Bug#13721257 RACE CONDITION IN UPDATES OR INSERTS OF WIDE RECORDS

This bug was originally filed and fixed as Bug#12612184. The original
fix was buggy, and it was patched by Bug#12704861. Also that patch was
buggy (potentially breaking crash recovery), and both fixes were
reverted.

This fix was not ported to the built-in InnoDB of MySQL 5.1, because
the function signatures of many core functions are different from
InnoDB Plugin and later versions. The block allocation routines and
their callers would have to changed so that they handle block
descriptors instead of page frames.

When a record is updated so that its size grows, non-updated columns
can be selected for external (off-page) storage. The bug is that the
initially inserted updated record contains an all-zero BLOB pointer to
the field that was not updated. Only after the BLOB pages have been
allocated and written, the valid pointer can be written to the record.

Between the release of the page latch in mtr_commit(mtr) after
btr_cur_pessimistic_update() and the re-latching of the page in
btr_pcur_restore_position(), other threads can see the invalid BLOB
pointer consisting of 20 zero bytes. Moreover, if the system crashes
at this point, the situation could persist after crash recovery, and
the contents of the non-updated column would be permanently lost.

The problem is amplified by the ROW_FORMAT=DYNAMIC and
ROW_FORMAT=COMPRESSED that were introduced in
innodb_file_format=barracuda in InnoDB Plugin, but the bug does exist
in all InnoDB versions.

The fix is as follows. After a pessimistic B-tree operation that needs
to write out off-page columns, allocate the pages for these columns in
the mini-transaction that performed the B-tree operation (btr_mtr),
but write the pages in a separate mini-transaction (blob_mtr). Do
mtr_commit(blob_mtr) before mtr_commit(btr_mtr). A quirk: Do not reuse
pages that were previously freed in btr_mtr. Only write the off-page
columns to 'fresh' pages.

In this way, crash recovery will see redo log entries for blob_mtr
before any redo log entry for btr_mtr. It will apply the BLOB page
writes to pages that were marked free at that point. If crash recovery
fails to see all of the btr_mtr redo log, there will be some
unreachable BLOB data in free pages, but the B-tree will be in a
consistent state.

btr_page_alloc_low(): Renamed from btr_page_alloc(). Add the parameter
init_mtr. Return an allocated block, or NULL. If init_mtr!=mtr but
the page was already X-latched in mtr, do not initialize the page.

btr_page_alloc(): Wrapper for btr_page_alloc_for_ibuf() and
btr_page_alloc_low().

btr_page_free(): Add a debug assertion that the page was a B-tree page.

btr_lift_page_up(): Return the father block.

btr_compress(), btr_cur_compress_if_useful(): Add the parameter ibool
adjust, for adjusting the cursor position.

btr_cur_pessimistic_update(): Preserve the cursor position when
big_rec will be written and the new flag BTR_KEEP_POS_FLAG is defined.
Remove a duplicate rec_get_offsets() call. Keep the X-latch on
index->lock when big_rec is needed.

btr_store_big_rec_extern_fields(): Replace update_inplace with
an operation code, and local_mtr with btr_mtr. When not doing a
fresh insert and btr_mtr has freed pages, put aside any pages that
were previously X-latched in btr_mtr, and free the pages after
writing out all data. The data must be written to 'fresh' pages,
because btr_mtr will be committed and written to the redo log after
the BLOB writes have been written to the redo log.

btr_blob_op_is_update(): Check if an operation passed to
btr_store_big_rec_extern_fields() is an update or insert-by-update.

fseg_alloc_free_page_low(), fsp_alloc_free_page(),
fseg_alloc_free_extent(), fseg_alloc_free_page_general(): Add the
parameter init_mtr. Return an allocated block, or NULL. If
init_mtr!=mtr but the page was already X-latched in mtr, do not
initialize the page.

xdes_get_descriptor_with_space_hdr(): Assert that the file space
header is being X-latched.

fsp_alloc_from_free_frag(): Refactored from fsp_alloc_free_page().

fsp_page_create(): New function, for allocating, X-latching and
potentially initializing a page. If init_mtr!=mtr but the page was
already X-latched in mtr, do not initialize the page.

fsp_free_page(): Add ut_ad(0) to the error outcomes.

fsp_free_page(), fseg_free_page_low(): Increment mtr->n_freed_pages.

fsp_alloc_seg_inode_page(), fseg_create_general(): Assert that the
page was not previously X-latched in the mini-transaction. A file
segment or inode page should never be allocated in the middle of an
mini-transaction that frees pages, such as btr_cur_pessimistic_delete().

fseg_alloc_free_page_low(): If the hinted page was allocated, skip the
check if the tablespace should be extended. Return NULL instead of
FIL_NULL on failure. Remove the flag frag_page_allocated. Instead,
return directly, because the page would already have been initialized.

fseg_find_free_frag_page_slot() would return ULINT_UNDEFINED on error,
not FIL_NULL. Correct a bogus assertion.

fseg_alloc_free_page(): Redefine as a wrapper macro around
fseg_alloc_free_page_general().

buf_block_buf_fix_inc(): Move the definition from the buf0buf.ic to
buf0buf.h, so that it can be called from other modules.

mtr_t: Add n_freed_pages (number of pages that have been freed).

page_rec_get_nth_const(), page_rec_get_nth(): The inverse function of
page_rec_get_n_recs_before(), get the nth record of the record
list. This is faster than iterating the linked list. Refactored from
page_get_middle_rec().

trx_undo_rec_copy(): Add a debug assertion for the length.

trx_undo_add_page(): Return a block descriptor or NULL instead of a
page number or FIL_NULL.

trx_undo_report_row_operation(): Add debug assertions.

trx_sys_create_doublewrite_buf(): Assert that each page was not
previously X-latched.

page_cur_insert_rec_zip_reorg(): Make use of page_rec_get_nth().

row_ins_clust_index_entry_by_modify(): Pass BTR_KEEP_POS_FLAG, so that
the repositioning of the cursor can be avoided.

row_ins_index_entry_low(): Add DEBUG_SYNC points before and after
writing off-page columns. If inserting by updating a delete-marked
record, do not reposition the cursor or commit the mini-transaction
before writing the off-page columns.

row_build(): Tighten a debug assertion about null BLOB pointers.

row_upd_clust_rec(): Add DEBUG_SYNC points before and after writing
off-page columns. Do not reposition the cursor or commit the
mini-transaction before writing the off-page columns.

rb:939 approved by Jimmy Yang

mysql-test/suite/innodb_plugin/r/innodb-blob.result

+CREATE TABLE t1 (a INT PRIMARY KEY, b TEXT) ENGINE=InnoDB;
+CREATE TABLE t2 (a INT PRIMARY KEY) ENGINE=InnoDB;
+CREATE TABLE t3 (a INT PRIMARY KEY, b TEXT, c TEXT) ENGINE=InnoDB;
+INSERT INTO t1 VALUES (1,REPEAT('a',30000)),(2,REPEAT('b',40000));
+SET DEBUG_SYNC='before_row_upd_extern SIGNAL have_latch WAIT_FOR go1';
+BEGIN;
+UPDATE t1 SET a=a+2;
+ROLLBACK;
+BEGIN;
+UPDATE t1 SET b=CONCAT(b,'foo');
+SET DEBUG_SYNC='now WAIT_FOR have_latch';
+SELECT a, RIGHT(b,20) FROM t1;
+SET DEBUG_SYNC='now SIGNAL go1';
+a	RIGHT(b,20)
+1	aaaaaaaaaaaaaaaaaaaa
+2	bbbbbbbbbbbbbbbbbbbb
+SET DEBUG='+d,row_ins_extern_checkpoint';
+SET DEBUG_SYNC='before_row_ins_extern_latch SIGNAL rec_not_blob WAIT_FOR crash';
+ROLLBACK;
+BEGIN;
+INSERT INTO t1 VALUES (3,REPEAT('c',50000));
+SET DEBUG_SYNC='now WAIT_FOR rec_not_blob';
+SET SESSION TRANSACTION ISOLATION LEVEL READ UNCOMMITTED;
+SELECT @@tx_isolation;
+@@tx_isolation
+READ-UNCOMMITTED
+SELECT a, RIGHT(b,20) FROM t1;
+a	RIGHT(b,20)
+1	aaaaaaaaaaaaaaaaaaaa
+2	bbbbbbbbbbbbbbbbbbbb
+SELECT a FROM t1;
+a
+1
+2
+3
+SET DEBUG='+d,crash_commit_before';
+INSERT INTO t2 VALUES (42);
+ERROR HY000: Lost connection to MySQL server during query
+ERROR HY000: Lost connection to MySQL server during query
+CHECK TABLE t1;
+Table	Op	Msg_type	Msg_text
+test.t1	check	status	OK
+INSERT INTO t3 VALUES
+(1,REPEAT('d',7000),REPEAT('e',100)),
+(2,REPEAT('g',7000),REPEAT('h',100));
+SET DEBUG_SYNC='before_row_upd_extern SIGNAL have_latch WAIT_FOR go';
+UPDATE t3 SET c=REPEAT('f',3000) WHERE a=1;
+SET DEBUG_SYNC='now WAIT_FOR have_latch';
+SET SESSION TRANSACTION ISOLATION LEVEL READ UNCOMMITTED;
+SELECT @@tx_isolation;
+@@tx_isolation
+READ-UNCOMMITTED
+SELECT a, RIGHT(b,20), RIGHT(c,20) FROM t3;
+SET DEBUG_SYNC='now SIGNAL go';
+a	RIGHT(b,20)	RIGHT(c,20)
+1	dddddddddddddddddddd	ffffffffffffffffffff
+2	gggggggggggggggggggg	hhhhhhhhhhhhhhhhhhhh
+CHECK TABLE t1,t2,t3;
+Table	Op	Msg_type	Msg_text
+test.t1	check	status	OK
+test.t2	check	status	OK
+test.t3	check	status	OK
+BEGIN;
+INSERT INTO t2 VALUES (347);
+SET DEBUG='+d,row_upd_extern_checkpoint';
+SET DEBUG_SYNC='before_row_upd_extern SIGNAL have_latch WAIT_FOR crash';
+UPDATE t3 SET c=REPEAT('i',3000) WHERE a=2;
+SET DEBUG_SYNC='now WAIT_FOR have_latch';
+SELECT info FROM information_schema.processlist
+WHERE state = 'debug sync point: before_row_upd_extern';
+info
+UPDATE t3 SET c=REPEAT('i',3000) WHERE a=2
+SET DEBUG='+d,crash_commit_before';
+COMMIT;
+ERROR HY000: Lost connection to MySQL server during query
+ERROR HY000: Lost connection to MySQL server during query
+CHECK TABLE t1,t2,t3;
+Table	Op	Msg_type	Msg_text
+test.t1	check	status	OK
+test.t2	check	status	OK
+test.t3	check	status	OK
+SELECT a, RIGHT(b,20), RIGHT(c,20) FROM t3;
+a	RIGHT(b,20)	RIGHT(c,20)
+1	dddddddddddddddddddd	ffffffffffffffffffff
+2	gggggggggggggggggggg	hhhhhhhhhhhhhhhhhhhh
+SELECT a FROM t3;
+a
+1
+2
+BEGIN;
+INSERT INTO t2 VALUES (33101);
+SET DEBUG='+d,row_upd_extern_checkpoint';
+SET DEBUG_SYNC='after_row_upd_extern SIGNAL have_latch WAIT_FOR crash';
+UPDATE t3 SET c=REPEAT('j',3000) WHERE a=2;
+SET DEBUG_SYNC='now WAIT_FOR have_latch';
+SELECT info FROM information_schema.processlist
+WHERE state = 'debug sync point: after_row_upd_extern';
+info
+UPDATE t3 SET c=REPEAT('j',3000) WHERE a=2
+SET DEBUG='+d,crash_commit_before';
+COMMIT;
+ERROR HY000: Lost connection to MySQL server during query
+ERROR HY000: Lost connection to MySQL server during query
+CHECK TABLE t1,t2,t3;
+Table	Op	Msg_type	Msg_text
+test.t1	check	status	OK
+test.t2	check	status	OK
+test.t3	check	status	OK
+SELECT a, RIGHT(b,20), RIGHT(c,20) FROM t3;
+a	RIGHT(b,20)	RIGHT(c,20)
+1	dddddddddddddddddddd	ffffffffffffffffffff
+2	gggggggggggggggggggg	hhhhhhhhhhhhhhhhhhhh
+SELECT a FROM t3;
+a
+1
+2
+SELECT * FROM t2;
+a
+DROP TABLE t1,t2,t3;

mysql-test/suite/innodb_plugin/t/innodb-blob.test

+# Bug#13721257 RACE CONDITION IN UPDATES OR INSERTS OF WIDE RECORDS
+# Test what happens when a record is inserted or updated so that some
+# columns are stored off-page.
+
+--source include/have_innodb_plugin.inc
+
+# DEBUG_SYNC must be compiled in.
+--source include/have_debug_sync.inc
+
+# Valgrind would complain about memory leaks when we crash on purpose.
+--source include/not_valgrind.inc
+# Embedded server does not support crashing
+--source include/not_embedded.inc
+# Avoid CrashReporter popup on Mac
+--source include/not_crashrep.inc
+# InnoDB Plugin cannot use DEBUG_SYNC on Windows
+--source include/not_windows.inc
+
+CREATE TABLE t1 (a INT PRIMARY KEY, b TEXT) ENGINE=InnoDB;
+CREATE TABLE t2 (a INT PRIMARY KEY) ENGINE=InnoDB;
+CREATE TABLE t3 (a INT PRIMARY KEY, b TEXT, c TEXT) ENGINE=InnoDB;
+
+INSERT INTO t1 VALUES (1,REPEAT('a',30000)),(2,REPEAT('b',40000));
+SET DEBUG_SYNC='before_row_upd_extern SIGNAL have_latch WAIT_FOR go1';
+BEGIN;
+# This will not block, because it will not store new BLOBs.
+UPDATE t1 SET a=a+2;
+ROLLBACK;
+BEGIN;
+--send
+UPDATE t1 SET b=CONCAT(b,'foo');
+
+connect (con1,localhost,root,,);
+SET DEBUG_SYNC='now WAIT_FOR have_latch';
+
+# this one should block due to the clustered index tree and leaf page latches
+--send
+SELECT a, RIGHT(b,20) FROM t1;
+
+connect (con2,localhost,root,,);
+
+# Check that the above SELECT is blocked
+let $wait_condition=
+  select count(*) = 1 from information_schema.processlist
+  where state = 'Sending data' and
+        info = 'SELECT a, RIGHT(b,20) FROM t1';
+--source include/wait_condition.inc
+
+SET DEBUG_SYNC='now SIGNAL go1';
+
+connection con1;
+reap;
+connection default;
+reap;
+SET DEBUG='+d,row_ins_extern_checkpoint';
+SET DEBUG_SYNC='before_row_ins_extern_latch SIGNAL rec_not_blob WAIT_FOR crash';
+ROLLBACK;
+BEGIN;
+--send
+INSERT INTO t1 VALUES (3,REPEAT('c',50000));
+
+connection con1;
+SET DEBUG_SYNC='now WAIT_FOR rec_not_blob';
+SET SESSION TRANSACTION ISOLATION LEVEL READ UNCOMMITTED;
+SELECT @@tx_isolation;
+
+# this one should see (3,NULL_BLOB)
+SELECT a, RIGHT(b,20) FROM t1;
+SELECT a FROM t1;
+
+# Request a crash, and restart the server.
+SET DEBUG='+d,crash_commit_before';
+--exec echo "restart" > $MYSQLTEST_VARDIR/tmp/mysqld.1.expect
+--error 2013
+INSERT INTO t2 VALUES (42);
+
+disconnect con1;
+disconnect con2;
+connection default;
+# This connection should notice the crash as well.
+--error 2013
+reap;
+
+# Write file to make mysql-test-run.pl restart the server
+--enable_reconnect
+--source include/wait_until_connected_again.inc
+--disable_reconnect
+
+CHECK TABLE t1;
+
+INSERT INTO t3 VALUES
+       (1,REPEAT('d',7000),REPEAT('e',100)),
+       (2,REPEAT('g',7000),REPEAT('h',100));
+SET DEBUG_SYNC='before_row_upd_extern SIGNAL have_latch WAIT_FOR go';
+# This should move column b off-page.
+--send
+UPDATE t3 SET c=REPEAT('f',3000) WHERE a=1;
+
+connect (con1,localhost,root,,);
+SET DEBUG_SYNC='now WAIT_FOR have_latch';
+SET SESSION TRANSACTION ISOLATION LEVEL READ UNCOMMITTED;
+SELECT @@tx_isolation;
+
+# this one should block
+-- send
+SELECT a, RIGHT(b,20), RIGHT(c,20) FROM t3;
+
+connect (con2,localhost,root,,);
+
+# Check that the above SELECT is blocked
+let $wait_condition=
+  select count(*) = 1 from information_schema.processlist
+  where state = 'Sending data' and
+        info = 'SELECT a, RIGHT(b,20), RIGHT(c,20) FROM t3';
+--source include/wait_condition.inc
+
+SET DEBUG_SYNC='now SIGNAL go';
+
+connection con1;
+reap;
+disconnect con1;
+
+connection default;
+reap;
+
+CHECK TABLE t1,t2,t3;
+
+connection con2;
+BEGIN;
+INSERT INTO t2 VALUES (347);
+connection default;
+
+# The row_upd_extern_checkpoint was removed in Bug#13721257,
+# because the mini-transaction of the B-tree modification would
+# remain open while we are writing the off-page columns and are
+# stuck in the DEBUG_SYNC. A checkpoint involves a flush, which
+# would wait for the buffer-fix to cease.
+SET DEBUG='+d,row_upd_extern_checkpoint';
+SET DEBUG_SYNC='before_row_upd_extern SIGNAL have_latch WAIT_FOR crash';
+# This should move column b off-page.
+--send
+UPDATE t3 SET c=REPEAT('i',3000) WHERE a=2;
+
+connection con2;
+SET DEBUG_SYNC='now WAIT_FOR have_latch';
+
+# Check that the above UPDATE is blocked
+SELECT info FROM information_schema.processlist
+WHERE state = 'debug sync point: before_row_upd_extern';
+
+# Request a crash, and restart the server.
+SET DEBUG='+d,crash_commit_before';
+--exec echo "restart" > $MYSQLTEST_VARDIR/tmp/mysqld.1.expect
+--error 2013
+COMMIT;
+
+disconnect con2;
+connection default;
+# This connection should notice the crash as well.
+--error 2013
+reap;
+
+# Write file to make mysql-test-run.pl restart the server
+--enable_reconnect
+--source include/wait_until_connected_again.inc
+--disable_reconnect
+
+CHECK TABLE t1,t2,t3;
+SELECT a, RIGHT(b,20), RIGHT(c,20) FROM t3;
+SELECT a FROM t3;
+
+connect (con2,localhost,root,,);
+BEGIN;
+INSERT INTO t2 VALUES (33101);
+connection default;
+
+# The row_upd_extern_checkpoint was removed in Bug#13721257,
+# because the mini-transaction of the B-tree modification would
+# remain open while we are writing the off-page columns and are
+# stuck in the DEBUG_SYNC. A checkpoint involves a flush, which
+# would wait for the buffer-fix to cease.
+SET DEBUG='+d,row_upd_extern_checkpoint';
+SET DEBUG_SYNC='after_row_upd_extern SIGNAL have_latch WAIT_FOR crash';
+# This should move column b off-page.
+--send
+UPDATE t3 SET c=REPEAT('j',3000) WHERE a=2;
+
+connection con2;
+SET DEBUG_SYNC='now WAIT_FOR have_latch';
+
+# Check that the above UPDATE is blocked
+SELECT info FROM information_schema.processlist
+WHERE state = 'debug sync point: after_row_upd_extern';
+
+# Request a crash, and restart the server.
+SET DEBUG='+d,crash_commit_before';
+--exec echo "restart" > $MYSQLTEST_VARDIR/tmp/mysqld.1.expect
+--error 2013
+COMMIT;
+
+disconnect con2;
+connection default;
+# This connection should notice the crash as well.
+--error 2013
+reap;
+
+# Write file to make mysql-test-run.pl restart the server
+--enable_reconnect
+--source include/wait_until_connected_again.inc
+--disable_reconnect
+
+CHECK TABLE t1,t2,t3;
+SELECT a, RIGHT(b,20), RIGHT(c,20) FROM t3;
+SELECT a FROM t3;
+
+SELECT * FROM t2;
+
+DROP TABLE t1,t2,t3;

storage/innodb_plugin/ChangeLog

+2012-02-15	The InnoDB Team
+
+	* btr/btr0btr.c, btr/btr0cur.c, fsp/fsp0fsp.c, ibuf/ibuf0ibuf.c,
+	include/btr0btr.h, include/btr0cur.h, include/btr0cur.ic,
+	include/buf0buf.h, include/buf0buf.ic, include/fsp0fsp.h,
+	include/mtr0mtr.h, include/mtr0mtr.ic, include/page0page.h,
+	include/page0page.ic, include/trx0rec.ic, include/trx0undo.h,
+	mtr/mtr0mtr.c, page/page0cur.c, page/page0page.c, row/row0ins.c,
+	row/row0row.c, row/row0upd.c, trx/trx0rec.c, trx/trx0sys.c,
+	trx/trx0undo.c:
+	Fix Bug#13721257 RACE CONDITION IN UPDATES OR INSERTS OF WIDE RECORDS
+
 2012-02-06	The InnoDB Team
 
 	* handler/ha_innodb.cc:
-	Fix Bug #11754376 45976: INNODB LOST FILES FOR TEMPORARY TABLES ON
+	Fix Bug#11754376 45976: INNODB LOST FILES FOR TEMPORARY TABLES ON
 	GRACEFUL SHUTDOWN
 
 2012-01-30	The InnoDB Team

storage/innodb_plugin/btr/btr0btr.c

 /*****************************************************************************
 
-Copyright (c) 1994, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1994, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -906,28 +906,31 @@ btr_page_alloc_for_ibuf(
 /**************************************************************//**
 Allocates a new file page to be used in an index tree. NOTE: we assume
 that the caller has made the reservation for free extents!
-@return	new allocated block, x-latched; NULL if out of space */
-UNIV_INTERN
+@retval NULL if no page could be allocated
+@retval block, rw_lock_x_lock_count(&block->lock) == 1 if allocation succeeded
+(init_mtr == mtr, or the page was not previously freed in mtr)
+@retval block (not allocated or initialized) otherwise */
+static __attribute__((nonnull, warn_unused_result))
 buf_block_t*
-btr_page_alloc(
-/*===========*/
+btr_page_alloc_low(
+/*===============*/
 	dict_index_t*	index,		/*!< in: index */
 	ulint		hint_page_no,	/*!< in: hint of a good page */
 	byte		file_direction,	/*!< in: direction where a possible
 					page split is made */
 	ulint		level,		/*!< in: level where the page is placed
 					in the tree */
-	mtr_t*		mtr)		/*!< in: mtr */
+	mtr_t*		mtr,		/*!< in/out: mini-transaction
+					for the allocation */
+	mtr_t*		init_mtr)	/*!< in/out: mtr or another
+					mini-transaction in which the
+					page should be initialized.
+					If init_mtr!=mtr, but the page
+					is already X-latched in mtr, do
+					not initialize the page. */
 {
 	fseg_header_t*	seg_header;
 	page_t*		root;
-	buf_block_t*	new_block;
-	ulint		new_page_no;
-
-	if (dict_index_is_ibuf(index)) {
-
-		return(btr_page_alloc_for_ibuf(index, mtr));
-	}
 
 	root = btr_root_get(index, mtr);
 
@@ -941,17 +944,47 @@ btr_page_alloc(
 	reservation for free extents, and thus we know that a page can
 	be allocated: */
 
-	new_page_no = fseg_alloc_free_page_general(seg_header, hint_page_no,
-						   file_direction, TRUE, mtr);
-	if (new_page_no == FIL_NULL) {
+	return(fseg_alloc_free_page_general(
+		       seg_header, hint_page_no, file_direction,
+		       TRUE, mtr, init_mtr));
+}
 
-		return(NULL);
+/**************************************************************//**
+Allocates a new file page to be used in an index tree. NOTE: we assume
+that the caller has made the reservation for free extents!
+@retval NULL if no page could be allocated
+@retval block, rw_lock_x_lock_count(&block->lock) == 1 if allocation succeeded
+(init_mtr == mtr, or the page was not previously freed in mtr)
+@retval block (not allocated or initialized) otherwise */
+UNIV_INTERN
+buf_block_t*
+btr_page_alloc(
+/*===========*/
+	dict_index_t*	index,		/*!< in: index */
+	ulint		hint_page_no,	/*!< in: hint of a good page */
+	byte		file_direction,	/*!< in: direction where a possible
+					page split is made */
+	ulint		level,		/*!< in: level where the page is placed
+					in the tree */
+	mtr_t*		mtr,		/*!< in/out: mini-transaction
+					for the allocation */
+	mtr_t*		init_mtr)	/*!< in/out: mini-transaction
+					for x-latching and initializing
+					the page */
+{
+	buf_block_t*	new_block;
+
+	if (dict_index_is_ibuf(index)) {
+
+		return(btr_page_alloc_for_ibuf(index, mtr));
 	}
 
-	new_block = buf_page_get(dict_index_get_space(index),
-				 dict_table_zip_size(index->table),
-				 new_page_no, RW_X_LATCH, mtr);
-	buf_block_dbg_add_level(new_block, SYNC_TREE_NODE_NEW);
+	new_block = btr_page_alloc_low(
+		index, hint_page_no, file_direction, level, mtr, init_mtr);
+
+	if (new_block) {
+		buf_block_dbg_add_level(new_block, SYNC_TREE_NODE_NEW);
+	}
 
 	return(new_block);
 }
@@ -1087,10 +1120,10 @@ btr_page_free(
 	buf_block_t*	block,	/*!< in: block to be freed, x-latched */
 	mtr_t*		mtr)	/*!< in: mtr */
 {
-	ulint		level;
-
-	level = btr_page_get_level(buf_block_get_frame(block), mtr);
+	const page_t*	page	= buf_block_get_frame(block);
+	ulint		level	= btr_page_get_level(page, mtr);
 
+	ut_ad(fil_page_get_type(block->frame) == FIL_PAGE_INDEX);
 	btr_page_free_low(index, block, level, mtr);
 }
 
@@ -1329,16 +1362,12 @@ btr_create(
 		/* Allocate then the next page to the segment: it will be the
 		tree root page */
 
-		page_no = fseg_alloc_free_page(buf_block_get_frame(
-						       ibuf_hdr_block)
-					       + IBUF_HEADER
-					       + IBUF_TREE_SEG_HEADER,
-					       IBUF_TREE_ROOT_PAGE_NO,
-					       FSP_UP, mtr);
-		ut_ad(page_no == IBUF_TREE_ROOT_PAGE_NO);
-
-		block = buf_page_get(space, zip_size, page_no,
-				     RW_X_LATCH, mtr);
+		block = fseg_alloc_free_page(
+			buf_block_get_frame(ibuf_hdr_block)
+			+ IBUF_HEADER + IBUF_TREE_SEG_HEADER,
+			IBUF_TREE_ROOT_PAGE_NO,
+			FSP_UP, mtr);
+		ut_ad(buf_block_get_page_no(block) == IBUF_TREE_ROOT_PAGE_NO);
 	} else {
 #ifdef UNIV_BLOB_DEBUG
 		if ((type & DICT_CLUSTERED) && !index->blobs) {
@@ -1815,7 +1844,7 @@ btr_root_raise_and_insert(
 
 	level = btr_page_get_level(root, mtr);
 
-	new_block = btr_page_alloc(index, 0, FSP_NO_DIR, level, mtr);
+	new_block = btr_page_alloc(index, 0, FSP_NO_DIR, level, mtr, mtr);
 	new_page = buf_block_get_frame(new_block);
 	new_page_zip = buf_block_get_page_zip(new_block);
 	ut_a(!new_page_zip == !root_page_zip);
@@ -2551,7 +2580,7 @@ btr_page_split_and_insert(
 
 	/* 2. Allocate a new page to the index */
 	new_block = btr_page_alloc(cursor->index, hint_page_no, direction,
-				   btr_page_get_level(page, mtr), mtr);
+				   btr_page_get_level(page, mtr), mtr, mtr);
 	new_page = buf_block_get_frame(new_block);
 	new_page_zip = buf_block_get_page_zip(new_block);
 	btr_page_create(new_block, new_page_zip, cursor->index,
@@ -3001,15 +3030,16 @@ btr_node_ptr_delete(
 	ut_a(err == DB_SUCCESS);
 
 	if (!compressed) {
-		btr_cur_compress_if_useful(&cursor, mtr);
+		btr_cur_compress_if_useful(&cursor, FALSE, mtr);
 	}
 }
 
 /*************************************************************//**
 If page is the only on its level, this function moves its records to the
-father page, thus reducing the tree height. */
+father page, thus reducing the tree height.
+@return father block */
 static
-void
+buf_block_t*
 btr_lift_page_up(
 /*=============*/
 	dict_index_t*	index,	/*!< in: index tree */
@@ -3126,6 +3156,8 @@ btr_lift_page_up(
 	}
 	ut_ad(page_validate(father_page, index));
 	ut_ad(btr_check_node_ptr(index, father_block, mtr));
+
+	return(father_block);
 }
 
 /*************************************************************//**
@@ -3142,11 +3174,13 @@ UNIV_INTERN
 ibool
 btr_compress(
 /*=========*/
-	btr_cur_t*	cursor,	/*!< in: cursor on the page to merge or lift;
-				the page must not be empty: in record delete
-				use btr_discard_page if the page would become
-				empty */
-	mtr_t*		mtr)	/*!< in: mtr */
+	btr_cur_t*	cursor,	/*!< in/out: cursor on the page to merge
+				or lift; the page must not be empty:
+				when deleting records, use btr_discard_page()
+				if the page would become empty */
+	ibool		adjust,	/*!< in: TRUE if should adjust the
+				cursor position even if compression occurs */
+	mtr_t*		mtr)	/*!< in/out: mini-transaction */
 {
 	dict_index_t*	index;
 	ulint		space;
@@ -3164,12 +3198,14 @@ btr_compress(
 	ulint*		offsets;
 	ulint		data_size;
 	ulint		n_recs;
+	ulint		nth_rec = 0; /* remove bogus warning */
 	ulint		max_ins_size;
 	ulint		max_ins_size_reorg;
 
 	block = btr_cur_get_block(cursor);
 	page = btr_cur_get_page(cursor);
 	index = btr_cur_get_index(cursor);
+
 	ut_a((ibool) !!page_is_comp(page) == dict_table_is_comp(index->table));
 
 	ut_ad(mtr_memo_contains(mtr, dict_index_get_lock(index),
@@ -3190,6 +3226,10 @@ btr_compress(
 	offsets = btr_page_get_father_block(NULL, heap, index, block, mtr,
 					    &father_cursor);
 
+	if (adjust) {
+		nth_rec = page_rec_get_n_recs_before(btr_cur_get_rec(cursor));
+	}
+
 	/* Decide the page to which we try to merge and which will inherit
 	the locks */
 
@@ -3216,9 +3256,9 @@ btr_compress(
 	} else {
 		/* The page is the only one on the level, lift the records
 		to the father */
-		btr_lift_page_up(index, block, mtr);
-		mem_heap_free(heap);
-		return(TRUE);
+
+		merge_block = btr_lift_page_up(index, block, mtr);
+		goto func_exit;
 	}
 
 	n_recs = page_get_n_recs(page);
@@ -3300,6 +3340,10 @@ btr_compress(
 
 		btr_node_ptr_delete(index, block, mtr);
 		lock_update_merge_left(merge_block, orig_pred, block);
+
+		if (adjust) {
+			nth_rec += page_rec_get_n_recs_before(orig_pred);
+		}
 	} else {
 		rec_t*		orig_succ;
 #ifdef UNIV_BTR_DEBUG
@@ -3364,7 +3408,6 @@ btr_compress(
 	}
 
 	btr_blob_dbg_remove(page, index, "btr_compress");
-	mem_heap_free(heap);
 
 	if (!dict_index_is_clust(index) && page_is_leaf(merge_page)) {
 		/* Update the free bits of the B-tree page in the
@@ -3416,6 +3459,16 @@ btr_compress(
 	btr_page_free(index, block, mtr);
 
 	ut_ad(btr_check_node_ptr(index, merge_block, mtr));
+func_exit:
+	mem_heap_free(heap);
+
+	if (adjust) {
+		btr_cur_position(
+			index,
+			page_rec_get_nth(merge_block->frame, nth_rec),
+			merge_block, cursor);
+	}
+
 	return(TRUE);
 }
 

storage/innodb_plugin/ibuf/ibuf0ibuf.c

@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -1722,14 +1722,14 @@ ulint
 ibuf_add_free_page(void)
 /*====================*/
 {
-	mtr_t	mtr;
-	page_t*	header_page;
-	ulint	flags;
-	ulint	zip_size;
-	ulint	page_no;
-	page_t*	page;
-	page_t*	root;
-	page_t*	bitmap_page;
+	mtr_t		mtr;
+	page_t*		header_page;
+	ulint		flags;
+	ulint		zip_size;
+	buf_block_t*	block;
+	page_t*		page;
+	page_t*		root;
+	page_t*		bitmap_page;
 
 	mtr_start(&mtr);
 
@@ -1750,32 +1750,23 @@ ibuf_add_free_page(void)
 	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(
+	block = fseg_alloc_free_page(
 		header_page + IBUF_HEADER + IBUF_TREE_SEG_HEADER, 0, FSP_UP,
 		&mtr);
 
-	if (page_no == FIL_NULL) {
+	if (block == NULL) {
 		mtr_commit(&mtr);
 
 		return(DB_STRONG_FAIL);
 	}
 
-	{
-		buf_block_t*	block;
-
-		block = buf_page_get(
-			IBUF_SPACE_ID, 0, page_no, RW_X_LATCH, &mtr);
-
-		ibuf_enter();
-
-		mutex_enter(&ibuf_mutex);
-
-		root = ibuf_tree_root_get(&mtr);
-
-		buf_block_dbg_add_level(block, SYNC_IBUF_TREE_NODE_NEW);
+	ut_ad(rw_lock_get_x_lock_count(&block->lock) == 1);
+	ibuf_enter();
+	mutex_enter(&ibuf_mutex);
+	root = ibuf_tree_root_get(&mtr);
 
-		page = buf_block_get_frame(block);
-	}
+	buf_block_dbg_add_level(block, SYNC_IBUF_TREE_NODE_NEW);
+	page = buf_block_get_frame(block);
 
 	/* Add the page to the free list and update the ibuf size data */
 
@@ -1792,10 +1783,11 @@ ibuf_add_free_page(void)
 	(level 2 page) */
 
 	bitmap_page = ibuf_bitmap_get_map_page(
-		IBUF_SPACE_ID, page_no, zip_size, &mtr);
+		IBUF_SPACE_ID, buf_block_get_page_no(block), zip_size, &mtr);
 
 	ibuf_bitmap_page_set_bits(
-		bitmap_page, page_no, zip_size, IBUF_BITMAP_IBUF, TRUE, &mtr);
+		bitmap_page, buf_block_get_page_no(block), zip_size,
+		IBUF_BITMAP_IBUF, TRUE, &mtr);
 
 	mtr_commit(&mtr);
 

storage/innodb_plugin/include/btr0btr.h

 /*****************************************************************************
 
-Copyright (c) 1994, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1994, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -488,11 +488,14 @@ UNIV_INTERN
 ibool
 btr_compress(
 /*=========*/
-	btr_cur_t*	cursor,	/*!< in: cursor on the page to merge or lift;
-				the page must not be empty: in record delete
-				use btr_discard_page if the page would become
-				empty */
-	mtr_t*		mtr);	/*!< in: mtr */
+	btr_cur_t*	cursor,	/*!< in/out: cursor on the page to merge
+				or lift; the page must not be empty:
+				when deleting records, use btr_discard_page()
+				if the page would become empty */
+	ibool		adjust,	/*!< in: TRUE if should adjust the
+				cursor position even if compression occurs */
+	mtr_t*		mtr)	/*!< in/out: mini-transaction */
+	__attribute__((nonnull));
 /*************************************************************//**
 Discards a page from a B-tree. This is used to remove the last record from
 a B-tree page: the whole page must be removed at the same time. This cannot
@@ -543,7 +546,10 @@ btr_get_size(
 /**************************************************************//**
 Allocates a new file page to be used in an index tree. NOTE: we assume
 that the caller has made the reservation for free extents!
-@return	new allocated block, x-latched; NULL if out of space */
+@retval NULL if no page could be allocated
+@retval block, rw_lock_x_lock_count(&block->lock) == 1 if allocation succeeded
+(init_mtr == mtr, or the page was not previously freed in mtr)
+@retval block (not allocated or initialized) otherwise */
 UNIV_INTERN
 buf_block_t*
 btr_page_alloc(
@@ -554,7 +560,12 @@ btr_page_alloc(
 					page split is made */
 	ulint		level,		/*!< in: level where the page is placed
 					in the tree */
-	mtr_t*		mtr);		/*!< in: mtr */
+	mtr_t*		mtr,		/*!< in/out: mini-transaction
+					for the allocation */
+	mtr_t*		init_mtr)	/*!< in/out: mini-transaction
+					for x-latching and initializing
+					the page */
+	__attribute__((nonnull, warn_unused_result));
 /**************************************************************//**
 Frees a file page used in an index tree. NOTE: cannot free field external
 storage pages because the page must contain info on its level. */

storage/innodb_plugin/include/btr0cur.h

 /*****************************************************************************
 
-Copyright (c) 1994, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1994, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -36,6 +36,9 @@ Created 10/16/1994 Heikki Tuuri
 #define BTR_NO_LOCKING_FLAG	2	/* do no record lock checking */
 #define BTR_KEEP_SYS_FLAG	4	/* sys fields will be found from the
 					update vector or inserted entry */
+#define BTR_KEEP_POS_FLAG	8	/* btr_cur_pessimistic_update()
+					must keep cursor position when
+					moving columns to big_rec */
 
 #ifndef UNIV_HOTBACKUP
 #include "que0types.h"
@@ -309,7 +312,9 @@ btr_cur_pessimistic_update(
 /*=======================*/
 	ulint		flags,	/*!< in: undo logging, locking, and rollback
 				flags */
-	btr_cur_t*	cursor,	/*!< in: cursor on the record to update */
+	btr_cur_t*	cursor,	/*!< in/out: cursor on the record to update;
+				cursor may become invalid if *big_rec == NULL
+				|| !(flags & BTR_KEEP_POS_FLAG) */
 	mem_heap_t**	heap,	/*!< in/out: pointer to memory heap, or NULL */
 	big_rec_t**	big_rec,/*!< out: big rec vector whose fields have to
 				be stored externally by the caller, or NULL */
@@ -376,10 +381,13 @@ UNIV_INTERN
 ibool
 btr_cur_compress_if_useful(
 /*=======================*/
-	btr_cur_t*	cursor,	/*!< in: cursor on the page to compress;
+	btr_cur_t*	cursor,	/*!< in/out: cursor on the page to compress;
 				cursor does not stay valid if compression
 				occurs */
-	mtr_t*		mtr);	/*!< in: mtr */
+	ibool		adjust,	/*!< in: TRUE if should adjust the
+				cursor position even if compression occurs */
+	mtr_t*		mtr)	/*!< in/out: mini-transaction */
+	__attribute__((nonnull));
 /*******************************************************//**
 Removes the record on which the tree cursor is positioned. It is assumed
 that the mtr has an x-latch on the page where the cursor is positioned,
@@ -504,6 +512,27 @@ btr_cur_disown_inherited_fields(
 	const upd_t*	update,	/*!< in: update vector */
 	mtr_t*		mtr)	/*!< in/out: mini-transaction */
 	__attribute__((nonnull(2,3,4,5,6)));
+
+/** Operation code for btr_store_big_rec_extern_fields(). */
+enum blob_op {
+	/** Store off-page columns for a freshly inserted record */
+	BTR_STORE_INSERT = 0,
+	/** Store off-page columns for an insert by update */
+	BTR_STORE_INSERT_UPDATE,
+	/** Store off-page columns for an update */
+	BTR_STORE_UPDATE
+};
+
+/*******************************************************************//**
+Determine if an operation on off-page columns is an update.
+@return TRUE if op != BTR_STORE_INSERT */
+UNIV_INLINE
+ibool
+btr_blob_op_is_update(
+/*==================*/
+	enum blob_op	op)	/*!< in: operation */
+	__attribute__((warn_unused_result));
+
 /*******************************************************************//**
 Stores the fields in big_rec_vec to the tablespace and puts pointers to
 them in rec.  The extern flags in rec will have to be set beforehand.
@@ -511,52 +540,23 @@ The fields are stored on pages allocated from leaf node
 file segment of the index tree.
 @return	DB_SUCCESS or DB_OUT_OF_FILE_SPACE */
 UNIV_INTERN
-ulint
-btr_store_big_rec_extern_fields_func(
-/*=================================*/
+enum db_err
+btr_store_big_rec_extern_fields(
+/*============================*/
 	dict_index_t*	index,		/*!< in: index of rec; the index tree
 					MUST be X-latched */
 	buf_block_t*	rec_block,	/*!< in/out: block containing rec */
-	rec_t*		rec,		/*!< in: record */
+	rec_t*		rec,		/*!< in/out: record */
 	const ulint*	offsets,	/*!< in: rec_get_offsets(rec, index);
 					the "external storage" flags in offsets
 					will not correspond to rec when
 					this function returns */
-#ifdef UNIV_DEBUG
-	mtr_t*		local_mtr,	/*!< in: mtr containing the
-					latch to rec and to the tree */
-#endif /* UNIV_DEBUG */
-#if defined UNIV_DEBUG || defined UNIV_BLOB_LIGHT_DEBUG
-	ibool		update_in_place,/*! in: TRUE if the record is updated
-					in place (not delete+insert) */
-#endif /* UNIV_DEBUG || UNIV_BLOB_LIGHT_DEBUG */
-	const big_rec_t*big_rec_vec)	/*!< in: vector containing fields
+	const big_rec_t*big_rec_vec,	/*!< in: vector containing fields
 					to be stored externally */
-	__attribute__((nonnull));
-
-/** Stores the fields in big_rec_vec to the tablespace and puts pointers to
-them in rec.  The extern flags in rec will have to be set beforehand.
-The fields are stored on pages allocated from leaf node
-file segment of the index tree.
-@param index	in: clustered index; MUST be X-latched by mtr
-@param b	in/out: block containing rec; MUST be X-latched by mtr
-@param rec	in/out: clustered index record
-@param offsets	in: rec_get_offsets(rec, index);
-		the "external storage" flags in offsets will not be adjusted
-@param mtr	in: mini-transaction that holds x-latch on index and b
-@param upd	in: TRUE if the record is updated in place (not delete+insert)
-@param big	in: vector containing fields to be stored externally
-@return	DB_SUCCESS or DB_OUT_OF_FILE_SPACE */
-#ifdef UNIV_DEBUG
-# define btr_store_big_rec_extern_fields(index,b,rec,offsets,mtr,upd,big) \
-	btr_store_big_rec_extern_fields_func(index,b,rec,offsets,mtr,upd,big)
-#elif defined UNIV_BLOB_LIGHT_DEBUG
-# define btr_store_big_rec_extern_fields(index,b,rec,offsets,mtr,upd,big) \
-	btr_store_big_rec_extern_fields_func(index,b,rec,offsets,upd,big)
-#else
-# define btr_store_big_rec_extern_fields(index,b,rec,offsets,mtr,upd,big) \
-	btr_store_big_rec_extern_fields_func(index,b,rec,offsets,big)
-#endif
+	mtr_t*		btr_mtr,	/*!< in: mtr containing the
+					latches to the clustered index */
+	enum blob_op	op)		/*! in: operation code */
+	__attribute__((nonnull, warn_unused_result));
 
 /*******************************************************************//**
 Frees the space in an externally stored field to the file space

storage/innodb_plugin/include/btr0cur.ic

 /*****************************************************************************
 
-Copyright (c) 1994, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1994, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -139,7 +139,7 @@ btr_cur_compress_recommendation(
 	btr_cur_t*	cursor,	/*!< in: btr cursor */
 	mtr_t*		mtr)	/*!< in: mtr */
 {
-	page_t*		page;
+	const page_t*	page;
 
 	ut_ad(mtr_memo_contains(mtr, btr_cur_get_block(cursor),
 				MTR_MEMO_PAGE_X_FIX));
@@ -197,4 +197,25 @@ btr_cur_can_delete_without_compress(
 
 	return(TRUE);
 }
+
+/*******************************************************************//**
+Determine if an operation on off-page columns is an update.
+@return TRUE if op != BTR_STORE_INSERT */
+UNIV_INLINE
+ibool
+btr_blob_op_is_update(
+/*==================*/
+	enum blob_op	op)	/*!< in: operation */
+{
+	switch (op) {
+	case BTR_STORE_INSERT:
+		return(FALSE);
+	case BTR_STORE_INSERT_UPDATE:
+	case BTR_STORE_UPDATE:
+		return(TRUE);
+	}
+
+	ut_ad(0);
+	return(FALSE);
+}
 #endif /* !UNIV_HOTBACKUP */

storage/innodb_plugin/include/buf0buf.h

 /*****************************************************************************
 
-Copyright (c) 1995, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1995, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -468,6 +468,31 @@ buf_block_get_modify_clock(
 #else /* !UNIV_HOTBACKUP */
 # define buf_block_modify_clock_inc(block) ((void) 0)
 #endif /* !UNIV_HOTBACKUP */
+/*******************************************************************//**
+Increments the bufferfix count. */
+UNIV_INLINE
+void
+buf_block_buf_fix_inc_func(
+/*=======================*/
+#ifdef UNIV_SYNC_DEBUG
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line */
+#endif /* UNIV_SYNC_DEBUG */
+	buf_block_t*	block)	/*!< in/out: block to bufferfix */
+	__attribute__((nonnull));
+#ifdef UNIV_SYNC_DEBUG
+/** Increments the bufferfix count.
+@param b	in/out: block to bufferfix
+@param f	in: file name where requested
+@param l	in: line number where requested */
+# define buf_block_buf_fix_inc(b,f,l) buf_block_buf_fix_inc_func(f,l,b)
+#else /* UNIV_SYNC_DEBUG */
+/** Increments the bufferfix count.
+@param b	in/out: block to bufferfix
+@param f	in: file name where requested
+@param l	in: line number where requested */
+# define buf_block_buf_fix_inc(b,f,l) buf_block_buf_fix_inc_func(b)
+#endif /* UNIV_SYNC_DEBUG */
 /********************************************************************//**
 Calculates a page checksum which is stored to the page when it is written
 to a file. Note that we must be careful to calculate the same value

storage/innodb_plugin/include/buf0buf.ic

 /*****************************************************************************
 
-Copyright (c) 1995, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1995, 2012, Oracle and/or its affiliates. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
@@ -18,8 +18,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -916,19 +916,6 @@ buf_block_buf_fix_inc_func(
 
 	block->page.buf_fix_count++;
 }
-#ifdef UNIV_SYNC_DEBUG
-/** Increments the bufferfix count.
-@param b	in/out: block to bufferfix
-@param f	in: file name where requested
-@param l	in: line number where requested */
-# define buf_block_buf_fix_inc(b,f,l) buf_block_buf_fix_inc_func(f,l,b)
-#else /* UNIV_SYNC_DEBUG */
-/** Increments the bufferfix count.
-@param b	in/out: block to bufferfix
-@param f	in: file name where requested
-@param l	in: line number where requested */
-# define buf_block_buf_fix_inc(b,f,l) buf_block_buf_fix_inc_func(b)
-#endif /* UNIV_SYNC_DEBUG */
 
 /*******************************************************************//**
 Decrements the bufferfix count. */

storage/innodb_plugin/include/fsp0fsp.h

 /*****************************************************************************
 
-Copyright (c) 1995, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1995, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -176,30 +176,33 @@ fseg_n_reserved_pages(
 Allocates a single free page from a segment. This function implements
 the intelligent allocation strategy which tries to minimize
 file space fragmentation.
-@return	the allocated page offset FIL_NULL if no page could be allocated */
-UNIV_INTERN
-ulint
-fseg_alloc_free_page(
-/*=================*/
-	fseg_header_t*	seg_header, /*!< in: segment header */
-	ulint		hint,	/*!< in: hint of which page would be desirable */
-	byte		direction, /*!< in: if the new page is needed because
+@param[in/out] seg_header	segment header
+@param[in] hint			hint of which page would be desirable
+@param[in] direction		if the new page is needed because
 				of an index page split, and records are
 				inserted there in order, into which
 				direction they go alphabetically: FSP_DOWN,
-				FSP_UP, FSP_NO_DIR */
-	mtr_t*		mtr);	/*!< in: mtr handle */
+				FSP_UP, FSP_NO_DIR
+@param[in/out] mtr		mini-transaction
+@return	X-latched block, or NULL if no page could be allocated */
+#define fseg_alloc_free_page(seg_header, hint, direction, mtr)		\
+	fseg_alloc_free_page_general(seg_header, hint, direction,	\
+				     FALSE, mtr, mtr)
 /**********************************************************************//**
 Allocates a single free page from a segment. This function implements
 the intelligent allocation strategy which tries to minimize file space
 fragmentation.
-@return	allocated page offset, FIL_NULL if no page could be allocated */
+@retval NULL if no page could be allocated
+@retval block, rw_lock_x_lock_count(&block->lock) == 1 if allocation succeeded
+(init_mtr == mtr, or the page was not previously freed in mtr)
+@retval block (not allocated or initialized) otherwise */
 UNIV_INTERN
-ulint
+buf_block_t*
 fseg_alloc_free_page_general(
 /*=========================*/
 	fseg_header_t*	seg_header,/*!< in/out: segment header */
-	ulint		hint,	/*!< in: hint of which page would be desirable */
+	ulint		hint,	/*!< in: hint of which page would be
+				desirable */
 	byte		direction,/*!< in: if the new page is needed because
 				of an index page split, and records are
 				inserted there in order, into which
@@ -210,8 +213,12 @@ fseg_alloc_free_page_general(
 				with fsp_reserve_free_extents, then there
 				is no need to do the check for this individual
 				page */
-	mtr_t*		mtr)	/*!< in/out: mini-transaction */
-	__attribute__((warn_unused_result, nonnull(1,5)));
+	mtr_t*		mtr,	/*!< in/out: mini-transaction */
+	mtr_t*		init_mtr)/*!< in/out: mtr or another mini-transaction
+				in which the page should be initialized.
+				If init_mtr!=mtr, but the page is already
+				latched in mtr, do not initialize the page. */
+	__attribute__((warn_unused_result, nonnull));
 /**********************************************************************//**
 Reserves free pages from a tablespace. All mini-transactions which may
 use several pages from the tablespace should call this function beforehand

storage/innodb_plugin/include/mtr0mtr.h

 /*****************************************************************************
 
-Copyright (c) 1995, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1995, 2012, 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
@@ -50,7 +50,9 @@ first 3 values must be RW_S_LATCH, RW_X_LATCH, RW_NO_LATCH */
 #define	MTR_MEMO_PAGE_S_FIX	RW_S_LATCH
 #define	MTR_MEMO_PAGE_X_FIX	RW_X_LATCH
 #define	MTR_MEMO_BUF_FIX	RW_NO_LATCH
-#define MTR_MEMO_MODIFY		54
+#ifdef UNIV_DEBUG
+# define MTR_MEMO_MODIFY	54
+#endif /* UNIV_DEBUG */
 #define	MTR_MEMO_S_LOCK		55
 #define	MTR_MEMO_X_LOCK		56
 
@@ -378,11 +380,14 @@ struct mtr_struct{
 	dyn_array_t	memo;	/*!< memo stack for locks etc. */
 	dyn_array_t	log;	/*!< mini-transaction log */
 	ibool		modifications;
-				/* TRUE if the mtr made modifications to
-				buffer pool pages */
+				/*!< TRUE if the mini-transaction
+				modified buffer pool pages */
 	ulint		n_log_recs;
 				/* count of how many page initial log records
 				have been written to the mtr log */
+	ulint		n_freed_pages;
+				/* number of pages that have been freed in
+				this mini-transaction */
 	ulint		log_mode; /* specifies which operations should be
 				logged; default value MTR_LOG_ALL */
 	ib_uint64_t	start_lsn;/* start lsn of the possible log entry for

storage/innodb_plugin/include/mtr0mtr.ic

 /*****************************************************************************
 
-Copyright (c) 1995, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1995, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -45,6 +45,7 @@ mtr_start(
 	mtr->log_mode = MTR_LOG_ALL;
 	mtr->modifications = FALSE;
 	mtr->n_log_recs = 0;
+	mtr->n_freed_pages = 0;
 
 	ut_d(mtr->state = MTR_ACTIVE);
 	ut_d(mtr->magic_n = MTR_MAGIC_N);

storage/innodb_plugin/include/page0page.h

 /*****************************************************************************
 
-Copyright (c) 1994, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1994, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -281,16 +281,42 @@ page_get_supremum_offset(
 	const page_t*	page);	/*!< in: page which must have record(s) */
 #define page_get_infimum_rec(page) ((page) + page_get_infimum_offset(page))
 #define page_get_supremum_rec(page) ((page) + page_get_supremum_offset(page))
+
 /************************************************************//**
-Returns the middle record of record list. If there are an even number
-of records in the list, returns the first record of upper half-list.
-@return	middle record */
+Returns the nth record of the record list.
+This is the inverse function of page_rec_get_n_recs_before().
+@return	nth record */
 UNIV_INTERN
+const rec_t*
+page_rec_get_nth_const(
+/*===================*/
+	const page_t*	page,	/*!< in: page */
+	ulint		nth)	/*!< in: nth record */
+	__attribute__((nonnull, warn_unused_result));
+/************************************************************//**
+Returns the nth record of the record list.
+This is the inverse function of page_rec_get_n_recs_before().
+@return	nth record */
+UNIV_INLINE
+rec_t*
+page_rec_get_nth(
+/*=============*/
+	page_t*	page,	/*< in: page */
+	ulint	nth)	/*!< in: nth record */
+	__attribute__((nonnull, warn_unused_result));
+
+#ifndef UNIV_HOTBACKUP
+/************************************************************//**
+Returns the middle record of the records on the page. If there is an
+even number of records in the list, returns the first record of the
+upper half-list.
+@return	middle record */
+UNIV_INLINE
 rec_t*
 page_get_middle_rec(
 /*================*/
-	page_t*	page);	/*!< in: page */
-#ifndef UNIV_HOTBACKUP
+	page_t*	page)	/*!< in: page */
+	__attribute__((nonnull, warn_unused_result));
 /*************************************************************//**
 Compares a data tuple to a physical record. Differs from the function
 cmp_dtuple_rec_with_match in the way that the record must reside on an
@@ -345,6 +371,7 @@ page_get_n_recs(
 /***************************************************************//**
 Returns the number of records before the given record in chain.
 The number includes infimum and supremum records.
+This is the inverse function of page_rec_get_nth().
 @return	number of records */
 UNIV_INTERN
 ulint

storage/innodb_plugin/include/page0page.ic

 /*****************************************************************************
 
-Copyright (c) 1994, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1994, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -420,7 +420,37 @@ page_rec_is_infimum(
 	return(page_rec_is_infimum_low(page_offset(rec)));
 }
 
+/************************************************************//**
+Returns the nth record of the record list.
+This is the inverse function of page_rec_get_n_recs_before().
+@return	nth record */
+UNIV_INLINE
+rec_t*
+page_rec_get_nth(
+/*=============*/
+	page_t*	page,	/*!< in: page */
+	ulint	nth)	/*!< in: nth record */
+{
+	return((rec_t*) page_rec_get_nth_const(page, nth));
+}
+
 #ifndef UNIV_HOTBACKUP
+/************************************************************//**
+Returns the middle record of the records on the page. If there is an
+even number of records in the list, returns the first record of the
+upper half-list.
+@return	middle record */
+UNIV_INLINE
+rec_t*
+page_get_middle_rec(
+/*================*/
+	page_t*	page)	/*!< in: page */
+{
+	ulint	middle = (page_get_n_recs(page) + PAGE_HEAP_NO_USER_LOW) / 2;
+
+	return(page_rec_get_nth(page, middle));
+}
+
 /*************************************************************//**
 Compares a data tuple to a physical record. Differs from the function
 cmp_dtuple_rec_with_match in the way that the record must reside on an

storage/innodb_plugin/include/trx0rec.ic

 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -107,6 +107,7 @@ trx_undo_rec_copy(
 
 	len = mach_read_from_2(undo_rec)
 		- ut_align_offset(undo_rec, UNIV_PAGE_SIZE);
+	ut_ad(len < UNIV_PAGE_SIZE);
 	return(mem_heap_dup(heap, undo_rec, len));
 }
 #endif /* !UNIV_HOTBACKUP */

storage/innodb_plugin/include/trx0undo.h

 /*****************************************************************************
 
-Copyright (c) 1996, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1996, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -194,16 +194,17 @@ trx_undo_get_first_rec(
 	mtr_t*	mtr);	/*!< in: mtr */
 /********************************************************************//**
 Tries to add a page to the undo log segment where the undo log is placed.
-@return	page number if success, else FIL_NULL */
+@return	X-latched block if success, else NULL */
 UNIV_INTERN
-ulint
+buf_block_t*
 trx_undo_add_page(
 /*==============*/
 	trx_t*		trx,	/*!< in: transaction */
 	trx_undo_t*	undo,	/*!< in: undo log memory object */
-	mtr_t*		mtr);	/*!< in: mtr which does not have a latch to any
+	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 */
+	__attribute__((nonnull, warn_unused_result));
 /********************************************************************//**
 Frees the last undo log page.
 The caller must hold the rollback segment mutex. */

storage/innodb_plugin/page/page0cur.c

 /*****************************************************************************
 
-Copyright (c) 1994, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1994, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -1180,14 +1180,15 @@ page_cur_insert_rec_zip_reorg(
 	/* Before trying to reorganize the page,
 	store the number of preceding records on the page. */
 	pos = page_rec_get_n_recs_before(rec);
+	ut_ad(pos > 0);
 
 	if (page_zip_reorganize(block, index, mtr)) {
 		/* The page was reorganized: Find rec by seeking to pos,
 		and update *current_rec. */
-		rec = page + PAGE_NEW_INFIMUM;
-
-		while (--pos) {
-			rec = page + rec_get_next_offs(rec, TRUE);
+		if (pos > 1) {
+			rec = page_rec_get_nth(page, pos - 1);
+		} else {
+			rec = page + PAGE_NEW_INFIMUM;
 		}
 
 		*current_rec = rec;
@@ -1283,6 +1284,12 @@ page_cur_insert_rec_zip(
 			insert_rec = page_cur_insert_rec_zip_reorg(
 				current_rec, block, index, insert_rec,
 				page, page_zip, mtr);
+#ifdef UNIV_DEBUG
+			if (insert_rec) {
+				rec_offs_make_valid(
+					insert_rec, index, offsets);
+			}
+#endif /* UNIV_DEBUG */
 		}
 
 		return(insert_rec);

storage/innodb_plugin/page/page0page.c

 /*****************************************************************************
 
-Copyright (c) 1994, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1994, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -1475,55 +1475,54 @@ page_dir_balance_slot(
 	}
 }
 
-#ifndef UNIV_HOTBACKUP
 /************************************************************//**
-Returns the middle record of the record list. If there are an even number
-of records in the list, returns the first record of the upper half-list.
-@return	middle record */
+Returns the nth record of the record list.
+This is the inverse function of page_rec_get_n_recs_before().
+@return	nth record */
 UNIV_INTERN
-rec_t*
-page_get_middle_rec(
-/*================*/
-	page_t*	page)	/*!< in: page */
+const rec_t*
+page_rec_get_nth_const(
+/*===================*/
+	const page_t*	page,	/*!< in: page */
+	ulint		nth)	/*!< in: nth record */
 {
-	page_dir_slot_t*	slot;
-	ulint			middle;
+	const page_dir_slot_t*	slot;
 	ulint			i;
 	ulint			n_owned;
-	ulint			count;
-	rec_t*			rec;
+	const rec_t*		rec;
 
-	/* This many records we must leave behind */
-	middle = (page_get_n_recs(page) + PAGE_HEAP_NO_USER_LOW) / 2;
-
-	count = 0;
+	ut_ad(nth < UNIV_PAGE_SIZE / (REC_N_NEW_EXTRA_BYTES + 1));
 
 	for (i = 0;; i++) {
 
 		slot = page_dir_get_nth_slot(page, i);
 		n_owned = page_dir_slot_get_n_owned(slot);
 
-		if (count + n_owned > middle) {
+		if (n_owned > nth) {
 			break;
 		} else {
-			count += n_owned;
+			nth -= n_owned;
 		}
 	}
 
 	ut_ad(i > 0);
 	slot = page_dir_get_nth_slot(page, i - 1);
-	rec = (rec_t*) page_dir_slot_get_rec(slot);
-	rec = page_rec_get_next(rec);
-
-	/* There are now count records behind rec */
+	rec = page_dir_slot_get_rec(slot);
 
-	for (i = 0; i < middle - count; i++) {
-		rec = page_rec_get_next(rec);
+	if (page_is_comp(page)) {
+		do {
+			rec = page_rec_get_next_low(rec, TRUE);
+			ut_ad(rec);
+		} while (nth--);
+	} else {
+		do {
+			rec = page_rec_get_next_low(rec, FALSE);
+			ut_ad(rec);
+		} while (nth--);
 	}
 
 	return(rec);
 }
-#endif /* !UNIV_HOTBACKUP */
 
 /***************************************************************//**
 Returns the number of records before the given record in chain.
@@ -1585,6 +1584,7 @@ page_rec_get_n_recs_before(
 	n--;
 
 	ut_ad(n >= 0);
+	ut_ad(n < UNIV_PAGE_SIZE / (REC_N_NEW_EXTRA_BYTES + 1));
 
 	return((ulint) n);
 }

storage/innodb_plugin/row/row0ins.c

 /*****************************************************************************
 
-Copyright (c) 1996, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1996, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -354,9 +354,9 @@ row_ins_clust_index_entry_by_modify(
 			return(DB_LOCK_TABLE_FULL);
 
 		}
-		err = btr_cur_pessimistic_update(0, cursor,
-						 heap, big_rec, update,
-						 0, thr, mtr);
+		err = btr_cur_pessimistic_update(
+			BTR_KEEP_POS_FLAG, cursor, heap, big_rec, update,
+			0, thr, mtr);
 	}
 
 	return(err);
@@ -1998,6 +1998,7 @@ row_ins_index_entry_low(
 	ulint		modify = 0; /* remove warning */
 	rec_t*		insert_rec;
 	rec_t*		rec;
+	ulint*		offsets;
 	ulint		err;
 	ulint		n_unique;
 	big_rec_t*	big_rec			= NULL;
@@ -2101,6 +2102,64 @@ row_ins_index_entry_low(
 			err = row_ins_clust_index_entry_by_modify(
 				mode, &cursor, &heap, &big_rec, entry,
 				thr, &mtr);
+
+			if (big_rec) {
+				ut_a(err == DB_SUCCESS);
+				/* Write out the externally stored
+				columns while still x-latching
+				index->lock and block->lock. Allocate
+				pages for big_rec in the mtr that
+				modified the B-tree, but be sure to skip
+				any pages that were freed in mtr. We will
+				write out the big_rec pages before
+				committing the B-tree mini-transaction. If
+				the system crashes so that crash recovery
+				will not replay the mtr_commit(&mtr), the
+				big_rec pages will be left orphaned until
+				the pages are allocated for something else.
+
+				TODO: If the allocation extends the
+				tablespace, it will not be redo
+				logged, in either mini-transaction.
+				Tablespace extension should be
+				redo-logged in the big_rec
+				mini-transaction, so that recovery
+				will not fail when the big_rec was
+				written to the extended portion of the
+				file, in case the file was somehow
+				truncated in the crash. */
+
+				rec = btr_cur_get_rec(&cursor);
+				offsets = rec_get_offsets(
+					rec, index, NULL,
+					ULINT_UNDEFINED, &heap);
+
+				DEBUG_SYNC_C("before_row_ins_upd_extern");
+				err = btr_store_big_rec_extern_fields(
+					index, btr_cur_get_block(&cursor),
+					rec, offsets, big_rec, &mtr,
+					BTR_STORE_INSERT_UPDATE);
+				DEBUG_SYNC_C("after_row_ins_upd_extern");
+				/* If writing big_rec fails (for
+				example, because of DB_OUT_OF_FILE_SPACE),
+				the record will be corrupted. Even if
+				we did not update any externally
+				stored columns, our update could cause
+				the record to grow so that a
+				non-updated column was selected for
+				external storage. This non-update
+				would not have been written to the
+				undo log, and thus the record cannot
+				be rolled back.
+
+				However, because we have not executed
+				mtr_commit(mtr) yet, the update will
+				not be replayed in crash recovery, and
+				the following assertion failure will
+				effectively "roll back" the operation. */
+				ut_a(err == DB_SUCCESS);
+				goto stored_big_rec;
+			}
 		} else {
 			ut_ad(!n_ext);
 			err = row_ins_sec_index_entry_by_modify(
@@ -2129,9 +2188,6 @@ row_ins_index_entry_low(
 	mtr_commit(&mtr);
 
 	if (UNIV_LIKELY_NULL(big_rec)) {
-		rec_t*	rec;
-		ulint*	offsets;
-
 		DBUG_EXECUTE_IF(
 			"row_ins_extern_checkpoint",
 			log_make_checkpoint_at(IB_ULONGLONG_MAX, TRUE););
@@ -2146,12 +2202,13 @@ row_ins_index_entry_low(
 		offsets = rec_get_offsets(rec, index, NULL,
 					  ULINT_UNDEFINED, &heap);
 
-		DEBUG_SYNC_C("before_row_ins_upd_extern");
+		DEBUG_SYNC_C("before_row_ins_extern");
 		err = btr_store_big_rec_extern_fields(
 			index, btr_cur_get_block(&cursor),
-			rec, offsets, &mtr, FALSE, big_rec);
-		DEBUG_SYNC_C("after_row_ins_upd_extern");
+			rec, offsets, big_rec, &mtr, BTR_STORE_INSERT);
+		DEBUG_SYNC_C("after_row_ins_extern");
 
+stored_big_rec:
 		if (modify) {
 			dtuple_big_rec_free(big_rec);
 		} else {

storage/innodb_plugin/row/row0row.c

 /*****************************************************************************
 
-Copyright (c) 1996, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1996, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -243,19 +243,20 @@ row_build(
 	}
 
 #if defined UNIV_DEBUG || defined UNIV_BLOB_LIGHT_DEBUG
-	/* This condition can occur during crash recovery before
-	trx_rollback_active() has completed execution.
-
-	This condition is possible if the server crashed
-	during an insert or update before
-	btr_store_big_rec_extern_fields() did mtr_commit() all
-	BLOB pointers to the clustered index record.
-
-	If the record contains a null BLOB pointer, look up the
-	transaction that holds the implicit lock on this record, and
-	assert that it was recovered (and will soon be rolled back). */
-	ut_a(!rec_offs_any_null_extern(rec, offsets)
-	     || trx_assert_recovered(row_get_rec_trx_id(rec, index, offsets)));
+	if (rec_offs_any_null_extern(rec, offsets)) {
+		/* This condition can occur during crash recovery
+		before trx_rollback_active() has completed execution.
+
+		This condition is possible if the server crashed
+		during an insert or update-by-delete-and-insert before
+		btr_store_big_rec_extern_fields() did mtr_commit() all
+		BLOB pointers to the freshly inserted clustered index
+		record. */
+		ut_a(trx_assert_recovered(
+			     row_get_rec_trx_id(rec, index, offsets)));
+		ut_a(trx_undo_roll_ptr_is_insert(
+			     row_get_rec_roll_ptr(rec, index, offsets)));
+	}
 #endif /* UNIV_DEBUG || UNIV_BLOB_LIGHT_DEBUG */
 
 	if (type != ROW_COPY_POINTERS) {

storage/innodb_plugin/row/row0upd.c

 /*****************************************************************************
 
-Copyright (c) 1996, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1996, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -1978,33 +1978,62 @@ row_upd_clust_rec(
 	ut_ad(!rec_get_deleted_flag(btr_pcur_get_rec(pcur),
 				    dict_table_is_comp(index->table)));
 
-	err = btr_cur_pessimistic_update(BTR_NO_LOCKING_FLAG, btr_cur,
-					 &heap, &big_rec, node->update,
-					 node->cmpl_info, thr, mtr);
-	mtr_commit(mtr);
-
-	if (err == DB_SUCCESS && big_rec) {
-		ulint		offsets_[REC_OFFS_NORMAL_SIZE];
-		rec_t*		rec;
+	err = btr_cur_pessimistic_update(
+		BTR_NO_LOCKING_FLAG | BTR_KEEP_POS_FLAG, btr_cur,
+		&heap, &big_rec, node->update, node->cmpl_info, thr, mtr);
+	if (big_rec) {
+		ulint	offsets_[REC_OFFS_NORMAL_SIZE];
+		rec_t*	rec;
 		rec_offs_init(offsets_);
 
-		DBUG_EXECUTE_IF(
-			"row_upd_extern_checkpoint",
-			log_make_checkpoint_at(IB_ULONGLONG_MAX, TRUE););
+		ut_a(err == DB_SUCCESS);
+		/* Write out the externally stored
+		columns while still x-latching
+		index->lock and block->lock. Allocate
+		pages for big_rec in the mtr that
+		modified the B-tree, but be sure to skip
+		any pages that were freed in mtr. We will
+		write out the big_rec pages before
+		committing the B-tree mini-transaction. If
+		the system crashes so that crash recovery
+		will not replay the mtr_commit(&mtr), the
+		big_rec pages will be left orphaned until
+		the pages are allocated for something else.
+
+		TODO: If the allocation extends the tablespace, it
+		will not be redo logged, in either mini-transaction.
+		Tablespace extension should be redo-logged in the
+		big_rec mini-transaction, so that recovery will not
+		fail when the big_rec was written to the extended
+		portion of the file, in case the file was somehow
+		truncated in the crash. */
 
-		mtr_start(mtr);
-		ut_a(btr_pcur_restore_position(BTR_MODIFY_TREE, pcur, mtr));
 		rec = btr_cur_get_rec(btr_cur);
 		DEBUG_SYNC_C("before_row_upd_extern");
 		err = btr_store_big_rec_extern_fields(
 			index, btr_cur_get_block(btr_cur), rec,
 			rec_get_offsets(rec, index, offsets_,
 					ULINT_UNDEFINED, &heap),
-			mtr, TRUE, big_rec);
+			big_rec, mtr, BTR_STORE_UPDATE);
 		DEBUG_SYNC_C("after_row_upd_extern");
-		mtr_commit(mtr);
+		/* If writing big_rec fails (for example, because of
+		DB_OUT_OF_FILE_SPACE), the record will be corrupted.
+		Even if we did not update any externally stored
+		columns, our update could cause the record to grow so
+		that a non-updated column was selected for external
+		storage. This non-update would not have been written
+		to the undo log, and thus the record cannot be rolled
+		back.
+
+		However, because we have not executed mtr_commit(mtr)
+		yet, the update will not be replayed in crash
+		recovery, and the following assertion failure will
+		effectively "roll back" the operation. */
+		ut_a(err == DB_SUCCESS);
 	}
 
+	mtr_commit(mtr);
+
 	if (UNIV_LIKELY_NULL(heap)) {
 		mem_heap_free(heap);
 	}

storage/innodb_plugin/trx/trx0rec.c

 /*****************************************************************************
 
-Copyright (c) 1996, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1996, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -1176,6 +1176,7 @@ trx_undo_report_row_operation(
 	trx_t*		trx;
 	trx_undo_t*	undo;
 	ulint		page_no;
+	buf_block_t*	undo_block;
 	trx_rseg_t*	rseg;
 	mtr_t		mtr;
 	ulint		err		= DB_SUCCESS;
@@ -1218,10 +1219,13 @@ trx_undo_report_row_operation(
 
 		if (UNIV_UNLIKELY(!undo)) {
 			/* Did not succeed */
+			ut_ad(err != DB_SUCCESS);
 			mutex_exit(&(trx->undo_mutex));
 
 			return(err);
 		}
+
+		ut_ad(err == DB_SUCCESS);
 	} else {
 		ut_ad(op_type == TRX_UNDO_MODIFY_OP);
 
@@ -1235,30 +1239,30 @@ trx_undo_report_row_operation(
 
 		if (UNIV_UNLIKELY(!undo)) {
 			/* Did not succeed */
+			ut_ad(err != DB_SUCCESS);
 			mutex_exit(&(trx->undo_mutex));
 			return(err);
 		}
 
+		ut_ad(err == DB_SUCCESS);
 		offsets = rec_get_offsets(rec, index, offsets,
 					  ULINT_UNDEFINED, &heap);
 	}
 
-	page_no = undo->last_page_no;
-
 	mtr_start(&mtr);
 
+	page_no = undo->last_page_no;
+	undo_block = buf_page_get_gen(
+		undo->space, undo->zip_size, page_no, RW_X_LATCH,
+		undo->guess_block, BUF_GET, __FILE__, __LINE__, &mtr);
+	buf_block_dbg_add_level(undo_block, SYNC_TRX_UNDO_PAGE);
+
 	do {
-		buf_block_t*	undo_block;
 		page_t*		undo_page;
 		ulint		offset;
 
-		undo_block = buf_page_get_gen(undo->space, undo->zip_size,
-					      page_no, RW_X_LATCH,
-					      undo->guess_block, BUF_GET,
-					      __FILE__, __LINE__, &mtr);
-		buf_block_dbg_add_level(undo_block, SYNC_TRX_UNDO_PAGE);
-
 		undo_page = buf_block_get_frame(undo_block);
+		ut_ad(page_no == buf_block_get_page_no(undo_block));
 
 		if (op_type == TRX_UNDO_INSERT_OP) {
 			offset = trx_undo_page_report_insert(
@@ -1335,12 +1339,12 @@ trx_undo_report_row_operation(
 		a pessimistic insert in a B-tree, and we must reserve the
 		counterpart of the tree latch, which is the rseg mutex. */
 
-		mutex_enter(&(rseg->mutex));
-
-		page_no = trx_undo_add_page(trx, undo, &mtr);
+		mutex_enter(&rseg->mutex);
+		undo_block = trx_undo_add_page(trx, undo, &mtr);
+		mutex_exit(&rseg->mutex);
 
-		mutex_exit(&(rseg->mutex));
-	} while (UNIV_LIKELY(page_no != FIL_NULL));
+		page_no = undo->last_page_no;
+	} while (undo_block != NULL);
 
 	/* Did not succeed: out of space */
 	err = DB_OUT_OF_FILE_SPACE;

storage/innodb_plugin/trx/trx0sys.c

 /*****************************************************************************
 
-Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -247,9 +247,7 @@ trx_sys_create_doublewrite_buf(void)
 {
 	buf_block_t*	block;
 	buf_block_t*	block2;
-#ifdef UNIV_SYNC_DEBUG
 	buf_block_t*	new_block;
-#endif /* UNIV_SYNC_DEBUG */
 	byte*	doublewrite;
 	byte*	fseg_header;
 	ulint	page_no;
@@ -328,10 +326,9 @@ trx_sys_create_doublewrite_buf(void)
 
 		for (i = 0; i < 2 * TRX_SYS_DOUBLEWRITE_BLOCK_SIZE
 			     + FSP_EXTENT_SIZE / 2; i++) {
-			page_no = fseg_alloc_free_page(fseg_header,
-						       prev_page_no + 1,
-						       FSP_UP, &mtr);
-			if (page_no == FIL_NULL) {
+			new_block = fseg_alloc_free_page(
+				fseg_header, prev_page_no + 1, FSP_UP, &mtr);
+			if (new_block == NULL) {
 				fprintf(stderr,
 					"InnoDB: Cannot create doublewrite"
 					" buffer: you must\n"
@@ -352,13 +349,8 @@ trx_sys_create_doublewrite_buf(void)
 			the page position in the tablespace, then the page
 			has not been written to in doublewrite. */
 
-#ifdef UNIV_SYNC_DEBUG
-			new_block =
-#endif /* UNIV_SYNC_DEBUG */
-			buf_page_get(TRX_SYS_SPACE, 0, page_no,
-				     RW_X_LATCH, &mtr);
-			buf_block_dbg_add_level(new_block,
-						SYNC_NO_ORDER_CHECK);
+			ut_ad(rw_lock_get_x_lock_count(&new_block->lock) == 1);
+			page_no = buf_block_get_page_no(new_block);
 
 			if (i == FSP_EXTENT_SIZE / 2) {
 				ut_a(page_no == FSP_EXTENT_SIZE);

storage/innodb_plugin/trx/trx0undo.c

 /*****************************************************************************
 
-Copyright (c) 1996, 2011, Oracle and/or its affiliates. All Rights Reserved.
+Copyright (c) 1996, 2012, 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
@@ -11,8 +11,8 @@ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
 FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
 
 You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
+this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
 
 *****************************************************************************/
 
@@ -870,9 +870,9 @@ trx_undo_discard_latest_update_undo(
 #ifndef UNIV_HOTBACKUP
 /********************************************************************//**
 Tries to add a page to the undo log segment where the undo log is placed.
-@return	page number if success, else FIL_NULL */
+@return	X-latched block if success, else NULL */
 UNIV_INTERN
-ulint
+buf_block_t*
 trx_undo_add_page(
 /*==============*/
 	trx_t*		trx,	/*!< in: transaction */
@@ -882,11 +882,10 @@ trx_undo_add_page(
 				the rollback segment mutex */
 {
 	page_t*		header_page;
+	buf_block_t*	new_block;
 	page_t*		new_page;
 	trx_rseg_t*	rseg;
-	ulint		page_no;
 	ulint		n_reserved;
-	ibool		success;
 
 	ut_ad(mutex_own(&(trx->undo_mutex)));
 	ut_ad(!mutex_own(&kernel_mutex));
@@ -896,37 +895,37 @@ trx_undo_add_page(
 
 	if (rseg->curr_size == rseg->max_size) {
 
-		return(FIL_NULL);
+		return(NULL);
 	}
 
 	header_page = trx_undo_page_get(undo->space, undo->zip_size,
 					undo->hdr_page_no, mtr);
 
-	success = fsp_reserve_free_extents(&n_reserved, undo->space, 1,
-					   FSP_UNDO, mtr);
-	if (!success) {
+	if (!fsp_reserve_free_extents(&n_reserved, undo->space, 1,
+				      FSP_UNDO, mtr)) {
 
-		return(FIL_NULL);
+		return(NULL);
 	}
 
-	page_no = fseg_alloc_free_page_general(header_page + TRX_UNDO_SEG_HDR
-					       + TRX_UNDO_FSEG_HEADER,
-					       undo->top_page_no + 1, FSP_UP,
-					       TRUE, mtr);
+	new_block = fseg_alloc_free_page_general(
+		TRX_UNDO_SEG_HDR + TRX_UNDO_FSEG_HEADER
+		+ header_page,
+		undo->top_page_no + 1, FSP_UP, TRUE, mtr, mtr);
 
 	fil_space_release_free_extents(undo->space, n_reserved);
 
-	if (page_no == FIL_NULL) {
+	if (new_block == NULL) {
 
 		/* No space left */
 
-		return(FIL_NULL);
+		return(NULL);
 	}
 
-	undo->last_page_no = page_no;
+	ut_ad(rw_lock_get_x_lock_count(&new_block->lock) == 1);
+	buf_block_dbg_add_level(new_block, SYNC_TRX_UNDO_PAGE);
+	undo->last_page_no = buf_block_get_page_no(new_block);
 
-	new_page = trx_undo_page_get(undo->space, undo->zip_size,
-				     page_no, mtr);
+	new_page = buf_block_get_frame(new_block);
 
 	trx_undo_page_init(new_page, undo->type, mtr);
 
@@ -935,7 +934,7 @@ trx_undo_add_page(
 	undo->size++;
 	rseg->curr_size++;
 
-	return(page_no);
+	return(new_block);
 }
 
 /********************************************************************//**