load_corrupted) {
+ fputs(" is corrupted, but"
+ " innodb_force_load_corrupted is set\n", stderr);
+ } else {
+ fputs(" is corrupted\n", stderr);
+ return(NULL);
+ }
+ }
+
if (table == NULL) {
table = dict_load_table(table_name, TRUE, DICT_ERR_IGNORE_NONE);
}
@@ -937,4 +950,35 @@ dict_max_field_len_store_undo(
return(prefix_len);
}
+/********************************************************************//**
+Check whether the table is corrupted.
+@return nonzero for corrupted table, zero for valid tables */
+UNIV_INLINE
+ulint
+dict_table_is_corrupted(
+/*====================*/
+ const dict_table_t* table) /*!< in: table */
+{
+ ut_ad(table);
+ ut_ad(table->magic_n == DICT_TABLE_MAGIC_N);
+
+ return(UNIV_UNLIKELY(table->corrupted));
+}
+
+/********************************************************************//**
+Check whether the index is corrupted.
+@return nonzero for corrupted index, zero for valid indexes */
+UNIV_INLINE
+ulint
+dict_index_is_corrupted(
+/*====================*/
+ const dict_index_t* index) /*!< in: index */
+{
+ ut_ad(index);
+ ut_ad(index->magic_n == DICT_INDEX_MAGIC_N);
+
+ return(UNIV_UNLIKELY((index->type & DICT_CORRUPT)
+ || (index->table && index->table->corrupted)));
+}
+
#endif /* !UNIV_HOTBACKUP */
diff --git a/storage/innobase/include/dict0mem.h b/storage/innobase/include/dict0mem.h
index 3a475fa85fc..9ded0dba39b 100644
--- a/storage/innobase/include/dict0mem.h
+++ b/storage/innobase/include/dict0mem.h
@@ -51,7 +51,12 @@ combination of types */
#define DICT_UNIQUE 2 /*!< unique index */
#define DICT_UNIVERSAL 4 /*!< index which can contain records from any
other index */
-#define DICT_IBUF 8 /*!< insert buffer tree */
+#define DICT_IBUF 8 /*!< insert buffer tree */
+#define DICT_CORRUPT 16 /*!< bit to store the corrupted flag
+ in SYS_INDEXES.TYPE */
+
+#define DICT_IT_BITS 5 /*!< number of bits used for
+ SYS_INDEXES.TYPE */
/* @} */
/** Types for a table object */
@@ -369,8 +374,9 @@ struct dict_index_struct{
/*!< space where the index tree is placed */
unsigned page:32;/*!< index tree root page number */
#endif /* !UNIV_HOTBACKUP */
- unsigned type:4; /*!< index type (DICT_CLUSTERED, DICT_UNIQUE,
- DICT_UNIVERSAL, DICT_IBUF) */
+ unsigned type:DICT_IT_BITS;
+ /*!< index type (DICT_CLUSTERED, DICT_UNIQUE,
+ DICT_UNIVERSAL, DICT_IBUF, DICT_CORRUPT) */
unsigned trx_id_offset:10;/*!< position of the trx id column
in a clustered index record, if the fields
before it are known to be of a fixed size,
@@ -391,8 +397,6 @@ struct dict_index_struct{
/*!< TRUE if this index is marked to be
dropped in ha_innobase::prepare_drop_index(),
otherwise FALSE */
- unsigned corrupted:1;
- /*!< TRUE if the index object is corrupted */
dict_field_t* fields; /*!< array of field descriptions */
#ifndef UNIV_HOTBACKUP
UT_LIST_NODE_T(dict_index_t)
diff --git a/storage/innobase/include/dict0types.h b/storage/innobase/include/dict0types.h
index 8cbd7cd5783..f0a05a38070 100644
--- a/storage/innobase/include/dict0types.h
+++ b/storage/innobase/include/dict0types.h
@@ -51,7 +51,8 @@ be or-ed together */
enum dict_err_ignore {
DICT_ERR_IGNORE_NONE = 0, /*!< no error to ignore */
DICT_ERR_IGNORE_INDEX_ROOT = 1, /*!< ignore error if index root
- page is FIL_NUL or incorrect value */
+ page is FIL_NULL or incorrect value */
+ DICT_ERR_IGNORE_CORRUPT = 2, /*!< skip corrupted indexes */
DICT_ERR_IGNORE_ALL = 0xFFFF /*!< ignore all errors */
};
diff --git a/storage/innobase/include/mtr0mtr.h b/storage/innobase/include/mtr0mtr.h
index 5582ad63039..7f608546cc2 100644
--- a/storage/innobase/include/mtr0mtr.h
+++ b/storage/innobase/include/mtr0mtr.h
@@ -213,16 +213,6 @@ ulint
mtr_set_savepoint(
/*==============*/
mtr_t* mtr); /*!< in: mtr */
-/**********************************************************//**
-Releases the latches stored in an mtr memo down to a savepoint.
-NOTE! The mtr must not have made changes to buffer pages after the
-savepoint, as these can be handled only by mtr_commit. */
-UNIV_INTERN
-void
-mtr_rollback_to_savepoint(
-/*======================*/
- mtr_t* mtr, /*!< in: mtr */
- ulint savepoint); /*!< in: savepoint */
#ifndef UNIV_HOTBACKUP
/**********************************************************//**
Releases the (index tree) s-latch stored in an mtr memo after a
diff --git a/storage/innobase/include/page0page.h b/storage/innobase/include/page0page.h
index 7091d8c2707..6a82e820312 100644
--- a/storage/innobase/include/page0page.h
+++ b/storage/innobase/include/page0page.h
@@ -68,10 +68,7 @@ typedef byte page_header_t;
#define PAGE_MAX_TRX_ID 18 /* highest id of a trx which may have modified
a record on the page; trx_id_t; defined only
in secondary indexes and in the insert buffer
- tree; NOTE: this may be modified only
- when the thread has an x-latch to the page,
- and ALSO an x-latch to btr_search_latch
- if there is a hash index to the page! */
+ tree */
#define PAGE_HEADER_PRIV_END 26 /* end of private data structure of the page
header which are set in a page create */
/*----*/
diff --git a/storage/innobase/include/row0row.h b/storage/innobase/include/row0row.h
index 75e15d67246..c2849be7c3e 100644
--- a/storage/innobase/include/row0row.h
+++ b/storage/innobase/include/row0row.h
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2011, Oracle and/or its affiliates. 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
@@ -38,27 +38,16 @@ Created 4/20/1996 Heikki Tuuri
#include "btr0types.h"
/*********************************************************************//**
-Gets the offset of the trx id field, in bytes relative to the origin of
+Gets the offset of the DB_TRX_ID field, in bytes relative to the origin of
a clustered index record.
@return offset of DATA_TRX_ID */
UNIV_INLINE
ulint
-row_get_trx_id_offset_func(
-/*=======================*/
-#ifdef UNIV_DEBUG
- const rec_t* rec, /*!< in: record */
-#endif /* UNIV_DEBUG */
+row_get_trx_id_offset(
+/*==================*/
const dict_index_t* index, /*!< in: clustered index */
- const ulint* offsets)/*!< in: rec_get_offsets(rec, index) */
+ const ulint* offsets)/*!< in: record offsets */
__attribute__((nonnull, warn_unused_result));
-#ifdef UNIV_DEBUG
-# define row_get_trx_id_offset(rec, index, offsets) \
- row_get_trx_id_offset_func(rec, index, offsets)
-#else /* UNIV_DEBUG */
-# define row_get_trx_id_offset(rec, index, offsets) \
- row_get_trx_id_offset_func(index, offsets)
-#endif /* UNIV_DEBUG */
-
/*********************************************************************//**
Reads the trx id field from a clustered index record.
@return value of the field */
@@ -77,9 +66,10 @@ UNIV_INLINE
roll_ptr_t
row_get_rec_roll_ptr(
/*=================*/
- const rec_t* rec, /*!< in: record */
- dict_index_t* index, /*!< in: clustered index */
- const ulint* offsets);/*!< in: rec_get_offsets(rec, index) */
+ const rec_t* rec, /*!< in: record */
+ const dict_index_t* index, /*!< in: clustered index */
+ const ulint* offsets)/*!< in: rec_get_offsets(rec, index) */
+ __attribute__((nonnull, warn_unused_result));
/*****************************************************************//**
When an insert or purge to a table is performed, this function builds
the entry to be inserted into or purged from an index on the table.
diff --git a/storage/innobase/include/row0row.ic b/storage/innobase/include/row0row.ic
index 9d19e430e16..0b9ca982af8 100644
--- a/storage/innobase/include/row0row.ic
+++ b/storage/innobase/include/row0row.ic
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2011, Oracle and/or its affiliates. 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
@@ -28,25 +28,22 @@ Created 4/20/1996 Heikki Tuuri
#include "trx0undo.h"
/*********************************************************************//**
-Gets the offset of trx id field, in bytes relative to the origin of
+Gets the offset of the DB_TRX_ID field, in bytes relative to the origin of
a clustered index record.
@return offset of DATA_TRX_ID */
UNIV_INLINE
ulint
-row_get_trx_id_offset_func(
-/*=======================*/
-#ifdef UNIV_DEBUG
- const rec_t* rec, /*!< in: record */
-#endif /* UNIV_DEBUG */
+row_get_trx_id_offset(
+/*==================*/
const dict_index_t* index, /*!< in: clustered index */
- const ulint* offsets)/*!< in: rec_get_offsets(rec, index) */
+ const ulint* offsets)/*!< in: record offsets */
{
ulint pos;
ulint offset;
ulint len;
ut_ad(dict_index_is_clust(index));
- ut_ad(rec_offs_validate(rec, index, offsets));
+ ut_ad(rec_offs_validate(NULL, index, offsets));
pos = dict_index_get_sys_col_pos(index, DATA_TRX_ID);
@@ -76,7 +73,7 @@ row_get_rec_trx_id(
offset = index->trx_id_offset;
if (!offset) {
- offset = row_get_trx_id_offset(rec, index, offsets);
+ offset = row_get_trx_id_offset(index, offsets);
}
return(trx_read_trx_id(rec + offset));
@@ -89,9 +86,9 @@ UNIV_INLINE
roll_ptr_t
row_get_rec_roll_ptr(
/*=================*/
- const rec_t* rec, /*!< in: record */
- dict_index_t* index, /*!< in: clustered index */
- const ulint* offsets)/*!< in: rec_get_offsets(rec, index) */
+ const rec_t* rec, /*!< in: record */
+ const dict_index_t* index, /*!< in: clustered index */
+ const ulint* offsets)/*!< in: rec_get_offsets(rec, index) */
{
ulint offset;
@@ -101,7 +98,7 @@ row_get_rec_roll_ptr(
offset = index->trx_id_offset;
if (!offset) {
- offset = row_get_trx_id_offset(rec, index, offsets);
+ offset = row_get_trx_id_offset(index, offsets);
}
return(trx_read_roll_ptr(rec + offset + DATA_TRX_ID_LEN));
diff --git a/storage/innobase/include/row0upd.ic b/storage/innobase/include/row0upd.ic
index 18e22f1eca9..11db82f64da 100644
--- a/storage/innobase/include/row0upd.ic
+++ b/storage/innobase/include/row0upd.ic
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2011, Oracle and/or its affiliates. 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
@@ -157,11 +157,6 @@ row_upd_rec_sys_fields(
{
ut_ad(dict_index_is_clust(index));
ut_ad(rec_offs_validate(rec, index, offsets));
-#ifdef UNIV_SYNC_DEBUG
- if (!rw_lock_own(&btr_search_latch, RW_LOCK_EX)) {
- ut_ad(!buf_block_align(rec)->is_hashed);
- }
-#endif /* UNIV_SYNC_DEBUG */
if (UNIV_LIKELY_NULL(page_zip)) {
ulint pos = dict_index_get_sys_col_pos(index, DATA_TRX_ID);
@@ -171,7 +166,7 @@ row_upd_rec_sys_fields(
ulint offset = index->trx_id_offset;
if (!offset) {
- offset = row_get_trx_id_offset(rec, index, offsets);
+ offset = row_get_trx_id_offset(index, offsets);
}
#if DATA_TRX_ID + 1 != DATA_ROLL_PTR
diff --git a/storage/innobase/include/srv0srv.h b/storage/innobase/include/srv0srv.h
index fc9401a12f5..dfe7397d189 100644
--- a/storage/innobase/include/srv0srv.h
+++ b/storage/innobase/include/srv0srv.h
@@ -141,6 +141,10 @@ extern ulint srv_log_buffer_size;
extern ulong srv_flush_log_at_trx_commit;
extern char srv_adaptive_flushing;
+/* If this flag is TRUE, then we will load the indexes' (and tables') metadata
+even if they are marked as "corrupted". Mostly it is for DBA to process
+corrupted index and table */
+extern my_bool srv_load_corrupted;
/* The sort order table of the MySQL latin1_swedish_ci character set
collation */
@@ -158,6 +162,7 @@ extern ulint srv_mem_pool_size;
extern ulint srv_lock_table_size;
extern ulint srv_n_file_io_threads;
+extern my_bool srv_random_read_ahead;
extern ulong srv_read_ahead_threshold;
extern ulint srv_n_read_io_threads;
extern ulint srv_n_write_io_threads;
@@ -703,6 +708,7 @@ struct export_var_struct{
ulint innodb_buffer_pool_wait_free; /*!< srv_buf_pool_wait_free */
ulint innodb_buffer_pool_pages_flushed; /*!< srv_buf_pool_flushed */
ulint innodb_buffer_pool_write_requests;/*!< srv_buf_pool_write_requests */
+ ulint innodb_buffer_pool_read_ahead_rnd;/*!< srv_read_ahead_rnd */
ulint innodb_buffer_pool_read_ahead; /*!< srv_read_ahead */
ulint innodb_buffer_pool_read_ahead_evicted;/*!< srv_read_ahead evicted*/
ulint innodb_dblwr_pages_written; /*!< srv_dblwr_pages_written */
diff --git a/storage/innobase/include/sync0rw.h b/storage/innobase/include/sync0rw.h
index 9e5c96fa323..dedf080d00a 100644
--- a/storage/innobase/include/sync0rw.h
+++ b/storage/innobase/include/sync0rw.h
@@ -542,7 +542,7 @@ mutex. */
UNIV_INTERN
void
rw_lock_debug_mutex_enter(void);
-/*==========================*/
+/*===========================*/
/******************************************************************//**
Releases the debug mutex. */
UNIV_INTERN
@@ -632,7 +632,8 @@ struct rw_lock_struct {
};
#ifdef UNIV_SYNC_DEBUG
-/** The structure for storing debug info of an rw-lock */
+/** The structure for storing debug info of an rw-lock. All access to this
+structure must be protected by rw_lock_debug_mutex_enter(). */
struct rw_lock_debug_struct {
os_thread_id_t thread_id; /*!< The thread id of the thread which
diff --git a/storage/innobase/include/sync0rw.ic b/storage/innobase/include/sync0rw.ic
index 2ffd9fdafb5..51cf7121b4c 100644
--- a/storage/innobase/include/sync0rw.ic
+++ b/storage/innobase/include/sync0rw.ic
@@ -406,6 +406,7 @@ rw_lock_s_lock_func(
#ifdef UNIV_SYNC_DEBUG
ut_ad(!rw_lock_own(lock, RW_LOCK_SHARED)); /* see NOTE above */
+ ut_ad(!rw_lock_own(lock, RW_LOCK_EX));
#endif /* UNIV_SYNC_DEBUG */
/* TODO: study performance of UNIV_LIKELY branch prediction hints. */
diff --git a/storage/innobase/include/sync0sync.h b/storage/innobase/include/sync0sync.h
index b823c9d5259..d9dea0aa63d 100644
--- a/storage/innobase/include/sync0sync.h
+++ b/storage/innobase/include/sync0sync.h
@@ -636,10 +636,6 @@ or row lock! */
#define SYNC_DICT_HEADER 995
#define SYNC_IBUF_HEADER 914
#define SYNC_IBUF_PESS_INSERT_MUTEX 912
-#define SYNC_IBUF_MUTEX 910 /* ibuf mutex is really below
- SYNC_FSP_PAGE: we assign a value this
- high only to make the program to pass
- the debug checks */
/*-------------------------------*/
#define SYNC_INDEX_TREE 900
#define SYNC_TREE_NODE_NEW 892
@@ -655,8 +651,11 @@ or row lock! */
#define SYNC_FSP 400
#define SYNC_FSP_PAGE 395
/*------------------------------------- Insert buffer headers */
-/*------------------------------------- ibuf_mutex */
+#define SYNC_IBUF_MUTEX 370 /* ibuf_mutex */
/*------------------------------------- Insert buffer tree */
+#define SYNC_IBUF_INDEX_TREE 360
+#define SYNC_IBUF_TREE_NODE_NEW 359
+#define SYNC_IBUF_TREE_NODE 358
#define SYNC_IBUF_BITMAP_MUTEX 351
#define SYNC_IBUF_BITMAP 350
/*------------------------------------- MySQL query cache mutex */
@@ -683,7 +682,6 @@ or row lock! */
#define SYNC_BUF_FLUSH_LIST 145 /* Buffer flush list mutex */
#define SYNC_DOUBLEWRITE 140
#define SYNC_ANY_LATCH 135
-#define SYNC_THR_LOCAL 133
#define SYNC_MEM_HASH 131
#define SYNC_MEM_POOL 130
diff --git a/storage/innobase/include/sync0sync.ic b/storage/innobase/include/sync0sync.ic
index 2977f71154a..eb21f44c65e 100644
--- a/storage/innobase/include/sync0sync.ic
+++ b/storage/innobase/include/sync0sync.ic
@@ -272,11 +272,10 @@ pfs_mutex_enter_nowait_func(
ulint ret;
struct PSI_mutex_locker* locker = NULL;
PSI_mutex_locker_state state;
- int result = 0;
if (UNIV_LIKELY(PSI_server && mutex->pfs_psi)) {
locker = PSI_server->get_thread_mutex_locker(
- &state, mutex->pfs_psi, PSI_MUTEX_LOCK);
+ &state, mutex->pfs_psi, PSI_MUTEX_TRYLOCK);
if (locker) {
PSI_server->start_mutex_wait(locker, file_name, line);
}
@@ -285,7 +284,7 @@ pfs_mutex_enter_nowait_func(
ret = mutex_enter_nowait_func(mutex, file_name, line);
if (locker) {
- PSI_server->end_mutex_wait(locker, result);
+ PSI_server->end_mutex_wait(locker, ret);
}
return(ret);
diff --git a/storage/innobase/include/trx0sys.h b/storage/innobase/include/trx0sys.h
index dc0ca2285b9..9cec1933e9d 100644
--- a/storage/innobase/include/trx0sys.h
+++ b/storage/innobase/include/trx0sys.h
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1996, 2011, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2011, Oracle and/or its affiliates. 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
@@ -278,6 +278,17 @@ ibool
trx_in_trx_list(
/*============*/
trx_t* in_trx);/*!< in: trx */
+#if defined UNIV_DEBUG || defined UNIV_BLOB_LIGHT_DEBUG
+/***********************************************************//**
+Assert that a transaction has been recovered.
+@return TRUE */
+UNIV_INLINE
+ibool
+trx_assert_recovered(
+/*=================*/
+ trx_id_t trx_id) /*!< in: transaction identifier */
+ __attribute__((warn_unused_result));
+#endif /* UNIV_DEBUG || UNIV_BLOB_LIGHT_DEBUG */
/*****************************************************************//**
Updates the offset information about the end of the MySQL binlog entry
which corresponds to the transaction just being committed. In a MySQL
diff --git a/storage/innobase/include/trx0sys.ic b/storage/innobase/include/trx0sys.ic
index 355f118a1ec..5e702b25325 100644
--- a/storage/innobase/include/trx0sys.ic
+++ b/storage/innobase/include/trx0sys.ic
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2011, Oracle and/or its affiliates. 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
@@ -277,6 +277,28 @@ trx_get_on_id(
return(NULL);
}
+#if defined UNIV_DEBUG || defined UNIV_BLOB_LIGHT_DEBUG
+/***********************************************************//**
+Assert that a transaction has been recovered.
+@return TRUE */
+UNIV_INLINE
+ibool
+trx_assert_recovered(
+/*=================*/
+ trx_id_t trx_id) /*!< in: transaction identifier */
+{
+ trx_t* trx;
+
+ mutex_enter(&kernel_mutex);
+ trx = trx_get_on_id(trx_id);
+ ut_a(trx);
+ ut_a(trx->is_recovered);
+ mutex_exit(&kernel_mutex);
+
+ return(TRUE);
+}
+#endif /* UNIV_DEBUG || UNIV_BLOB_LIGHT_DEBUG */
+
/****************************************************************//**
Returns the minumum trx id in trx list. This is the smallest id for which
the trx can possibly be active. (But, you must look at the trx->conc_state to
diff --git a/storage/innobase/include/trx0undo.h b/storage/innobase/include/trx0undo.h
index df16c939070..50aa6d0ac09 100644
--- a/storage/innobase/include/trx0undo.h
+++ b/storage/innobase/include/trx0undo.h
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2011, Oracle and/or its affiliates. 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
@@ -204,17 +204,51 @@ trx_undo_add_page(
mtr_t* mtr); /*!< in: mtr which does not have a latch to any
undo log page; the caller must have reserved
the rollback segment mutex */
+/********************************************************************//**
+Frees the last undo log page.
+The caller must hold the rollback segment mutex. */
+UNIV_INTERN
+void
+trx_undo_free_last_page_func(
+/*==========================*/
+#ifdef UNIV_DEBUG
+ const trx_t* trx, /*!< in: transaction */
+#endif /* UNIV_DEBUG */
+ trx_undo_t* undo, /*!< in/out: undo log memory copy */
+ mtr_t* mtr) /*!< in/out: mini-transaction which does not
+ have a latch to any undo log page or which
+ has allocated the undo log page */
+ __attribute__((nonnull));
+#ifdef UNIV_DEBUG
+# define trx_undo_free_last_page(trx,undo,mtr) \
+ trx_undo_free_last_page_func(trx,undo,mtr)
+#else /* UNIV_DEBUG */
+# define trx_undo_free_last_page(trx,undo,mtr) \
+ trx_undo_free_last_page_func(undo,mtr)
+#endif /* UNIV_DEBUG */
+
/***********************************************************************//**
Truncates an undo log from the end. This function is used during a rollback
to free space from an undo log. */
UNIV_INTERN
void
-trx_undo_truncate_end(
-/*==================*/
- trx_t* trx, /*!< in: transaction whose undo log it is */
- trx_undo_t* undo, /*!< in: undo log */
- undo_no_t limit); /*!< in: all undo records with undo number
+trx_undo_truncate_end_func(
+/*=======================*/
+#ifdef UNIV_DEBUG
+ const trx_t* trx, /*!< in: transaction whose undo log it is */
+#endif /* UNIV_DEBUG */
+ trx_undo_t* undo, /*!< in/out: undo log */
+ undo_no_t limit) /*!< in: all undo records with undo number
>= this value should be truncated */
+ __attribute__((nonnull));
+#ifdef UNIV_DEBUG
+# define trx_undo_truncate_end(trx,undo,limit) \
+ trx_undo_truncate_end_func(trx,undo,limit)
+#else /* UNIV_DEBUG */
+# define trx_undo_truncate_end(trx,undo,limit) \
+ trx_undo_truncate_end_func(undo,limit)
+#endif /* UNIV_DEBUG */
+
/***********************************************************************//**
Truncates an undo log from the start. This function is used during a purge
operation. */
diff --git a/storage/innobase/include/ut0mem.h b/storage/innobase/include/ut0mem.h
index 57dfb08f41c..faf6f242883 100644
--- a/storage/innobase/include/ut0mem.h
+++ b/storage/innobase/include/ut0mem.h
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2011, Oracle and/or its affiliates. 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
@@ -210,43 +210,6 @@ ut_strlcpy_rev(
ulint size); /*!< in: size of destination buffer */
/**********************************************************************//**
-Compute strlen(ut_strcpyq(str, q)).
-@return length of the string when quoted */
-UNIV_INLINE
-ulint
-ut_strlenq(
-/*=======*/
- const char* str, /*!< in: null-terminated string */
- char q); /*!< in: the quote character */
-
-/**********************************************************************//**
-Make a quoted copy of a NUL-terminated string. Leading and trailing
-quotes will not be included; only embedded quotes will be escaped.
-See also ut_strlenq() and ut_memcpyq().
-@return pointer to end of dest */
-UNIV_INTERN
-char*
-ut_strcpyq(
-/*=======*/
- char* dest, /*!< in: output buffer */
- char q, /*!< in: the quote character */
- const char* src); /*!< in: null-terminated string */
-
-/**********************************************************************//**
-Make a quoted copy of a fixed-length string. Leading and trailing
-quotes will not be included; only embedded quotes will be escaped.
-See also ut_strlenq() and ut_strcpyq().
-@return pointer to end of dest */
-UNIV_INTERN
-char*
-ut_memcpyq(
-/*=======*/
- char* dest, /*!< in: output buffer */
- char q, /*!< in: the quote character */
- const char* src, /*!< in: string to be quoted */
- ulint len); /*!< in: length of src */
-
-/**********************************************************************//**
Return the number of times s2 occurs in s1. Overlapping instances of s2
are only counted once.
@return the number of times s2 occurs in s1 */
diff --git a/storage/innobase/include/ut0mem.ic b/storage/innobase/include/ut0mem.ic
index f36c28f1989..c06e2b3ae81 100644
--- a/storage/innobase/include/ut0mem.ic
+++ b/storage/innobase/include/ut0mem.ic
@@ -1,6 +1,6 @@
/*****************************************************************************
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2011, Oracle and/or its affiliates. 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
@@ -99,27 +99,6 @@ ut_strcmp(const char* str1, const char* str2)
}
/**********************************************************************//**
-Compute strlen(ut_strcpyq(str, q)).
-@return length of the string when quoted */
-UNIV_INLINE
-ulint
-ut_strlenq(
-/*=======*/
- const char* str, /*!< in: null-terminated string */
- char q) /*!< in: the quote character */
-{
- ulint len;
-
- for (len = 0; *str; len++, str++) {
- if (*str == q) {
- len++;
- }
- }
-
- return(len);
-}
-
-/**********************************************************************//**
Converts a raw binary data to a NUL-terminated hex string. The output is
truncated if there is not enough space in "hex", make sure "hex_size" is at
least (2 * raw_size + 1) if you do not want this to happen. Returns the