/****************************************************** Insert buffer (c) 1997 Innobase Oy Created 7/19/1997 Heikki Tuuri *******************************************************/ #include "ibuf0ibuf.h" #ifdef UNIV_NONINL #include "ibuf0ibuf.ic" #endif #include "buf0buf.h" #include "buf0rea.h" #include "fsp0fsp.h" #include "trx0sys.h" #include "fil0fil.h" #include "thr0loc.h" #include "rem0rec.h" #include "btr0cur.h" #include "btr0pcur.h" #include "btr0btr.h" #include "sync0sync.h" #include "dict0boot.h" #include "fut0lst.h" #include "lock0lock.h" #include "log0recv.h" #include "que0que.h" /* PREVENTING DEADLOCKS IN THE INSERT BUFFER SYSTEM If an OS thread performs any operation that brings in disk pages from non-system tablespaces into the buffer pool, or creates such a page there, then the operation may have as a side effect an insert buffer index tree compression. Thus, the tree latch of the insert buffer tree may be acquired in the x-mode, and also the file space latch of the system tablespace may be acquired in the x-mode. Also, an insert to an index in a non-system tablespace can have the same effect. How do we know this cannot lead to a deadlock of OS threads? There is a problem with the i\o-handler threads: they break the latching order because they own x-latches to pages which are on a lower level than the insert buffer tree latch, its page latches, and the tablespace latch an insert buffer operation can reserve. The solution is the following: We put into each tablespace an insert buffer of its own. Let all the tree and page latches connected with the insert buffer be later in the latching order than the fsp latch and fsp page latches. Insert buffer pages must be such that the insert buffer is never invoked when these pages area accessed as this would result in a recursion violating the latching order. We let a special i/o-handler thread take care of i/o to the insert buffer pages and the ibuf bitmap pages, as well as the fsp bitmap pages and the first inode page, which contains the inode of the ibuf tree: let us call all these ibuf pages. If the OS does not support asynchronous i/o, then there is no special i/o thread, but to prevent deadlocks, we do not let a read-ahead access both non-ibuf and ibuf pages. Then an i/o-handler for the insert buffer never needs to access the insert buffer tree and thus obeys the latching order. On the other hand, other i/o-handlers for other tablespaces may require access to the insert buffer, but because all kinds of latches they need to access there are later in the latching order, no violation of the latching order occurs in this case, either. A problem is how to grow and contract an insert buffer tree. As it is later in the latching order than the fsp management, we have to reserve the fsp latch first, before adding or removing pages from the insert buffer tree. We let the insert buffer tree have its own file space management: a free list of pages linked to the tree root. To prevent recursive using of the insert buffer when adding pages to the tree, we must first load these pages to memory, obtaining a latch on them, and only after that add them to the free list of the insert buffer tree. More difficult is removing of pages from the free list. If there is an excess of pages in the free list of the ibuf tree, they might be needed if some thread reserves the fsp latch, intending to allocate more file space. So we do the following: if a thread reserves the fsp latch, we check the writer count field of the latch. If this field has value 1, it means that the thread did not own the latch before entering the fsp system, and the mtr of the thread contains no modifications to the fsp pages. Now we are free to reserve the ibuf latch, and check if there is an excess of pages in the free list. We can then, in a separate mini-transaction, take them out of the free list and free them to the fsp system. To avoid deadlocks in the ibuf system, we divide file pages into three levels: (1) non-ibuf pages, (2) ibuf tree pages and the pages in the ibuf tree free list, and (3) ibuf bitmap pages. No OS thread is allowed to access higher level pages if it has latches to lower level pages; even if the thread owns a B-tree latch it must not access the B-tree non-leaf pages if it has latches on lower level pages. Read-ahead is only allowed for level 1 and 2 pages. Dedicated i/o-handler threads handle exclusively level 1 i/o. A dedicated i/o handler thread handles exclusively level 2 i/o. However, if an OS thread does the i/o handling for itself, i.e., it uses synchronous aio or the OS does not support aio, it can access any pages, as long as it obeys the access order rules. */ /* Buffer pool size per the maximum insert buffer size */ #define IBUF_POOL_SIZE_PER_MAX_SIZE 2 /* The insert buffer control structure */ ibuf_t* ibuf = NULL; ulint ibuf_rnd = 986058871; ulint ibuf_flush_count = 0; /* Dimensions for the ibuf_count array */ #define IBUF_COUNT_N_SPACES 10 #define IBUF_COUNT_N_PAGES 10000 /* Buffered entry counts for file pages, used in debugging */ ulint* ibuf_counts[IBUF_COUNT_N_SPACES]; ibool ibuf_counts_inited = FALSE; /* The start address for an insert buffer bitmap page bitmap */ #define IBUF_BITMAP PAGE_DATA /* Offsets in bits for the bits describing a single page in the bitmap */ #define IBUF_BITMAP_FREE 0 #define IBUF_BITMAP_BUFFERED 2 #define IBUF_BITMAP_IBUF 3 /* TRUE if page is a part of the ibuf tree, excluding the root page, or is in the free list of the ibuf */ /* Number of bits describing a single page */ #define IBUF_BITS_PER_PAGE 4 /* The mutex used to block pessimistic inserts to ibuf trees */ mutex_t ibuf_pessimistic_insert_mutex; /* The mutex protecting the insert buffer structs */ mutex_t ibuf_mutex; /* The mutex protecting the insert buffer bitmaps */ mutex_t ibuf_bitmap_mutex; /* The area in pages from which contract looks for page numbers for merge */ #define IBUF_MERGE_AREA 8 /* Inside the merge area, pages which have at most 1 per this number less buffered entries compared to maximum volume that can buffered for a single page are merged along with the page whose buffer became full */ #define IBUF_MERGE_THRESHOLD 4 /* In ibuf_contract at most this number of pages is read to memory in one batch, in order to merge the entries for them in the insert buffer */ #define IBUF_MAX_N_PAGES_MERGED IBUF_MERGE_AREA /* If the combined size of the ibuf trees exceeds ibuf->max_size by this many pages, we start to contract it in connection to inserts there, using non-synchronous contract */ #define IBUF_CONTRACT_ON_INSERT_NON_SYNC 0 /* Same as above, but use synchronous contract */ #define IBUF_CONTRACT_ON_INSERT_SYNC 5 /* Same as above, but no insert is done, only contract is called */ #define IBUF_CONTRACT_DO_NOT_INSERT 10 /* TODO: how to cope with drop table if there are records in the insert buffer for the indexes of the table? Is there actually any problem, because ibuf merge is done to a page when it is read in, and it is still physically like the index page even if the index would have been dropped! So, there seems to be no problem. */ /********************************************************************** Validates the ibuf data structures when the caller owns ibuf_mutex. */ static ibool ibuf_validate_low(void); /*===================*/ /* out: TRUE if ok */ /********************************************************************** Sets the flag in the current OS thread local storage denoting that it is inside an insert buffer routine. */ UNIV_INLINE void ibuf_enter(void) /*============*/ { ibool* ptr; ptr = thr_local_get_in_ibuf_field(); ut_ad(*ptr == FALSE); *ptr = TRUE; } /********************************************************************** Sets the flag in the current OS thread local storage denoting that it is exiting an insert buffer routine. */ UNIV_INLINE void ibuf_exit(void) /*===========*/ { ibool* ptr; ptr = thr_local_get_in_ibuf_field(); ut_ad(*ptr == TRUE); *ptr = FALSE; } /********************************************************************** Returns TRUE if the current OS thread is performing an insert buffer routine. */ ibool ibuf_inside(void) /*=============*/ /* out: TRUE if inside an insert buffer routine: for instance, a read-ahead of non-ibuf pages is then forbidden */ { return(*thr_local_get_in_ibuf_field()); } /********************************************************************** Gets the ibuf header page and x-latches it. */ static page_t* ibuf_header_page_get( /*=================*/ /* out: insert buffer header page */ ulint space, /* in: space id */ mtr_t* mtr) /* in: mtr */ { page_t* page; ut_ad(!ibuf_inside()); page = buf_page_get(space, FSP_IBUF_HEADER_PAGE_NO, RW_X_LATCH, mtr); buf_page_dbg_add_level(page, SYNC_IBUF_HEADER); return(page); } /********************************************************************** Gets the root page and x-latches it. */ static page_t* ibuf_tree_root_get( /*===============*/ /* out: insert buffer tree root page */ ibuf_data_t* data, /* in: ibuf data */ ulint space, /* in: space id */ mtr_t* mtr) /* in: mtr */ { page_t* page; ut_ad(ibuf_inside()); mtr_x_lock(dict_tree_get_lock((data->index)->tree), mtr); page = buf_page_get(space, FSP_IBUF_TREE_ROOT_PAGE_NO, RW_X_LATCH, mtr); buf_page_dbg_add_level(page, SYNC_TREE_NODE); return(page); } /********************************************************************** Gets the ibuf count for a given page. */ ulint ibuf_count_get( /*===========*/ /* out: number of entries in the insert buffer currently buffered for this page */ ulint space, /* in: space id */ ulint page_no)/* in: page number */ { ut_ad(space < IBUF_COUNT_N_SPACES); ut_ad(page_no < IBUF_COUNT_N_PAGES); if (!ibuf_counts_inited) { return(0); } return(*(ibuf_counts[space] + page_no)); } /********************************************************************** Sets the ibuf count for a given page. */ static void ibuf_count_set( /*===========*/ ulint space, /* in: space id */ ulint page_no,/* in: page number */ ulint val) /* in: value to set */ { ut_ad(space < IBUF_COUNT_N_SPACES); ut_ad(page_no < IBUF_COUNT_N_PAGES); ut_ad(val < UNIV_PAGE_SIZE); *(ibuf_counts[space] + page_no) = val; } /********************************************************************** Creates the insert buffer data structure at a database startup and initializes the data structures for the insert buffer of each tablespace. */ void ibuf_init_at_db_start(void) /*=======================*/ { ibuf = mem_alloc(sizeof(ibuf_t)); /* Note that also a pessimistic delete can sometimes make a B-tree grow in size, as the references on the upper levels of the tree can change */ ibuf->max_size = buf_pool_get_curr_size() / UNIV_PAGE_SIZE / IBUF_POOL_SIZE_PER_MAX_SIZE; ibuf->meter = IBUF_THRESHOLD + 1; UT_LIST_INIT(ibuf->data_list); ibuf->size = 0; #ifdef UNIV_IBUF_DEBUG { ulint i, j; for (i = 0; i < IBUF_COUNT_N_SPACES; i++) { ibuf_counts[i] = mem_alloc(sizeof(ulint) * IBUF_COUNT_N_PAGES); for (j = 0; j < IBUF_COUNT_N_PAGES; j++) { ibuf_count_set(i, j, 0); } } } #endif mutex_create(&ibuf_pessimistic_insert_mutex); mutex_set_level(&ibuf_pessimistic_insert_mutex, SYNC_IBUF_PESS_INSERT_MUTEX); mutex_create(&ibuf_mutex); mutex_set_level(&ibuf_mutex, SYNC_IBUF_MUTEX); mutex_create(&ibuf_bitmap_mutex); mutex_set_level(&ibuf_bitmap_mutex, SYNC_IBUF_BITMAP_MUTEX); fil_ibuf_init_at_db_start(); ibuf_counts_inited = TRUE; } /********************************************************************** Updates the size information in an ibuf data, assuming the segment size has not changed. */ static void ibuf_data_sizes_update( /*===================*/ ibuf_data_t* data, /* in: ibuf data struct */ page_t* root, /* in: ibuf tree root */ mtr_t* mtr) /* in: mtr */ { ulint old_size; ut_ad(mutex_own(&ibuf_mutex)); old_size = data->size; data->free_list_len = flst_get_len(root + PAGE_HEADER + PAGE_BTR_IBUF_FREE_LIST, mtr); data->height = 1 + btr_page_get_level(root, mtr); data->size = data->seg_size - (1 + data->free_list_len); /* the '1 +' is the ibuf header page */ ut_ad(data->size < data->seg_size); if (page_get_n_recs(root) == 0) { data->empty = TRUE; } else { data->empty = FALSE; } ut_ad(ibuf->size + data->size >= old_size); ibuf->size = ibuf->size + data->size - old_size; /* printf("ibuf size %lu, space ibuf size %lu\n", ibuf->size, data->size); */ } /********************************************************************** Creates the insert buffer data struct for a single tablespace. Reads the root page of the insert buffer tree in the tablespace. This function can be called only after the dictionary system has been initialized, as this creates also the insert buffer table and index for this tablespace. */ ibuf_data_t* ibuf_data_init_for_space( /*=====================*/ /* out, own: ibuf data struct, linked to the list in ibuf control structure. */ ulint space) /* in: space id */ { ibuf_data_t* data; page_t* root; page_t* header_page; mtr_t mtr; char buf[50]; dict_table_t* table; dict_index_t* index; ulint n_used; #ifdef UNIV_LOG_DEBUG if (space % 2 == 1) { printf("No ibuf op in replicate space\n"); return(NULL); } #endif data = mem_alloc(sizeof(ibuf_data_t)); data->space = space; mtr_start(&mtr); mutex_enter(&ibuf_mutex); mtr_x_lock(fil_space_get_latch(space), &mtr); header_page = ibuf_header_page_get(space, &mtr); fseg_n_reserved_pages(header_page + IBUF_HEADER + IBUF_TREE_SEG_HEADER, &n_used, &mtr); ibuf_enter(); ut_ad(n_used >= 2); data->seg_size = n_used; root = buf_page_get(space, FSP_IBUF_TREE_ROOT_PAGE_NO, RW_X_LATCH, &mtr); buf_page_dbg_add_level(root, SYNC_TREE_NODE); data->size = 0; data->n_inserts = 0; data->n_merges = 0; data->n_merged_recs = 0; ibuf_data_sizes_update(data, root, &mtr); mutex_exit(&ibuf_mutex); mtr_commit(&mtr); ibuf_exit(); sprintf(buf, "SYS_IBUF_TABLE_%lu", space); table = dict_mem_table_create(buf, space, 2); dict_mem_table_add_col(table, "PAGE_NO", DATA_BINARY, 0, 0, 0); dict_mem_table_add_col(table, "TYPES", DATA_BINARY, 0, 0, 0); table->id = ut_dulint_add(DICT_IBUF_ID_MIN, space); dict_table_add_to_cache(table); index = dict_mem_index_create(buf, "CLUST_IND", space, DICT_CLUSTERED | DICT_UNIVERSAL | DICT_IBUF, 2); dict_mem_index_add_field(index, "PAGE_NO", 0); dict_mem_index_add_field(index, "TYPES", 0); index->page_no = FSP_IBUF_TREE_ROOT_PAGE_NO; index->id = ut_dulint_add(DICT_IBUF_ID_MIN, space); dict_index_add_to_cache(table, index); data->index = dict_table_get_first_index(table); mutex_enter(&ibuf_mutex); UT_LIST_ADD_LAST(data_list, ibuf->data_list, data); mutex_exit(&ibuf_mutex); return(data); } /************************************************************************* Initializes an ibuf bitmap page. */ void ibuf_bitmap_page_init( /*==================*/ page_t* page, /* in: bitmap page */ mtr_t* mtr) /* in: mtr */ { ulint bit_offset; ulint byte_offset; ulint i; /* Write all zeros to the bitmap */ bit_offset = XDES_DESCRIBED_PER_PAGE * IBUF_BITS_PER_PAGE; byte_offset = bit_offset / 8 + 1; for (i = IBUF_BITMAP; i < IBUF_BITMAP + byte_offset; i++) { *(page + i) = (byte)0; } mlog_write_initial_log_record(page, MLOG_IBUF_BITMAP_INIT, mtr); } /************************************************************************* Parses a redo log record of an ibuf bitmap page init. */ byte* ibuf_parse_bitmap_init( /*===================*/ /* out: end of log record or NULL */ byte* ptr, /* in: buffer */ byte* end_ptr,/* in: buffer end */ page_t* page, /* in: page or NULL */ mtr_t* mtr) /* in: mtr or NULL */ { ut_ad(ptr && end_ptr); if (page) { ibuf_bitmap_page_init(page, mtr); } return(ptr); } /************************************************************************ Gets the desired bits for a given page from a bitmap page. */ UNIV_INLINE ulint ibuf_bitmap_page_get_bits( /*======================*/ /* out: value of bits */ page_t* page, /* in: bitmap page */ ulint page_no,/* in: page whose bits to get */ ulint bit, /* in: IBUF_BITMAP_FREE, IBUF_BITMAP_BUFFERED, ... */ mtr_t* mtr) /* in: mtr containing an x-latch to the bitmap page */ { ulint byte_offset; ulint bit_offset; ulint map_byte; ulint value; ut_ad(bit < IBUF_BITS_PER_PAGE); ut_ad(IBUF_BITS_PER_PAGE % 2 == 0); ut_ad(mtr_memo_contains(mtr, buf_block_align(page), MTR_MEMO_PAGE_X_FIX)); bit_offset = (page_no % XDES_DESCRIBED_PER_PAGE) * IBUF_BITS_PER_PAGE + bit; byte_offset = bit_offset / 8; bit_offset = bit_offset % 8; ut_ad(byte_offset + IBUF_BITMAP < UNIV_PAGE_SIZE); map_byte = mach_read_from_1(page + IBUF_BITMAP + byte_offset); value = ut_bit_get_nth(map_byte, bit_offset); if (bit == IBUF_BITMAP_FREE) { ut_ad(bit_offset + 1 < 8); value = value * 2 + ut_bit_get_nth(map_byte, bit_offset + 1); } return(value); } /************************************************************************ Sets the desired bit for a given page in a bitmap page. */ static void ibuf_bitmap_page_set_bits( /*======================*/ page_t* page, /* in: bitmap page */ ulint page_no,/* in: page whose bits to set */ ulint bit, /* in: IBUF_BITMAP_FREE, IBUF_BITMAP_BUFFERED, ... */ ulint val, /* in: value to set */ mtr_t* mtr) /* in: mtr containing an x-latch to the bitmap page */ { ulint byte_offset; ulint bit_offset; ulint map_byte; ut_ad(bit < IBUF_BITS_PER_PAGE); ut_ad(IBUF_BITS_PER_PAGE % 2 == 0); ut_ad(mtr_memo_contains(mtr, buf_block_align(page), MTR_MEMO_PAGE_X_FIX)); #ifdef UNIV_IBUF_DEBUG ut_a((bit != IBUF_BITMAP_BUFFERED) || (val != FALSE) || (0 == ibuf_count_get(buf_frame_get_space_id(page), page_no))); #endif bit_offset = (page_no % XDES_DESCRIBED_PER_PAGE) * IBUF_BITS_PER_PAGE + bit; byte_offset = bit_offset / 8; bit_offset = bit_offset % 8; ut_ad(byte_offset + IBUF_BITMAP < UNIV_PAGE_SIZE); map_byte = mach_read_from_1(page + IBUF_BITMAP + byte_offset); if (bit == IBUF_BITMAP_FREE) { ut_ad(bit_offset + 1 < 8); ut_ad(val <= 3); map_byte = ut_bit_set_nth(map_byte, bit_offset, val / 2); map_byte = ut_bit_set_nth(map_byte, bit_offset + 1, val % 2); } else { ut_ad(val <= 1); map_byte = ut_bit_set_nth(map_byte, bit_offset, val); } mlog_write_ulint(page + IBUF_BITMAP + byte_offset, map_byte, MLOG_1BYTE, mtr); } /************************************************************************ Calculates the bitmap page number for a given page number. */ UNIV_INLINE ulint ibuf_bitmap_page_no_calc( /*=====================*/ /* out: the bitmap page number where the file page is mapped */ ulint page_no) /* in: tablespace page number */ { return(FSP_IBUF_BITMAP_OFFSET + XDES_DESCRIBED_PER_PAGE * (page_no / XDES_DESCRIBED_PER_PAGE)); } /************************************************************************ Gets the ibuf bitmap page where the bits describing a given file page are stored. */ static page_t* ibuf_bitmap_get_map_page( /*=====================*/ /* out: bitmap page where the file page is mapped, that is, the bitmap page containing the descriptor bits for the file page; the bitmap page is x-latched */ ulint space, /* in: space id of the file page */ ulint page_no,/* in: page number of the file page */ mtr_t* mtr) /* in: mtr */ { page_t* page; page = buf_page_get(space, ibuf_bitmap_page_no_calc(page_no), RW_X_LATCH, mtr); buf_page_dbg_add_level(page, SYNC_IBUF_BITMAP); return(page); } /**************************************************************************** Sets the free bits of the page in the ibuf bitmap. This is done in a separate mini-transaction, hence this operation does not restrict further work to only ibuf bitmap operations, which would result if the latch to the bitmap pag were kept. */ UNIV_INLINE void ibuf_set_free_bits_low( /*===================*/ ulint type, /* in: index type */ page_t* page, /* in: index page; free bit is reset if the index is a non-clustered non-unique, and page level is 0 */ ulint val, /* in: value to set: < 4 */ mtr_t* mtr) /* in: mtr */ { page_t* bitmap_page; if (type & (DICT_CLUSTERED | DICT_UNIQUE)) { return; } if (btr_page_get_level_low(page) != 0) { return; } bitmap_page = ibuf_bitmap_get_map_page(buf_frame_get_space_id(page), buf_frame_get_page_no(page), mtr); #ifdef UNIV_IBUF_DEBUG /* printf("Setting page no %lu free bits to %lu should be %lu\n", buf_frame_get_page_no(page), val, ibuf_index_page_calc_free(page)); */ ut_a(val <= ibuf_index_page_calc_free(page)); #endif ibuf_bitmap_page_set_bits(bitmap_page, buf_frame_get_page_no(page), IBUF_BITMAP_FREE, val, mtr); } /**************************************************************************** Sets the free bit of the page in the ibuf bitmap. This is done in a separate mini-transaction, hence this operation does not restrict further work to only ibuf bitmap operations, which would result if the latch to the bitmap page were kept. */ void ibuf_set_free_bits( /*===============*/ ulint type, /* in: index type */ page_t* page, /* in: index page; free bit is reset if the index is a non-clustered non-unique, and page level is 0 */ ulint val, /* in: value to set: < 4 */ ulint max_val)/* in: ULINT_UNDEFINED or a maximum value which the bits must have before setting; this is for debugging */ { mtr_t mtr; page_t* bitmap_page; if (type & (DICT_CLUSTERED | DICT_UNIQUE)) { return; } if (btr_page_get_level_low(page) != 0) { return; } mtr_start(&mtr); bitmap_page = ibuf_bitmap_get_map_page(buf_frame_get_space_id(page), buf_frame_get_page_no(page), &mtr); if (max_val != ULINT_UNDEFINED) { #ifdef UNIV_IBUF_DEBUG ulint old_val; old_val = ibuf_bitmap_page_get_bits(bitmap_page, buf_frame_get_page_no(page), IBUF_BITMAP_FREE, &mtr); if (old_val != max_val) { /* printf( "Ibuf: page %lu old val %lu max val %lu\n", buf_frame_get_page_no(page), old_val, max_val); */ } ut_a(old_val <= max_val); #endif } #ifdef UNIV_IBUF_DEBUG /* printf("Setting page no %lu free bits to %lu should be %lu\n", buf_frame_get_page_no(page), val, ibuf_index_page_calc_free(page)); */ ut_a(val <= ibuf_index_page_calc_free(page)); #endif ibuf_bitmap_page_set_bits(bitmap_page, buf_frame_get_page_no(page), IBUF_BITMAP_FREE, val, &mtr); mtr_commit(&mtr); } /**************************************************************************** Resets the free bits of the page in the ibuf bitmap. This is done in a separate mini-transaction, hence this operation does not restrict further work to only ibuf bitmap operations, which would result if the latch to the bitmap page were kept. */ void ibuf_reset_free_bits_with_type( /*===========================*/ ulint type, /* in: index type */ page_t* page) /* in: index page; free bits are set to 0 if the index is non-clustered and non-unique and the page level is 0 */ { ibuf_set_free_bits(type, page, 0, ULINT_UNDEFINED); } /**************************************************************************** Resets the free bits of the page in the ibuf bitmap. This is done in a separate mini-transaction, hence this operation does not restrict further work to solely ibuf bitmap operations, which would result if the latch to the bitmap page were kept. */ void ibuf_reset_free_bits( /*=================*/ dict_index_t* index, /* in: index */ page_t* page) /* in: index page; free bits are set to 0 if the index is non-clustered and non-unique and the page level is 0 */ { ibuf_set_free_bits(index->type, page, 0, ULINT_UNDEFINED); } /************************************************************************** Updates the free bits for a page to reflect the present state. Does this in the mtr given, which means that the latching order rules virtually prevent any further operations for this OS thread until mtr is committed. */ void ibuf_update_free_bits_low( /*======================*/ dict_index_t* index, /* in: index */ page_t* page, /* in: index page */ ulint max_ins_size, /* in: value of maximum insert size with reorganize before the latest operation performed to the page */ mtr_t* mtr) /* in: mtr */ { ulint before; ulint after; before = ibuf_index_page_calc_free_bits(max_ins_size); after = ibuf_index_page_calc_free(page); if (before != after) { ibuf_set_free_bits_low(index->type, page, after, mtr); } } /************************************************************************** Updates the free bits for the two pages to reflect the present state. Does this in the mtr given, which means that the latching order rules virtually prevent any further operations until mtr is committed. */ void ibuf_update_free_bits_for_two_pages_low( /*====================================*/ dict_index_t* index, /* in: index */ page_t* page1, /* in: index page */ page_t* page2, /* in: index page */ mtr_t* mtr) /* in: mtr */ { ulint state; /* As we have to x-latch two random bitmap pages, we have to acquire the bitmap mutex to prevent a deadlock with a similar operation performed by another OS thread. */ mutex_enter(&ibuf_bitmap_mutex); state = ibuf_index_page_calc_free(page1); ibuf_set_free_bits_low(index->type, page1, state, mtr); state = ibuf_index_page_calc_free(page2); ibuf_set_free_bits_low(index->type, page2, state, mtr); mutex_exit(&ibuf_bitmap_mutex); } /************************************************************************** Returns TRUE if the page is one of the fixed address ibuf pages. */ UNIV_INLINE ibool ibuf_fixed_addr_page( /*=================*/ /* out: TRUE if a fixed address ibuf i/o page */ ulint page_no)/* in: page number */ { if ((ibuf_bitmap_page(page_no)) || (page_no == IBUF_TREE_ROOT_PAGE_NO)) { return(TRUE); } return(FALSE); } /*************************************************************************** Checks if a page is a level 2 or 3 page in the ibuf hierarchy of pages. */ ibool ibuf_page( /*======*/ /* out: TRUE if level 2 or level 3 page */ ulint space, /* in: space id */ ulint page_no)/* in: page number */ { page_t* bitmap_page; mtr_t mtr; ibool ret; if (recv_no_ibuf_operations) { /* Recovery is running: no ibuf operations should be performed */ return(FALSE); } if (ibuf_fixed_addr_page(page_no)) { return(TRUE); } ut_ad(fil_space_get_type(space) == FIL_TABLESPACE); mtr_start(&mtr); bitmap_page = ibuf_bitmap_get_map_page(space, page_no, &mtr); ret = ibuf_bitmap_page_get_bits(bitmap_page, page_no, IBUF_BITMAP_IBUF, &mtr); mtr_commit(&mtr); return(ret); } /*************************************************************************** Checks if a page is a level 2 or 3 page in the ibuf hierarchy of pages. */ ibool ibuf_page_low( /*==========*/ /* out: TRUE if level 2 or level 3 page */ ulint space, /* in: space id */ ulint page_no,/* in: page number */ mtr_t* mtr) /* in: mtr which will contain an x-latch to the bitmap page if the page is not one of the fixed address ibuf pages */ { page_t* bitmap_page; ibool ret; #ifdef UNIV_LOG_DEBUG if (space % 2 != 0) { printf("No ibuf in a replicate space\n"); return(FALSE); } #endif if (ibuf_fixed_addr_page(page_no)) { return(TRUE); } bitmap_page = ibuf_bitmap_get_map_page(space, page_no, mtr); ret = ibuf_bitmap_page_get_bits(bitmap_page, page_no, IBUF_BITMAP_IBUF, mtr); return(ret); } /************************************************************************ Returns the page number field of an ibuf record. */ static ulint ibuf_rec_get_page_no( /*=================*/ /* out: page number */ rec_t* rec) /* in: ibuf record */ { byte* field; ulint len; ut_ad(ibuf_inside()); ut_ad(rec_get_n_fields(rec) > 2); field = rec_get_nth_field(rec, 0, &len); ut_ad(len == 4); return(mach_read_from_4(field)); } /************************************************************************ Returns the space taken by a stored non-clustered index entry if converted to an index record. */ static ulint ibuf_rec_get_volume( /*================*/ /* out: size of index record in bytes + an upper limit of the space taken in the page directory */ rec_t* ibuf_rec)/* in: ibuf record */ { dtype_t dtype; ulint data_size = 0; ulint n_fields; byte* types; byte* data; ulint len; ulint i; ut_ad(ibuf_inside()); ut_ad(rec_get_n_fields(ibuf_rec) > 2); n_fields = rec_get_n_fields(ibuf_rec) - 2; types = rec_get_nth_field(ibuf_rec, 1, &len); ut_ad(len == n_fields * DATA_ORDER_NULL_TYPE_BUF_SIZE); for (i = 0; i < n_fields; i++) { data = rec_get_nth_field(ibuf_rec, i + 2, &len); dtype_read_for_order_and_null_size(&dtype, types + i * DATA_ORDER_NULL_TYPE_BUF_SIZE); if (len == UNIV_SQL_NULL) { data_size += dtype_get_sql_null_size(&dtype); } else { data_size += len; } } return(data_size + rec_get_converted_extra_size(data_size, n_fields) + page_dir_calc_reserved_space(1)); } /************************************************************************* Builds the tuple to insert to an ibuf tree when we have an entry for a non-clustered index. */ static dtuple_t* ibuf_entry_build( /*=============*/ /* out, own: entry to insert into an ibuf index tree; NOTE that the original entry must be kept because we copy pointers to its fields */ dtuple_t* entry, /* in: entry for a non-clustered index */ ulint page_no,/* in: index page number where entry should be inserted */ mem_heap_t* heap) /* in: heap into which to build */ { dtuple_t* tuple; dfield_t* field; dfield_t* entry_field; ulint n_fields; byte* buf; byte* buf2; ulint i; /* We have to build a tuple whose first field is the page number, the second field contains the original type information for entry, and the rest of the fields are copied from entry. All fields in the tuple are of the type binary. */ n_fields = dtuple_get_n_fields(entry); tuple = dtuple_create(heap, n_fields + 2); /* Store the page number in tuple */ field = dtuple_get_nth_field(tuple, 0); buf = mem_heap_alloc(heap, 4); mach_write_to_4(buf, page_no); dfield_set_data(field, buf, 4); /* Store the type info in tuple */ buf2 = mem_heap_alloc(heap, n_fields * DATA_ORDER_NULL_TYPE_BUF_SIZE); for (i = 0; i < n_fields; i++) { field = dtuple_get_nth_field(tuple, i + 2); entry_field = dtuple_get_nth_field(entry, i); dfield_copy(field, entry_field); dtype_store_for_order_and_null_size( buf2 + i * DATA_ORDER_NULL_TYPE_BUF_SIZE, dfield_get_type(entry_field)); } field = dtuple_get_nth_field(tuple, 1); dfield_set_data(field, buf2, n_fields * DATA_ORDER_NULL_TYPE_BUF_SIZE); /* Set the types in the new tuple binary */ dtuple_set_types_binary(tuple, n_fields + 2); return(tuple); } /************************************************************************* Builds the entry to insert into a non-clustered index when we have the corresponding record in an ibuf index. */ static dtuple_t* ibuf_build_entry_from_ibuf_rec( /*===========================*/ /* out, own: entry to insert to a non-clustered index; NOTE that as we copy pointers to fields in ibuf_rec, the caller must hold a latch to the ibuf_rec page as long as the entry is used! */ rec_t* ibuf_rec, /* in: record in an insert buffer */ mem_heap_t* heap) /* in: heap where built */ { dtuple_t* tuple; dfield_t* field; ulint n_fields; byte* types; byte* data; ulint len; ulint i; n_fields = rec_get_n_fields(ibuf_rec) - 2; tuple = dtuple_create(heap, n_fields); types = rec_get_nth_field(ibuf_rec, 1, &len); ut_ad(len == n_fields * DATA_ORDER_NULL_TYPE_BUF_SIZE); for (i = 0; i < n_fields; i++) { field = dtuple_get_nth_field(tuple, i); data = rec_get_nth_field(ibuf_rec, i + 2, &len); dfield_set_data(field, data, len); dtype_read_for_order_and_null_size(dfield_get_type(field), types + i * DATA_ORDER_NULL_TYPE_BUF_SIZE); } return(tuple); } /************************************************************************* Builds a search tuple used to search buffered inserts for an index page. */ static dtuple_t* ibuf_search_tuple_build( /*====================*/ /* out, own: search tuple */ ulint page_no,/* in: index page number */ mem_heap_t* heap) /* in: heap into which to build */ { dtuple_t* tuple; dfield_t* field; byte* buf; tuple = dtuple_create(heap, 1); /* Store the page number in tuple */ field = dtuple_get_nth_field(tuple, 0); buf = mem_heap_alloc(heap, 4); mach_write_to_4(buf, page_no); dfield_set_data(field, buf, 4); dtuple_set_types_binary(tuple, 1); return(tuple); } /************************************************************************* Checks if there are enough pages in the free list of the ibuf tree that we dare to start a pessimistic insert to the insert buffer. */ UNIV_INLINE ibool ibuf_data_enough_free_for_insert( /*=============================*/ /* out: TRUE if enough free pages in list */ ibuf_data_t* data) /* in: ibuf data for the space */ { ut_ad(mutex_own(&ibuf_mutex)); /* We want a big margin of free pages, because a B-tree can sometimes grow in size also if records are deleted from it, as the node pointers can change, and we must make sure that we are able to delete the inserts buffered for pages that we read to the buffer pool, without any risk of running out of free space in the insert buffer. */ if (data->free_list_len >= data->size / 2 + 3 * data->height) { return(TRUE); } return(FALSE); } /************************************************************************* Checks if there are enough pages in the free list of the ibuf tree that we should remove them and free to the file space management. */ UNIV_INLINE ibool ibuf_data_too_much_free( /*====================*/ /* out: TRUE if enough free pages in list */ ibuf_data_t* data) /* in: ibuf data for the space */ { ut_ad(mutex_own(&ibuf_mutex)); if (data->free_list_len >= 3 + data->size / 2 + 3 * data->height) { return(TRUE); } return(FALSE); } /************************************************************************* Allocates a new page from the ibuf file segment and adds it to the free list. */ static ulint ibuf_add_free_page( /*===============*/ /* out: DB_SUCCESS, or DB_STRONG_FAIL if no space left */ ulint space, /* in: space id */ ibuf_data_t* ibuf_data) /* in: ibuf data for the space */ { mtr_t mtr; page_t* header_page; ulint page_no; page_t* page; page_t* root; page_t* bitmap_page; mtr_start(&mtr); /* Acquire the fsp latch before the ibuf header, obeying the latching order */ mtr_x_lock(fil_space_get_latch(space), &mtr); header_page = ibuf_header_page_get(space, &mtr); /* Allocate a new page: NOTE that if the page has been a part of a non-clustered index which has subsequently been dropped, then the page may have buffered inserts in the insert buffer, and these should be deleted from there. These get deleted when the page allocation creates the page in buffer. Thus the call below may end up calling the insert buffer routines and, as we yet have no latches to insert buffer tree pages, these routines can run without a risk of a deadlock. This is the reason why we created a special ibuf header page apart from the ibuf tree. */ page_no = fseg_alloc_free_page(header_page + IBUF_HEADER + IBUF_TREE_SEG_HEADER, 0, FSP_UP, &mtr); if (page_no == FIL_NULL) { mtr_commit(&mtr); return(DB_STRONG_FAIL); } page = buf_page_get(space, page_no, RW_X_LATCH, &mtr); buf_page_dbg_add_level(page, SYNC_TREE_NODE_NEW); ibuf_enter(); mutex_enter(&ibuf_mutex); root = ibuf_tree_root_get(ibuf_data, space, &mtr); /* Add the page to the free list and update the ibuf size data */ flst_add_last(root + PAGE_HEADER + PAGE_BTR_IBUF_FREE_LIST, page + PAGE_HEADER + PAGE_BTR_IBUF_FREE_LIST_NODE, &mtr); ibuf_data->seg_size++; ibuf_data->free_list_len++; /* Set the bit indicating that this page is now an ibuf tree page (level 2 page) */ bitmap_page = ibuf_bitmap_get_map_page(space, page_no, &mtr); ibuf_bitmap_page_set_bits(bitmap_page, page_no, IBUF_BITMAP_IBUF, TRUE, &mtr); mtr_commit(&mtr); mutex_exit(&ibuf_mutex); ibuf_exit(); return(DB_SUCCESS); } /************************************************************************* Removes a page from the free list and frees it to the fsp system. */ static void ibuf_remove_free_page( /*==================*/ ulint space, /* in: space id */ ibuf_data_t* ibuf_data) /* in: ibuf data for the space */ { mtr_t mtr; mtr_t mtr2; page_t* header_page; ulint page_no; page_t* page; page_t* root; page_t* bitmap_page; mtr_start(&mtr); /* Acquire the fsp latch before the ibuf header, obeying the latching order */ mtr_x_lock(fil_space_get_latch(space), &mtr); header_page = ibuf_header_page_get(space, &mtr); /* Prevent pessimistic inserts to insert buffer trees for a while */ mutex_enter(&ibuf_pessimistic_insert_mutex); ibuf_enter(); mutex_enter(&ibuf_mutex); if (!ibuf_data_too_much_free(ibuf_data)) { mutex_exit(&ibuf_mutex); ibuf_exit(); mutex_exit(&ibuf_pessimistic_insert_mutex); mtr_commit(&mtr); return; } mtr_start(&mtr2); root = ibuf_tree_root_get(ibuf_data, space, &mtr2); page_no = flst_get_last(root + PAGE_HEADER + PAGE_BTR_IBUF_FREE_LIST, &mtr2) .page; /* NOTE that we must release the latch on the ibuf tree root because in fseg_free_page we access level 1 pages, and the root is a level 2 page. */ mtr_commit(&mtr2); mutex_exit(&ibuf_mutex); ibuf_exit(); /* Since pessimistic inserts were prevented, we know that the page is still in the free list. NOTE that also deletes may take pages from the free list, but they take them from the start, and the free list was so long that they cannot have taken the last page from it. */ fseg_free_page(header_page + IBUF_HEADER + IBUF_TREE_SEG_HEADER, space, page_no, &mtr); #ifdef UNIV_DEBUG_FILE_ACCESSES buf_page_reset_file_page_was_freed(space, page_no); #endif ibuf_enter(); mutex_enter(&ibuf_mutex); root = ibuf_tree_root_get(ibuf_data, space, &mtr); ut_ad(page_no == flst_get_last(root + PAGE_HEADER + PAGE_BTR_IBUF_FREE_LIST, &mtr) .page); page = buf_page_get(space, page_no, RW_X_LATCH, &mtr); buf_page_dbg_add_level(page, SYNC_TREE_NODE); /* Remove the page from the free list and update the ibuf size data */ flst_remove(root + PAGE_HEADER + PAGE_BTR_IBUF_FREE_LIST, page + PAGE_HEADER + PAGE_BTR_IBUF_FREE_LIST_NODE, &mtr); ibuf_data->seg_size--; ibuf_data->free_list_len--; mutex_exit(&ibuf_pessimistic_insert_mutex); /* Set the bit indicating that this page is no more an ibuf tree page (level 2 page) */ bitmap_page = ibuf_bitmap_get_map_page(space, page_no, &mtr); ibuf_bitmap_page_set_bits(bitmap_page, page_no, IBUF_BITMAP_IBUF, FALSE, &mtr); #ifdef UNIV_DEBUG_FILE_ACCESSES buf_page_set_file_page_was_freed(space, page_no); #endif mtr_commit(&mtr); mutex_exit(&ibuf_mutex); ibuf_exit(); } /*************************************************************************** Frees excess pages from the ibuf free list. This function is called when an OS thread calls fsp services to allocate a new file segment, or a new page to a file segment, and the thread did not own the fsp latch before this call. */ void ibuf_free_excess_pages( /*===================*/ ulint space) /* in: space id */ { ibuf_data_t* ibuf_data; ulint i; ut_ad(rw_lock_own(fil_space_get_latch(space), RW_LOCK_EX)); ut_ad(rw_lock_get_x_lock_count(fil_space_get_latch(space)) == 1); ut_ad(!ibuf_inside()); /* NOTE: We require that the thread did not own the latch before, because then we know that we can obey the correct latching order for ibuf latches */ ibuf_data = fil_space_get_ibuf_data(space); if (ibuf_data == NULL) { /* Not yet initialized */ #ifdef UNIV_DEBUG /*printf("Ibuf for space %lu not yet initialized\n", space); */ #endif return; } /* Free at most a few pages at a time, so that we do not delay the requested service too much */ for (i = 0; i < 4; i++) { mutex_enter(&ibuf_mutex); if (!ibuf_data_too_much_free(ibuf_data)) { mutex_exit(&ibuf_mutex); return; } mutex_exit(&ibuf_mutex); ibuf_remove_free_page(space, ibuf_data); } } /************************************************************************* Reads page numbers from a leaf in an ibuf tree. */ static ulint ibuf_get_merge_page_nos( /*====================*/ /* out: a lower limit for the combined volume of records which will be merged */ ibool contract,/* in: TRUE if this function is called to contract the tree, FALSE if this is called when a single page becomes full and we look if it pays to read also nearby pages */ rec_t* first_rec,/* in: record from which we read down and up in the chain of records */ ulint* page_nos,/* in/out: buffer for at least IBUF_MAX_N_PAGES_MERGED many page numbers; the page numbers are in an ascending order */ ulint* n_stored)/* out: number of page numbers stored to page_nos in this function */ { ulint prev_page_no; ulint first_page_no; ulint rec_page_no; rec_t* rec; ulint sum_volumes; ulint volume_for_page; ulint rec_volume; ulint limit; page_t* page; ulint n_pages; *n_stored = 0; limit = ut_min(IBUF_MAX_N_PAGES_MERGED, buf_pool->curr_size / 4); page = buf_frame_align(first_rec); if (first_rec == page_get_supremum_rec(page)) { first_rec = page_rec_get_prev(first_rec); } if (first_rec == page_get_infimum_rec(page)) { first_rec = page_rec_get_next(first_rec); } if (first_rec == page_get_supremum_rec(page)) { return(0); } rec = first_rec; first_page_no = ibuf_rec_get_page_no(first_rec); n_pages = 0; prev_page_no = 0; while ((rec != page_get_infimum_rec(page)) && (n_pages < limit)) { rec_page_no = ibuf_rec_get_page_no(rec); ut_ad(rec_page_no != 0); if (rec_page_no / IBUF_MERGE_AREA != first_page_no / IBUF_MERGE_AREA) { break; } if (rec_page_no != prev_page_no) { n_pages++; } prev_page_no = rec_page_no; rec = page_rec_get_prev(rec); } rec = page_rec_get_next(rec); prev_page_no = 0; sum_volumes = 0; volume_for_page = 0; while (*n_stored < limit) { if (rec == page_get_supremum_rec(page)) { rec_page_no = 1; } else { rec_page_no = ibuf_rec_get_page_no(rec); ut_ad(rec_page_no > IBUF_TREE_ROOT_PAGE_NO); } #ifdef UNIV_IBUF_DEBUG ut_a(*n_stored < IBUF_MAX_N_PAGES_MERGED); #endif if (rec_page_no != prev_page_no) { if ((prev_page_no == first_page_no) || contract || (volume_for_page > ((IBUF_MERGE_THRESHOLD - 1) * 4 * UNIV_PAGE_SIZE / IBUF_PAGE_SIZE_PER_FREE_SPACE) / IBUF_MERGE_THRESHOLD)) { page_nos[*n_stored] = prev_page_no; (*n_stored)++; sum_volumes += volume_for_page; } if (rec_page_no / IBUF_MERGE_AREA != first_page_no / IBUF_MERGE_AREA) { break; } volume_for_page = 0; } if (rec_page_no == 1) { /* Supremum record */ break; } rec_volume = ibuf_rec_get_volume(rec); volume_for_page += rec_volume; prev_page_no = rec_page_no; rec = page_rec_get_next(rec); } #ifdef UNIV_IBUF_DEBUG ut_a(*n_stored <= IBUF_MAX_N_PAGES_MERGED); #endif /* printf("Ibuf merge batch %lu pages %lu volume\n", *n_stored, sum_volumes); */ return(sum_volumes); } /************************************************************************* Contracts insert buffer trees by reading pages to the buffer pool. */ static ulint ibuf_contract_ext( /*==============*/ /* out: a lower limit for the combined size in bytes of entries which will be merged from ibuf trees to the pages read, 0 if ibuf is empty */ ulint* n_pages,/* out: number of pages to which merged */ ibool sync) /* in: TRUE if the caller wants to wait for the issued read with the highest tablespace address to complete */ { ulint rnd_pos; ibuf_data_t* data; btr_pcur_t pcur; ulint space; ibool all_trees_empty; ulint page_nos[IBUF_MAX_N_PAGES_MERGED]; ulint n_stored; ulint sum_sizes; mtr_t mtr; *n_pages = 0; loop: ut_ad(!ibuf_inside()); mutex_enter(&ibuf_mutex); ut_ad(ibuf_validate_low()); /* Choose an ibuf tree at random */ ibuf_rnd += 865558671; rnd_pos = ibuf_rnd % ibuf->size; all_trees_empty = TRUE; data = UT_LIST_GET_FIRST(ibuf->data_list); for (;;) { if (!data->empty) { all_trees_empty = FALSE; if (rnd_pos < data->size) { break; } rnd_pos -= data->size; } data = UT_LIST_GET_NEXT(data_list, data); if (data == NULL) { if (all_trees_empty) { mutex_exit(&ibuf_mutex); return(0); } data = UT_LIST_GET_FIRST(ibuf->data_list); } } ut_ad(data); space = (data->index)->space; mtr_start(&mtr); ibuf_enter(); /* Open a cursor to a randomly chosen leaf of the tree, at a random position within the leaf */ btr_pcur_open_at_rnd_pos(data->index, BTR_SEARCH_LEAF, &pcur, &mtr); if (0 == page_get_n_recs(btr_pcur_get_page(&pcur))) { /* This tree is empty */ data->empty = TRUE; ibuf_exit(); mtr_commit(&mtr); btr_pcur_close(&pcur); mutex_exit(&ibuf_mutex); goto loop; } mutex_exit(&ibuf_mutex); sum_sizes = ibuf_get_merge_page_nos(TRUE, btr_pcur_get_rec(&pcur), page_nos, &n_stored); #ifdef UNIV_IBUF_DEBUG /* printf("Ibuf contract sync %lu pages %lu volume %lu\n", sync, n_stored, sum_sizes); */ #endif ibuf_exit(); mtr_commit(&mtr); btr_pcur_close(&pcur); buf_read_ibuf_merge_pages(sync, space, page_nos, n_stored); *n_pages = n_stored; return(sum_sizes + 1); } /************************************************************************* Contracts insert buffer trees by reading pages to the buffer pool. */ ulint ibuf_contract( /*==========*/ /* out: a lower limit for the combined size in bytes of entries which will be merged from ibuf trees to the pages read, 0 if ibuf is empty */ ibool sync) /* in: TRUE if the caller wants to wait for the issued read with the highest tablespace address to complete */ { ulint n_pages; return(ibuf_contract_ext(&n_pages, sync)); } /************************************************************************* Contracts insert buffer trees by reading pages to the buffer pool. */ ulint ibuf_contract_for_n_pages( /*======================*/ /* out: a lower limit for the combined size in bytes of entries which will be merged from ibuf trees to the pages read, 0 if ibuf is empty */ ibool sync, /* in: TRUE if the caller wants to wait for the issued read with the highest tablespace address to complete */ ulint n_pages)/* in: try to read at least this many pages to the buffer pool and merge the ibuf contents to them */ { ulint sum_bytes = 0; ulint sum_pages = 0; ulint n_bytes; ulint n_pag2; while (sum_pages < n_pages) { n_bytes = ibuf_contract_ext(&n_pag2, sync); if (n_bytes == 0) { return(sum_bytes); } sum_bytes += n_bytes; sum_pages += n_pag2; } return(sum_bytes); } /************************************************************************* Contract insert buffer trees after insert if they are too big. */ UNIV_INLINE void ibuf_contract_after_insert( /*=======================*/ ulint entry_size) /* in: size of a record which was inserted into an ibuf tree */ { ibool sync; ulint sum_sizes; ulint size; mutex_enter(&ibuf_mutex); if (ibuf->size < ibuf->max_size + IBUF_CONTRACT_ON_INSERT_NON_SYNC) { mutex_exit(&ibuf_mutex); return; } sync = FALSE; if (ibuf->size >= ibuf->max_size + IBUF_CONTRACT_ON_INSERT_SYNC) { sync = TRUE; } mutex_exit(&ibuf_mutex); /* Contract at least entry_size many bytes */ sum_sizes = 0; size = 1; while ((size > 0) && (sum_sizes < entry_size)) { size = ibuf_contract(sync); sum_sizes += size; } } /************************************************************************* Gets an upper limit for the combined size of entries buffered in the insert buffer for a given page. */ ulint ibuf_get_volume_buffered( /*=====================*/ /* out: upper limit for the volume of buffered inserts for the index page, in bytes; we may also return UNIV_PAGE_SIZE, if the entries for the index page span on several pages in the insert buffer */ btr_pcur_t* pcur, /* in: pcur positioned at a place in an insert buffer tree where we would insert an entry for the index page whose number is page_no, latch mode has to be BTR_MODIFY_PREV or BTR_MODIFY_TREE */ ulint space, /* in: space id */ ulint page_no,/* in: page number of an index page */ mtr_t* mtr) /* in: mtr */ { ulint volume; rec_t* rec; page_t* page; ulint prev_page_no; page_t* prev_page; ulint next_page_no; page_t* next_page; ut_ad((pcur->latch_mode == BTR_MODIFY_PREV) || (pcur->latch_mode == BTR_MODIFY_TREE)); /* Count the volume of records earlier in the alphabetical order than pcur */ volume = 0; rec = btr_pcur_get_rec(pcur); page = buf_frame_align(rec); if (rec == page_get_supremum_rec(page)) { rec = page_rec_get_prev(rec); } for (;;) { if (rec == page_get_infimum_rec(page)) { break; } if (page_no != ibuf_rec_get_page_no(rec)) { goto count_later; } volume += ibuf_rec_get_volume(rec); rec = page_rec_get_prev(rec); } /* Look at the previous page */ prev_page_no = btr_page_get_prev(page, mtr); if (prev_page_no == FIL_NULL) { goto count_later; } prev_page = buf_page_get(space, prev_page_no, RW_X_LATCH, mtr); buf_page_dbg_add_level(prev_page, SYNC_TREE_NODE); rec = page_get_supremum_rec(prev_page); rec = page_rec_get_prev(rec); for (;;) { if (rec == page_get_infimum_rec(prev_page)) { /* We cannot go to yet a previous page, because we do not have the x-latch on it, and cannot acquire one because of the latching order: we have to give up */ return(UNIV_PAGE_SIZE); } if (page_no != ibuf_rec_get_page_no(rec)) { goto count_later; } volume += ibuf_rec_get_volume(rec); rec = page_rec_get_prev(rec); } count_later: rec = btr_pcur_get_rec(pcur); if (rec != page_get_supremum_rec(page)) { rec = page_rec_get_next(rec); } for (;;) { if (rec == page_get_supremum_rec(page)) { break; } if (page_no != ibuf_rec_get_page_no(rec)) { return(volume); } volume += ibuf_rec_get_volume(rec); rec = page_rec_get_next(rec); } /* Look at the next page */ next_page_no = btr_page_get_next(page, mtr); if (next_page_no == FIL_NULL) { return(volume); } next_page = buf_page_get(space, next_page_no, RW_X_LATCH, mtr); buf_page_dbg_add_level(next_page, SYNC_TREE_NODE); rec = page_get_infimum_rec(next_page); rec = page_rec_get_next(rec); for (;;) { if (rec == page_get_supremum_rec(next_page)) { /* We give up */ return(UNIV_PAGE_SIZE); } if (page_no != ibuf_rec_get_page_no(rec)) { return(volume); } volume += ibuf_rec_get_volume(rec); rec = page_rec_get_next(rec); } } /************************************************************************* Makes an index insert to the insert buffer, instead of directly to the disk page, if this is possible. */ static ulint ibuf_insert_low( /*============*/ /* out: DB_SUCCESS, DB_FAIL, DB_STRONG_FAIL */ ulint mode, /* in: BTR_MODIFY_PREV or BTR_MODIFY_TREE */ dtuple_t* entry, /* in: index entry to insert */ dict_index_t* index, /* in: index where to insert; must not be unique or clustered */ ulint space, /* in: space id where to insert */ ulint page_no,/* in: page number where to insert */ que_thr_t* thr) /* in: query thread */ { big_rec_t* dummy_big_rec; ulint entry_size; btr_pcur_t pcur; btr_cur_t* cursor; mtr_t mtr; mtr_t bitmap_mtr; dtuple_t* ibuf_entry; mem_heap_t* heap; ulint buffered; rec_t* ins_rec; ibool old_bit_value; page_t* bitmap_page; ibuf_data_t* ibuf_data; dict_index_t* ibuf_index; page_t* root; ulint err; ibool do_merge; ulint page_nos[IBUF_MAX_N_PAGES_MERGED]; ulint n_stored; ulint bits; ut_ad(!(index->type & (DICT_UNIQUE | DICT_CLUSTERED))); ut_ad(dtuple_check_typed(entry)); do_merge = FALSE; ibuf_data = fil_space_get_ibuf_data(space); ibuf_index = ibuf_data->index; mutex_enter(&ibuf_mutex); if (ibuf->size >= ibuf->max_size + IBUF_CONTRACT_DO_NOT_INSERT) { /* Insert buffer is now too big, contract it but do not try to insert */ mutex_exit(&ibuf_mutex); #ifdef UNIV_IBUF_DEBUG printf("Ibuf too big\n"); #endif /* Use synchronous contract (== TRUE) */ ibuf_contract(TRUE); return(DB_STRONG_FAIL); } mutex_exit(&ibuf_mutex); if (mode == BTR_MODIFY_TREE) { mutex_enter(&ibuf_pessimistic_insert_mutex); ibuf_enter(); mutex_enter(&ibuf_mutex); while (!ibuf_data_enough_free_for_insert(ibuf_data)) { mutex_exit(&ibuf_mutex); ibuf_exit(); mutex_exit(&ibuf_pessimistic_insert_mutex); err = ibuf_add_free_page(space, ibuf_data); if (err == DB_STRONG_FAIL) { return(err); } mutex_enter(&ibuf_pessimistic_insert_mutex); ibuf_enter(); mutex_enter(&ibuf_mutex); } } else { ibuf_enter(); } entry_size = rec_get_converted_size(entry); heap = mem_heap_create(512); /* Build the entry which contains the space id and the page number as the first fields and the type information for other fields, and which will be inserted to the insert buffer. */ ibuf_entry = ibuf_entry_build(entry, page_no, heap); /* Open a cursor to the insert buffer tree to calculate if we can add the new entry to it without exceeding the free space limit for the page. */ mtr_start(&mtr); btr_pcur_open(ibuf_index, ibuf_entry, PAGE_CUR_LE, mode, &pcur, &mtr); /* Find out the volume of already buffered inserts for the same index page */ buffered = ibuf_get_volume_buffered(&pcur, space, page_no, &mtr); #ifdef UNIV_IBUF_DEBUG ut_a((buffered == 0) || ibuf_count_get(space, page_no)); #endif mtr_start(&bitmap_mtr); bitmap_page = ibuf_bitmap_get_map_page(space, page_no, &bitmap_mtr); /* We check if the index page is suitable for buffered entries */ if (buf_page_peek(space, page_no) || lock_rec_expl_exist_on_page(space, page_no)) { err = DB_STRONG_FAIL; mtr_commit(&bitmap_mtr); goto function_exit; } bits = ibuf_bitmap_page_get_bits(bitmap_page, page_no, IBUF_BITMAP_FREE, &bitmap_mtr); if (buffered + entry_size + page_dir_calc_reserved_space(1) > ibuf_index_page_calc_free_from_bits(bits)) { mtr_commit(&bitmap_mtr); /* It may not fit */ err = DB_STRONG_FAIL; do_merge = TRUE; ibuf_get_merge_page_nos(FALSE, btr_pcur_get_rec(&pcur), page_nos, &n_stored); goto function_exit; } /* Set the bitmap bit denoting that the insert buffer contains buffered entries for this index page, if the bit is not set yet */ old_bit_value = ibuf_bitmap_page_get_bits(bitmap_page, page_no, IBUF_BITMAP_BUFFERED, &bitmap_mtr); if (!old_bit_value) { ibuf_bitmap_page_set_bits(bitmap_page, page_no, IBUF_BITMAP_BUFFERED, TRUE, &bitmap_mtr); } mtr_commit(&bitmap_mtr); cursor = btr_pcur_get_btr_cur(&pcur); if (mode == BTR_MODIFY_PREV) { err = btr_cur_optimistic_insert(BTR_NO_LOCKING_FLAG, cursor, ibuf_entry, &ins_rec, &dummy_big_rec, thr, &mtr); if (err == DB_SUCCESS) { /* Update the page max trx id field */ page_update_max_trx_id(buf_frame_align(ins_rec), thr_get_trx(thr)->id); } } else { ut_ad(mode == BTR_MODIFY_TREE); /* We acquire an x-latch to the root page before the insert, because a pessimistic insert releases the tree x-latch, which would cause the x-latching of the root after that to break the latching order. */ root = ibuf_tree_root_get(ibuf_data, space, &mtr); err = btr_cur_pessimistic_insert(BTR_NO_LOCKING_FLAG | BTR_NO_UNDO_LOG_FLAG, cursor, ibuf_entry, &ins_rec, &dummy_big_rec, thr, &mtr); if (err == DB_SUCCESS) { /* Update the page max trx id field */ page_update_max_trx_id(buf_frame_align(ins_rec), thr_get_trx(thr)->id); } ibuf_data_sizes_update(ibuf_data, root, &mtr); } function_exit: #ifdef UNIV_IBUF_DEBUG if (err == DB_SUCCESS) { ibuf_count_set(space, page_no, ibuf_count_get(space, page_no) + 1); } #endif if (mode == BTR_MODIFY_TREE) { ut_ad(ibuf_validate_low()); mutex_exit(&ibuf_mutex); mutex_exit(&ibuf_pessimistic_insert_mutex); } mtr_commit(&mtr); btr_pcur_close(&pcur); ibuf_exit(); mem_heap_free(heap); mutex_enter(&ibuf_mutex); if (err == DB_SUCCESS) { ibuf_data->empty = FALSE; ibuf_data->n_inserts++; } mutex_exit(&ibuf_mutex); if ((mode == BTR_MODIFY_TREE) && (err == DB_SUCCESS)) { ibuf_contract_after_insert(entry_size); } if (do_merge) { #ifdef UNIV_IBUF_DEBUG ut_a(n_stored <= IBUF_MAX_N_PAGES_MERGED); #endif buf_read_ibuf_merge_pages(FALSE, space, page_nos, n_stored); } return(err); } /************************************************************************* Makes an index insert to the insert buffer, instead of directly to the disk page, if this is possible. Does not do insert if the index is clustered or unique. */ ibool ibuf_insert( /*========*/ /* out: TRUE if success */ dtuple_t* entry, /* in: index entry to insert */ dict_index_t* index, /* in: index where to insert */ ulint space, /* in: space id where to insert */ ulint page_no,/* in: page number where to insert */ que_thr_t* thr) /* in: query thread */ { ulint err; ut_ad(dtuple_check_typed(entry)); if (index->type & DICT_CLUSTERED || index->type & DICT_UNIQUE) { return(FALSE); } if (rec_get_converted_size(entry) >= page_get_free_space_of_empty() / 2) { return(FALSE); } err = ibuf_insert_low(BTR_MODIFY_PREV, entry, index, space, page_no, thr); if (err == DB_FAIL) { err = ibuf_insert_low(BTR_MODIFY_TREE, entry, index, space, page_no, thr); } if (err == DB_SUCCESS) { #ifdef UNIV_IBUF_DEBUG /* printf("Ibuf insert for page no %lu of index %s\n", page_no, index->name); */ #endif return(TRUE); } else { ut_a(err == DB_STRONG_FAIL); return(FALSE); } } /************************************************************************ During merge, inserts to an index page a secondary index entry extracted from the insert buffer. */ static void ibuf_insert_to_index_page( /*======================*/ dtuple_t* entry, /* in: buffered entry to insert */ page_t* page, /* in: index page where the buffered entry should be placed */ mtr_t* mtr) /* in: mtr */ { page_cur_t page_cur; ulint low_match; rec_t* rec; page_t* bitmap_page; ulint old_bits; ut_ad(ibuf_inside()); ut_ad(dtuple_check_typed(entry)); low_match = page_cur_search(page, entry, PAGE_CUR_LE, &page_cur); if (low_match == dtuple_get_n_fields(entry)) { rec = page_cur_get_rec(&page_cur); btr_cur_del_unmark_for_ibuf(rec, mtr); } else { rec = page_cur_tuple_insert(&page_cur, entry, mtr); if (rec == NULL) { /* If the record did not fit, reorganize */ btr_page_reorganize(page, mtr); page_cur_search(page, entry, PAGE_CUR_LE, &page_cur); /* This time the record must fit */ if (!page_cur_tuple_insert(&page_cur, entry, mtr)) { printf( "Ibuf insert fails; page free %lu, dtuple size %lu\n", page_get_max_insert_size(page, 1), rec_get_converted_size(entry)); bitmap_page = ibuf_bitmap_get_map_page( buf_frame_get_space_id(page), buf_frame_get_page_no(page), mtr); old_bits = ibuf_bitmap_page_get_bits( bitmap_page, buf_frame_get_page_no(page), IBUF_BITMAP_FREE, mtr); printf("Bitmap bits %lu\n", old_bits); ut_error; } } } } /************************************************************************* Deletes from ibuf the record on which pcur is positioned. If we have to resort to a pessimistic delete, this function commits mtr and closes the cursor. */ static ibool ibuf_delete_rec( /*============*/ /* out: TRUE if mtr was committed and pcur closed in this operation */ ulint space, /* in: space id */ ulint page_no,/* in: index page number where the record should belong */ btr_pcur_t* pcur, /* in: pcur positioned on the record to delete, having latch mode BTR_MODIFY_LEAF */ dtuple_t* search_tuple, /* in: search tuple for entries of page_no */ mtr_t* mtr) /* in: mtr */ { ibool success; ibuf_data_t* ibuf_data; page_t* root; ulint err; ut_ad(ibuf_inside()); success = btr_cur_optimistic_delete(btr_pcur_get_btr_cur(pcur), mtr); if (success) { #ifdef UNIV_IBUF_DEBUG ibuf_count_set(space, page_no, ibuf_count_get(space, page_no) - 1); #endif return(FALSE); } /* We have to resort to a pessimistic delete from ibuf */ btr_pcur_store_position(pcur, mtr); btr_pcur_commit_specify_mtr(pcur, mtr); ibuf_data = fil_space_get_ibuf_data(space); mutex_enter(&ibuf_mutex); mtr_start(mtr); success = btr_pcur_restore_position(BTR_MODIFY_TREE, pcur, mtr); if (!success) { fprintf(stderr, "InnoDB: ERROR: Send the output to heikki.tuuri@innodb.com\n"); fprintf(stderr, "InnoDB: ibuf cursor restoration fails!\n"); fprintf(stderr, "InnoDB: ibuf record inserted to page %lu\n", page_no); rec_print(btr_pcur_get_rec(pcur)); rec_print(pcur->old_rec); dtuple_print(search_tuple); rec_print(page_rec_get_next(btr_pcur_get_rec(pcur))); mtr_commit(mtr); fprintf(stderr, "InnoDB: Validating insert buffer tree:\n"); ut_a(btr_validate_tree(ibuf_data->index->tree)); fprintf(stderr, "InnoDB: Ibuf tree ok\n"); } ut_a(success); root = ibuf_tree_root_get(ibuf_data, space, mtr); btr_cur_pessimistic_delete(&err, TRUE, btr_pcur_get_btr_cur(pcur), FALSE, mtr); ut_a(err == DB_SUCCESS); #ifdef UNIV_IBUF_DEBUG ibuf_count_set(space, page_no, ibuf_count_get(space, page_no) - 1); #endif ibuf_data_sizes_update(ibuf_data, root, mtr); ut_ad(ibuf_validate_low()); btr_pcur_commit_specify_mtr(pcur, mtr); btr_pcur_close(pcur); mutex_exit(&ibuf_mutex); return(TRUE); } /************************************************************************* When an index page is read from a disk to the buffer pool, this function inserts to the page the possible index entries buffered in the insert buffer. The entries are deleted from the insert buffer. If the page is not read, but created in the buffer pool, this function deletes its buffered entries from the insert buffer; there can exist entries for such a page if the page belonged to an index which subsequently was dropped. */ void ibuf_merge_or_delete_for_page( /*==========================*/ page_t* page, /* in: if page has been read from disk, pointer to the page x-latched, else NULL */ ulint space, /* in: space id of the index page */ ulint page_no)/* in: page number of the index page */ { mem_heap_t* heap; btr_pcur_t pcur; dtuple_t* entry; dtuple_t* search_tuple; rec_t* ibuf_rec; ibool closed; buf_block_t* block; page_t* bitmap_page; ibuf_data_t* ibuf_data; ibool success; ulint n_inserts; ulint volume; ulint old_bits; ulint new_bits; dulint max_trx_id; mtr_t mtr; if (srv_force_recovery >= SRV_FORCE_NO_IBUF_MERGE) { return; } #ifdef UNIV_LOG_DEBUG if (space % 2 != 0) { printf("No ibuf operation in a replicate space\n"); return; } #endif if (ibuf_fixed_addr_page(page_no) || fsp_descr_page(page_no) || trx_sys_hdr_page(space, page_no)) { return; } mtr_start(&mtr); bitmap_page = ibuf_bitmap_get_map_page(space, page_no, &mtr); if (!ibuf_bitmap_page_get_bits(bitmap_page, page_no, IBUF_BITMAP_BUFFERED, &mtr)) { /* No inserts buffered for this page */ mtr_commit(&mtr); return; } mtr_commit(&mtr); ibuf_data = fil_space_get_ibuf_data(space); ibuf_enter(); heap = mem_heap_create(512); search_tuple = ibuf_search_tuple_build(page_no, heap); if (page) { /* Move the ownership of the x-latch on the page to this OS thread, so that we can acquire a second x-latch on it. This is needed for the insert operations to the index page to pass the debug checks. */ block = buf_block_align(page); rw_lock_x_lock_move_ownership(&(block->lock)); ut_a(fil_page_get_type(page) == FIL_PAGE_INDEX); } n_inserts = 0; volume = 0; loop: mtr_start(&mtr); if (page) { success = buf_page_get_known_nowait(RW_X_LATCH, page, BUF_KEEP_OLD, IB__FILE__, __LINE__, &mtr); ut_a(success); buf_page_dbg_add_level(page, SYNC_TREE_NODE); } /* Position pcur in the insert buffer at the first entry for this index page */ btr_pcur_open_on_user_rec(ibuf_data->index, search_tuple, PAGE_CUR_GE, BTR_MODIFY_LEAF, &pcur, &mtr); if (!btr_pcur_is_on_user_rec(&pcur, &mtr)) { ut_ad(btr_pcur_is_after_last_in_tree(&pcur, &mtr)); goto reset_bit; } for (;;) { ut_ad(btr_pcur_is_on_user_rec(&pcur, &mtr)); ibuf_rec = btr_pcur_get_rec(&pcur); /* Check if the entry is for this index page */ if (ibuf_rec_get_page_no(ibuf_rec) != page_no) { if (page) { page_header_reset_last_insert(page, &mtr); } goto reset_bit; } if (page) { /* Now we have at pcur a record which should be inserted to the index page; NOTE that the call below copies pointers to fields in ibuf_rec, and we must keep the latch to the ibuf_rec page until the insertion is finished! */ max_trx_id = page_get_max_trx_id( buf_frame_align(ibuf_rec)); page_update_max_trx_id(page, max_trx_id); entry = ibuf_build_entry_from_ibuf_rec(ibuf_rec, heap); #ifdef UNIV_IBUF_DEBUG volume += rec_get_converted_size(entry) + page_dir_calc_reserved_space(1); ut_a(volume <= 4 * UNIV_PAGE_SIZE / IBUF_PAGE_SIZE_PER_FREE_SPACE); #endif ibuf_insert_to_index_page(entry, page, &mtr); } n_inserts++; /* Delete the record from ibuf */ closed = ibuf_delete_rec(space, page_no, &pcur, search_tuple, &mtr); if (closed) { /* Deletion was pessimistic and mtr was committed: we start from the beginning again */ goto loop; } if (btr_pcur_is_after_last_on_page(&pcur, &mtr)) { mtr_commit(&mtr); btr_pcur_close(&pcur); goto loop; } } reset_bit: #ifdef UNIV_IBUF_DEBUG if (ibuf_count_get(space, page_no) > 0) { /* btr_print_tree(ibuf_data->index->tree, 100); ibuf_print(); */ } #endif bitmap_page = ibuf_bitmap_get_map_page(space, page_no, &mtr); ibuf_bitmap_page_set_bits(bitmap_page, page_no, IBUF_BITMAP_BUFFERED, FALSE, &mtr); if (page) { old_bits = ibuf_bitmap_page_get_bits(bitmap_page, page_no, IBUF_BITMAP_FREE, &mtr); new_bits = ibuf_index_page_calc_free(page); #ifdef UNIV_IBUF_DEBUG /* printf("Old bits %lu new bits %lu max size %lu\n", old_bits, new_bits, page_get_max_insert_size_after_reorganize(page, 1)); */ #endif if (old_bits != new_bits) { ibuf_bitmap_page_set_bits(bitmap_page, page_no, IBUF_BITMAP_FREE, new_bits, &mtr); } } ibuf_data->n_merges++; ibuf_data->n_merged_recs += n_inserts; #ifdef UNIV_IBUF_DEBUG /* printf("Ibuf merge %lu records volume %lu to page no %lu\n", n_inserts, volume, page_no); */ #endif mtr_commit(&mtr); btr_pcur_close(&pcur); mem_heap_free(heap); ibuf_exit(); #ifdef UNIV_IBUF_DEBUG ut_a(ibuf_count_get(space, page_no) == 0); #endif } /********************************************************************** Validates the ibuf data structures when the caller owns ibuf_mutex. */ static ibool ibuf_validate_low(void) /*===================*/ /* out: TRUE if ok */ { ibuf_data_t* data; ulint sum_sizes; ut_ad(mutex_own(&ibuf_mutex)); sum_sizes = 0; data = UT_LIST_GET_FIRST(ibuf->data_list); while (data) { sum_sizes += data->size; data = UT_LIST_GET_NEXT(data_list, data); } ut_a(sum_sizes == ibuf->size); return(TRUE); } /********************************************************************** Prints info of ibuf. */ void ibuf_print(void) /*============*/ { ibuf_data_t* data; #ifdef UNIV_IBUF_DEBUG ulint i; #endif mutex_enter(&ibuf_mutex); data = UT_LIST_GET_FIRST(ibuf->data_list); while (data) { printf( "Ibuf for space %lu: size %lu, free list len %lu, seg size %lu,\n", data->space, data->size, data->free_list_len, data->seg_size); printf("%lu inserts, %lu merged recs, %lu merges\n", data->n_inserts, data->n_merged_recs, data->n_merges); #ifdef UNIV_IBUF_DEBUG for (i = 0; i < IBUF_COUNT_N_PAGES; i++) { if (ibuf_count_get(data->space, i) > 0) { printf("Ibuf count for page %lu is %lu\n", i, ibuf_count_get(data->space, i)); } } #endif data = UT_LIST_GET_NEXT(data_list, data); } mutex_exit(&ibuf_mutex); }