merge from 5.1 main minus merge jam pushes

scripts/mysql_install_db.pl.in

@@ -61,10 +61,15 @@ Usage: $0 [OPTIONS]
   --cross-bootstrap    For internal use.  Used when building the MySQL system
                        tables on a different host than the target.
   --datadir=path       The path to the MySQL data directory.
+  --defaults-extra-file=name
+                       Read this file after the global files are read.
+  --defaults-file=name Only read default options from the given file name.
   --force              Causes mysql_install_db to run even if DNS does not
                        work.  In that case, grant table entries that normally
                        use hostnames will use IP addresses.
+  --help               Display this help and exit.                     
   --ldata=path         The path to the MySQL data directory. Same as --datadir.
+  --no-defaults        Don't read default options from any option file.
   --rpm                For internal use.  This option is used by RPM files
                        during the MySQL installation process.
   --skip-name-resolve  Use IP addresses rather than hostnames when creating

scripts/mysql_install_db.sh

@@ -46,10 +46,15 @@ Usage: $0 [OPTIONS]
   --cross-bootstrap    For internal use.  Used when building the MySQL system
                        tables on a different host than the target.
   --datadir=path       The path to the MySQL data directory.
+  --defaults-extra-file=name
+                       Read this file after the global files are read.
+  --defaults-file=name Only read default options from the given file name.
   --force              Causes mysql_install_db to run even if DNS does not
                        work.  In that case, grant table entries that normally
                        use hostnames will use IP addresses.
+  --help               Display this help and exit.                     
   --ldata=path         The path to the MySQL data directory. Same as --datadir.
+  --no-defaults        Don't read default options from any option file.
   --rpm                For internal use.  This option is used by RPM files
                        during the MySQL installation process.
   --skip-name-resolve  Use IP addresses rather than hostnames when creating

storage/innobase/ibuf/ibuf0ibuf.c

@@ -1683,16 +1683,16 @@ ibuf_add_free_page(
 
 	page = buf_page_get(space, page_no, RW_X_LATCH, &mtr);
 
-#ifdef UNIV_SYNC_DEBUG
-	buf_page_dbg_add_level(page, SYNC_TREE_NODE_NEW);
-#endif /* UNIV_SYNC_DEBUG */
-
 	ibuf_enter();
 
 	mutex_enter(&ibuf_mutex);
 
 	root = ibuf_tree_root_get(ibuf_data, space, &mtr);
 
+#ifdef UNIV_SYNC_DEBUG
+	buf_page_dbg_add_level(page, SYNC_TREE_NODE_NEW);
+#endif /* UNIV_SYNC_DEBUG */
+
 	/* Add the page to the free list and update the ibuf size data */
 
 	flst_add_last(root + PAGE_HEADER + PAGE_BTR_IBUF_FREE_LIST,

storage/innodb_plugin/ChangeLog

+2011-09-12	The InnoDB Team
+
+	* row/row0sel.c:
+	Fix Bug#12601439 CONSISTENT READ FAILURE IN COLUMN PREFIX INDEX
+
+2011-09-08	The InnoDB Team
+
+	* btr/btr0cur.c, include/page0page.h, include/row0upd.ic:
+	Fix Bug#12948130 UNNECESSARY X-LOCKING OF ADAPTIVE HASH INDEX
+
+2011-09-06	The InnoDB Team
+
+	* buf/buf0buddy.c:
+	Fix Bug#12950803 62294: BUF_BUDDY_RELOCATE CALLS GETTIMEOFDAY
+	WHILE HOLDING BUFFER POOL MUTEX
+
+2011-09-06	The InnoDB Team
+
+	* include/trx0undo.h, trx/trx0rec.c, trx/trx0undo.c:
+	Fix Bug#12547647 UPDATE LOGGING COULD EXCEED LOG PAGE SIZE
+
 2011-08-29	The InnoDB Team
 
 	* btr/btr0btr.c, btr/btr0cur.c, fsp/fsp0fsp.c,
@@ -7,11 +28,6 @@
 	Fix Bug#12704861 Corruption after a crash during BLOB update
 	and other regressions from the fix of Bug#12612184
 
-2011-08-23	The InnoDB Team
-
-	* include/trx0undo.h, trx/trx0rec.c, trx/trx0undo.c:
-	Fix Bug#12547647 UPDATE LOGGING COULD EXCEED LOG PAGE SIZE
-
 2011-08-15	The InnoDB Team
 
 	* btr/btr0btr.c, btr/btr0cur.c, btr/btr0pcur.c, btr/btr0sea.c,

storage/innodb_plugin/btr/btr0cur.c

@@ -1727,6 +1727,7 @@ btr_cur_update_in_place(
 	roll_ptr_t	roll_ptr	= ut_dulint_zero;
 	trx_t*		trx;
 	ulint		was_delete_marked;
+	ibool		is_hashed;
 	mem_heap_t*	heap		= NULL;
 	ulint		offsets_[REC_OFFS_NORMAL_SIZE];
 	ulint*		offsets		= offsets_;
@@ -1768,7 +1769,21 @@ btr_cur_update_in_place(
 		return(err);
 	}
 
-	if (block->is_hashed) {
+	if (!(flags & BTR_KEEP_SYS_FLAG)) {
+		row_upd_rec_sys_fields(rec, NULL,
+				       index, offsets, trx, roll_ptr);
+	}
+
+	was_delete_marked = rec_get_deleted_flag(
+		rec, page_is_comp(buf_block_get_frame(block)));
+
+	is_hashed = block->is_hashed;
+
+	if (is_hashed) {
+		/* TO DO: Can we skip this if none of the fields
+		index->search_info->curr_n_fields
+		are being updated? */
+
 		/* The function row_upd_changes_ord_field_binary works only
 		if the update vector was built for a clustered index, we must
 		NOT call it if index is secondary */
@@ -1784,17 +1799,9 @@ btr_cur_update_in_place(
 		rw_lock_x_lock(&btr_search_latch);
 	}
 
-	if (!(flags & BTR_KEEP_SYS_FLAG)) {
-		row_upd_rec_sys_fields(rec, NULL,
-				       index, offsets, trx, roll_ptr);
-	}
-
-	was_delete_marked = rec_get_deleted_flag(
-		rec, page_is_comp(buf_block_get_frame(block)));
-
 	row_upd_rec_in_place(rec, index, offsets, update, page_zip);
 
-	if (block->is_hashed) {
+	if (is_hashed) {
 		rw_lock_x_unlock(&btr_search_latch);
 	}
 
@@ -2520,7 +2527,8 @@ btr_cur_parse_del_mark_set_clust_rec(
 
 		/* We do not need to reserve btr_search_latch, as the page
 		is only being recovered, and there cannot be a hash index to
-		it. */
+		it. Besides, these fields are being updated in place
+		and the adaptive hash index does not depend on them. */
 
 		btr_rec_set_deleted_flag(rec, page_zip, val);
 
@@ -2600,9 +2608,9 @@ btr_cur_del_mark_set_clust_rec(
 		return(err);
 	}
 
-	if (block->is_hashed) {
-		rw_lock_x_lock(&btr_search_latch);
-	}
+	/* The btr_search_latch is not needed here, because
+	the adaptive hash index does not depend on the delete-mark
+	and the delete-mark is being updated in place. */
 
 	page_zip = buf_block_get_page_zip(block);
 
@@ -2616,10 +2624,6 @@ btr_cur_del_mark_set_clust_rec(
 				       index, offsets, trx, roll_ptr);
 	}
 
-	if (block->is_hashed) {
-		rw_lock_x_unlock(&btr_search_latch);
-	}
-
 	btr_cur_del_mark_set_clust_rec_log(flags, rec, index, val, trx,
 					   roll_ptr, mtr);
 
@@ -2695,7 +2699,8 @@ btr_cur_parse_del_mark_set_sec_rec(
 
 		/* We do not need to reserve btr_search_latch, as the page
 		is only being recovered, and there cannot be a hash index to
-		it. */
+		it. Besides, the delete-mark flag is being updated in place
+		and the adaptive hash index does not depend on it. */
 
 		btr_rec_set_deleted_flag(rec, page_zip, val);
 	}
@@ -2743,16 +2748,11 @@ btr_cur_del_mark_set_sec_rec(
 	ut_ad(!!page_rec_is_comp(rec)
 	      == dict_table_is_comp(cursor->index->table));
 
-	if (block->is_hashed) {
-		rw_lock_x_lock(&btr_search_latch);
-	}
-
+	/* We do not need to reserve btr_search_latch, as the
+	delete-mark flag is being updated in place and the adaptive
+	hash index does not depend on it. */
 	btr_rec_set_deleted_flag(rec, buf_block_get_page_zip(block), val);
 
-	if (block->is_hashed) {
-		rw_lock_x_unlock(&btr_search_latch);
-	}
-
 	btr_cur_del_mark_set_sec_rec_log(rec, val, mtr);
 
 	return(DB_SUCCESS);
@@ -2772,8 +2772,11 @@ btr_cur_del_unmark_for_ibuf(
 					uncompressed */
 	mtr_t*		mtr)		/*!< in: mtr */
 {
-	/* We do not need to reserve btr_search_latch, as the page has just
-	been read to the buffer pool and there cannot be a hash index to it. */
+	/* We do not need to reserve btr_search_latch, as the page
+	has just been read to the buffer pool and there cannot be
+	a hash index to it.  Besides, the delete-mark flag is being
+	updated in place and the adaptive hash index does not depend
+	on it. */
 
 	btr_rec_set_deleted_flag(rec, page_zip, FALSE);
 

storage/innodb_plugin/buf/buf0buddy.c

@@ -327,7 +327,6 @@ buf_buddy_relocate(
 {
 	buf_page_t*	bpage;
 	const ulint	size	= BUF_BUDDY_LOW << i;
-	ullint		usec	= ut_time_us(NULL);
 	mutex_t*	mutex;
 	ulint		space;
 	ulint		page_no;
@@ -394,6 +393,7 @@ buf_buddy_relocate(
 
 	if (buf_page_can_relocate(bpage)) {
 		/* Relocate the compressed page. */
+		ullint	usec	= ut_time_us(NULL);
 		ut_a(bpage->zip.data == src);
 		memcpy(dst, src, size);
 		bpage->zip.data = dst;

storage/innodb_plugin/ibuf/ibuf0ibuf.c

@@ -1766,17 +1766,16 @@ ibuf_add_free_page(void)
 		block = buf_page_get(
 			IBUF_SPACE_ID, 0, page_no, RW_X_LATCH, &mtr);
 
-		buf_block_dbg_add_level(block, SYNC_IBUF_TREE_NODE_NEW);
-
+		ibuf_enter();
 
-		page = buf_block_get_frame(block);
-	}
+		mutex_enter(&ibuf_mutex);
 
-	ibuf_enter();
+		root = ibuf_tree_root_get(&mtr);
 
-	mutex_enter(&ibuf_mutex);
+		buf_block_dbg_add_level(block, SYNC_IBUF_TREE_NODE_NEW);
 
-	root = ibuf_tree_root_get(&mtr);
+		page = buf_block_get_frame(block);
+	}
 
 	/* Add the page to the free list and update the ibuf size data */
 

storage/innodb_plugin/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; a dulint; 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 */
 /*----*/

storage/innodb_plugin/include/row0upd.ic

@@ -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);

storage/innodb_plugin/row/row0mysql.c

@@ -3221,6 +3221,19 @@ check_next_foreign:
 			   "index_id CHAR;\n"
 			   "foreign_id CHAR;\n"
 			   "found INT;\n"
+
+			   "DECLARE CURSOR cur_fk IS\n"
+			   "SELECT ID FROM SYS_FOREIGN\n"
+			   "WHERE FOR_NAME = :table_name\n"
+			   "AND TO_BINARY(FOR_NAME)\n"
+			   "  = TO_BINARY(:table_name)\n"
+			   "LOCK IN SHARE MODE;\n"
+
+			   "DECLARE CURSOR cur_idx IS\n"
+			   "SELECT ID FROM SYS_INDEXES\n"
+			   "WHERE TABLE_ID = table_id\n"
+			   "LOCK IN SHARE MODE;\n"
+
 			   "BEGIN\n"
 			   "SELECT ID INTO table_id\n"
 			   "FROM SYS_TABLES\n"
@@ -3243,13 +3256,9 @@ check_next_foreign:
 			   "IF (:table_name = 'SYS_FOREIGN_COLS') THEN\n"
 			   "       found := 0;\n"
 			   "END IF;\n"
+			   "OPEN cur_fk;\n"
 			   "WHILE found = 1 LOOP\n"
-			   "       SELECT ID INTO foreign_id\n"
-			   "       FROM SYS_FOREIGN\n"
-			   "       WHERE FOR_NAME = :table_name\n"
-			   "               AND TO_BINARY(FOR_NAME)\n"
-			   "                 = TO_BINARY(:table_name)\n"
-			   "               LOCK IN SHARE MODE;\n"
+			   "       FETCH cur_fk INTO foreign_id;\n"
 			   "       IF (SQL % NOTFOUND) THEN\n"
 			   "               found := 0;\n"
 			   "       ELSE\n"
@@ -3259,12 +3268,11 @@ check_next_foreign:
 			   "               WHERE ID = foreign_id;\n"
 			   "       END IF;\n"
 			   "END LOOP;\n"
+			   "CLOSE cur_fk;\n"
 			   "found := 1;\n"
+			   "OPEN cur_idx;\n"
 			   "WHILE found = 1 LOOP\n"
-			   "       SELECT ID INTO index_id\n"
-			   "       FROM SYS_INDEXES\n"
-			   "       WHERE TABLE_ID = table_id\n"
-			   "       LOCK IN SHARE MODE;\n"
+			   "       FETCH cur_idx INTO index_id;\n"
 			   "       IF (SQL % NOTFOUND) THEN\n"
 			   "               found := 0;\n"
 			   "       ELSE\n"
@@ -3275,6 +3283,7 @@ check_next_foreign:
 			   "               AND TABLE_ID = table_id;\n"
 			   "       END IF;\n"
 			   "END LOOP;\n"
+			   "CLOSE cur_idx;\n"
 			   "DELETE FROM SYS_COLUMNS\n"
 			   "WHERE TABLE_ID = table_id;\n"
 			   "DELETE FROM SYS_TABLES\n"

storage/innodb_plugin/row/row0sel.c

 /*****************************************************************************
 
-Copyright (c) 1997, 2010, Innobase Oy. All Rights Reserved.
+Copyright (c) 1997, 2011, Oracle and/or its affiliates. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
@@ -101,12 +101,17 @@ row_sel_sec_rec_is_for_blob(
 	ulint		clust_len,	/*!< in: length of clust_field */
 	const byte*	sec_field,	/*!< in: column in secondary index */
 	ulint		sec_len,	/*!< in: length of sec_field */
+	ulint		prefix_len,	/*!< in: index column prefix length
+					in bytes */
 	ulint		zip_size)	/*!< in: compressed page size, or 0 */
 {
 	ulint	len;
 	byte	buf[DICT_MAX_INDEX_COL_LEN];
 
 	ut_a(clust_len >= BTR_EXTERN_FIELD_REF_SIZE);
+	ut_ad(prefix_len >= sec_len);
+	ut_ad(prefix_len > 0);
+	ut_a(prefix_len <= sizeof buf);
 
 	if (UNIV_UNLIKELY
 	    (!memcmp(clust_field + clust_len - BTR_EXTERN_FIELD_REF_SIZE,
@@ -118,7 +123,7 @@ row_sel_sec_rec_is_for_blob(
 		return(FALSE);
 	}
 
-	len = btr_copy_externally_stored_field_prefix(buf, sizeof buf,
+	len = btr_copy_externally_stored_field_prefix(buf, prefix_len,
 						      zip_size,
 						      clust_field, clust_len);
 
@@ -132,7 +137,7 @@ row_sel_sec_rec_is_for_blob(
 	}
 
 	len = dtype_get_at_most_n_mbchars(prtype, mbminlen, mbmaxlen,
-					  sec_len, len, (const char*) buf);
+					  prefix_len, len, (const char*) buf);
 
 	return(!cmp_data_data(mtype, prtype, buf, len, sec_field, sec_len));
 }
@@ -219,11 +224,20 @@ row_sel_sec_rec_is_for_clust_rec(
 
 			if (rec_offs_nth_extern(clust_offs, clust_pos)
 			    && len < sec_len) {
+				/* This function should never be
+				invoked on an Antelope format table,
+				because they should always contain
+				enough prefix in the clustered index
+				record. */
+				ut_ad(dict_table_get_format(clust_index->table)
+				      >= DICT_TF_FORMAT_ZIP);
+
 				if (!row_sel_sec_rec_is_for_blob(
 					    col->mtype, col->prtype,
 					    col->mbminlen, col->mbmaxlen,
 					    clust_field, clust_len,
 					    sec_field, sec_len,
+					    ifield->prefix_len,
 					    dict_table_zip_size(
 						    clust_index->table))) {
 					goto inequal;
@@ -494,7 +508,7 @@ sel_col_prefetch_buf_alloc(
 		sel_buf = column->prefetch_buf + i;
 
 		sel_buf->data = NULL;
-
+		sel_buf->len = 0;
 		sel_buf->val_buf_size = 0;
 	}
 }
@@ -519,6 +533,8 @@ sel_col_prefetch_buf_free(
 			mem_free(sel_buf->data);
 		}
 	}
+
+	mem_free(prefetch_buf);
 }
 
 /*********************************************************************//**

storage/innodb_plugin/sync/sync0sync.c

@@ -1258,8 +1258,7 @@ sync_thread_add_level(
 		break;
 	case SYNC_IBUF_INDEX_TREE:
 		if (sync_thread_levels_contain(array, SYNC_FSP)) {
-			ut_a(sync_thread_levels_g(
-				     array, SYNC_FSP_PAGE - 1, TRUE));
+			ut_a(sync_thread_levels_g(array, level - 1, TRUE));
 		} else {
 			ut_a(sync_thread_levels_g(
 				     array, SYNC_IBUF_TREE_NODE - 1, TRUE));

storage/innodb_plugin/trx/trx0rec.c

@@ -1099,7 +1099,7 @@ trx_undo_rec_get_partial_row(
 /***********************************************************************//**
 Erases the unused undo log page end.
 @return TRUE if the page contained something, FALSE if it was empty */
-static __attribute__((nonnull, warn_unused_result))
+static __attribute__((nonnull))
 ibool
 trx_undo_erase_page_end(
 /*====================*/
@@ -1110,16 +1110,11 @@ trx_undo_erase_page_end(
 
 	first_free = mach_read_from_2(undo_page + TRX_UNDO_PAGE_HDR
 				      + TRX_UNDO_PAGE_FREE);
-	if (first_free == TRX_UNDO_PAGE_HDR + TRX_UNDO_PAGE_HDR_SIZE) {
-		/* This was an empty page to begin with.
-		Do nothing here; the caller should free the page. */
-		return(FALSE);
-	}
 	memset(undo_page + first_free, 0xff,
 	       (UNIV_PAGE_SIZE - FIL_PAGE_DATA_END) - first_free);
 
 	mlog_write_initial_log_record(undo_page, MLOG_UNDO_ERASE_END, mtr);
-	return(TRUE);
+	return(first_free != TRX_UNDO_PAGE_HDR + TRX_UNDO_PAGE_HDR_SIZE);
 }
 
 /***********************************************************//**
@@ -1141,11 +1136,7 @@ trx_undo_parse_erase_page_end(
 		return(ptr);
 	}
 
-	if (!trx_undo_erase_page_end(page, mtr)) {
-		/* The function trx_undo_erase_page_end() should not
-		have done anything to an empty page. */
-		ut_ad(0);
-	}
+	trx_undo_erase_page_end(page, mtr);
 
 	return(ptr);
 }
@@ -1290,6 +1281,18 @@ trx_undo_report_row_operation(
 				undo page. Discard the freshly allocated
 				page and return an error. */
 
+				/* When we remove a page from an undo
+				log, this is analogous to a
+				pessimistic insert in a B-tree, and we
+				must reserve the counterpart of the
+				tree latch, which is the rseg
+				mutex. We must commit the mini-transaction
+				first, because it may be holding lower-level
+				latches, such as SYNC_FSP and SYNC_FSP_PAGE. */
+
+				mtr_commit(&mtr);
+				mtr_start(&mtr);
+
 				mutex_enter(&rseg->mutex);
 				trx_undo_free_last_page(trx, undo, &mtr);
 				mutex_exit(&rseg->mutex);