From 7946d6ca45f2489507066e3fe6f0ec73379ec54a Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 14:44:32 +0400
Subject: [PATCH 001/116] Applying InnoDB snapshot

Detailed revision comments:

r6248 | marko | 2009-11-30 12:19:50 +0200 (Mon, 30 Nov 2009) | 1 line
branches/zip: ChangeLog: Document r4922 that was forgotten.
---
 storage/innodb_plugin/ChangeLog | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 1a6e07fd147..5ea2995a974 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -535,6 +535,12 @@
 	Fix Bug#44320 InnoDB: missing DB_ROLL_PTR in Table Monitor COLUMNS
 	output
 
+2009-04-29	The InnoDB Team
+
+	* fil/fil0fil.c, include/fil0fil.h, include/mtr0mtr.h,
+	log/log0recv.c:
+	Fix Bug#41609 Crash recovery does not work for InnoDB temporary tables
+
 2009-04-23	The InnoDB Team
 
 	* row/row0mysql.c:

From 91111174ee920d797494403cfc88df34fddda4b5 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 14:45:58 +0400
Subject: [PATCH 002/116] Applying InnoDB snapshot, fixes BUG#41609.

Detailed revision comments:

r6252 | marko | 2009-11-30 12:50:11 +0200 (Mon, 30 Nov 2009) | 23 lines
branches/zip: Suppress errors about non-found temporary tables.
Write the is_temp flag to SYS_TABLES.MIX_LEN.

dict_table_t::flags: Add a flag for is_temporary, DICT_TF2_TEMPORARY.
Unlike other flags, this will not be written to the tablespace flags
or SYS_TABLES.TYPE, but only to SYS_TABLES.MIX_LEN.

dict_build_table_def_step(): Only pass DICT_TF_BITS to tablespaces.

dict_check_tablespaces_and_store_max_id(), dict_load_table():
Suppress errors about temporary tables not being found.

dict_create_sys_tables_tuple(): Write the DICT_TF2_TEMPORARY flag
to SYS_TABLES.MIX_LEN.

fil_space_create(), fil_create_new_single_table_tablespace(): Add assertions
about space->flags.

row_drop_table_for_mysql(): Do not complain about non-found temporary tables.

rb://160 approved by Heikki Tuuri.  This addresses the second part of
Bug #41609 Crash recovery does not work for InnoDB temporary tables.
---
 storage/innodb_plugin/ChangeLog            |  10 ++
 storage/innodb_plugin/dict/dict0boot.c     |   3 +
 storage/innodb_plugin/dict/dict0crea.c     |  17 ++--
 storage/innodb_plugin/dict/dict0load.c     | 113 +++++++++++++++------
 storage/innodb_plugin/dict/dict0mem.c      |   2 +-
 storage/innodb_plugin/fil/fil0fil.c        |   9 +-
 storage/innodb_plugin/handler/ha_innodb.cc |   4 +
 storage/innodb_plugin/include/dict0mem.h   |  24 ++++-
 storage/innodb_plugin/row/row0mysql.c      |  11 +-
 9 files changed, 143 insertions(+), 50 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 5ea2995a974..ac66ff10963 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,13 @@
+2009-11-30	The InnoDB Team
+
+	* dict/dict0crea.c, dict/dict0mem.c, dict/dict0load.c,
+	dict/dict0boot.c, fil/fil0fil.c, handler/ha_innodb.cc,
+	include/dict0mem.h, row/row0mysql.c:
+	Fix the bogus warning messages for non-existing temporary
+	tables that were reported in
+	Bug#41609 Crash recovery does not work for InnoDB temporary tables.
+	The actual crash recovery bug was corrected on 2009-04-29.
+
 2009-11-20	The InnoDB Team
 
 	* handler/ha_innodb.cc:
diff --git a/storage/innodb_plugin/dict/dict0boot.c b/storage/innodb_plugin/dict/dict0boot.c
index e55de30481b..8f948c06c51 100644
--- a/storage/innodb_plugin/dict/dict0boot.c
+++ b/storage/innodb_plugin/dict/dict0boot.c
@@ -274,6 +274,9 @@ dict_boot(void)
 	and (TYPE & DICT_TF_FORMAT_MASK) are nonzero and TYPE = table->flags */
 	dict_mem_table_add_col(table, heap, "TYPE", DATA_INT, 0, 4);
 	dict_mem_table_add_col(table, heap, "MIX_ID", DATA_BINARY, 0, 0);
+	/* MIX_LEN may contain additional table flags when
+	ROW_FORMAT!=REDUNDANT.  Currently, these flags include
+	DICT_TF2_TEMPORARY. */
 	dict_mem_table_add_col(table, heap, "MIX_LEN", DATA_INT, 0, 4);
 	dict_mem_table_add_col(table, heap, "CLUSTER_NAME", DATA_BINARY, 0, 0);
 	dict_mem_table_add_col(table, heap, "SPACE", DATA_INT, 0, 4);
diff --git a/storage/innodb_plugin/dict/dict0crea.c b/storage/innodb_plugin/dict/dict0crea.c
index 96a9bd8152e..b0341e5eeab 100644
--- a/storage/innodb_plugin/dict/dict0crea.c
+++ b/storage/innodb_plugin/dict/dict0crea.c
@@ -94,13 +94,13 @@ dict_create_sys_tables_tuple(
 	dfield = dtuple_get_nth_field(entry, 3);
 
 	ptr = mem_heap_alloc(heap, 4);
-	if (table->flags & ~DICT_TF_COMPACT) {
+	if (table->flags & (~DICT_TF_COMPACT & ~(~0 << DICT_TF_BITS))) {
 		ut_a(table->flags & DICT_TF_COMPACT);
 		ut_a(dict_table_get_format(table) >= DICT_TF_FORMAT_ZIP);
 		ut_a((table->flags & DICT_TF_ZSSIZE_MASK)
 		     <= (DICT_TF_ZSSIZE_MAX << DICT_TF_ZSSIZE_SHIFT));
-		ut_a(!(table->flags & (~0 << DICT_TF_BITS)));
-		mach_write_to_4(ptr, table->flags);
+		ut_a(!(table->flags & (~0 << DICT_TF2_BITS)));
+		mach_write_to_4(ptr, table->flags & ~(~0 << DICT_TF_BITS));
 	} else {
 		mach_write_to_4(ptr, DICT_TABLE_ORDINARY);
 	}
@@ -112,11 +112,12 @@ dict_create_sys_tables_tuple(
 	ptr = mem_heap_zalloc(heap, 8);
 
 	dfield_set_data(dfield, ptr, 8);
-	/* 7: MIX_LEN (obsolete) --------------------------*/
+	/* 7: MIX_LEN (additional flags) --------------------------*/
 
 	dfield = dtuple_get_nth_field(entry, 5);
 
-	ptr = mem_heap_zalloc(heap, 4);
+	ptr = mem_heap_alloc(heap, 4);
+	mach_write_to_4(ptr, table->flags >> DICT_TF2_SHIFT);
 
 	dfield_set_data(dfield, ptr, 4);
 	/* 8: CLUSTER_NAME ---------------------*/
@@ -230,6 +231,7 @@ dict_build_table_def_step(
 	dict_table_t*	table;
 	dtuple_t*	row;
 	ulint		error;
+	ulint		flags;
 	const char*	path_or_name;
 	ibool		is_path;
 	mtr_t		mtr;
@@ -268,9 +270,10 @@ dict_build_table_def_step(
 		ut_ad(!dict_table_zip_size(table)
 		      || dict_table_get_format(table) >= DICT_TF_FORMAT_ZIP);
 
+		flags = table->flags & ~(~0 << DICT_TF_BITS);
 		error = fil_create_new_single_table_tablespace(
 			&space, path_or_name, is_path,
-			table->flags == DICT_TF_COMPACT ? 0 : table->flags,
+			flags == DICT_TF_COMPACT ? 0 : flags,
 			FIL_IBD_FILE_INITIAL_SIZE);
 		table->space = (unsigned int) space;
 
@@ -286,7 +289,7 @@ dict_build_table_def_step(
 		mtr_commit(&mtr);
 	} else {
 		/* Create in the system tablespace: disallow new features */
-		table->flags &= DICT_TF_COMPACT;
+		table->flags &= (~0 << DICT_TF_BITS) | DICT_TF_COMPACT;
 	}
 
 	row = dict_create_sys_tables_tuple(table, node->heap);
diff --git a/storage/innodb_plugin/dict/dict0load.c b/storage/innodb_plugin/dict/dict0load.c
index 842a129c1a6..2867125e39d 100644
--- a/storage/innodb_plugin/dict/dict0load.c
+++ b/storage/innodb_plugin/dict/dict0load.c
@@ -390,15 +390,35 @@ loop:
 
 		mtr_commit(&mtr);
 
-		if (space_id != 0 && in_crash_recovery) {
+		if (space_id == 0) {
+			/* The system tablespace always exists. */
+		} else if (in_crash_recovery) {
 			/* Check that the tablespace (the .ibd file) really
-			exists; print a warning to the .err log if not */
+			exists; print a warning to the .err log if not.
+			Do not print warnings for temporary tables. */
+			ibool	is_temp;
 
-			fil_space_for_table_exists_in_mem(space_id, name,
-							  FALSE, TRUE, TRUE);
-		}
+			field = rec_get_nth_field_old(rec, 4, &len);
+			if (0x80000000UL &  mach_read_from_4(field)) {
+				/* ROW_FORMAT=COMPACT: read the is_temp
+				flag from SYS_TABLES.MIX_LEN. */
+				field = rec_get_nth_field_old(rec, 7, &len);
+				is_temp = mach_read_from_4(field)
+					& DICT_TF2_TEMPORARY;
+			} else {
+				/* For tables created with old versions
+				of InnoDB, SYS_TABLES.MIX_LEN may contain
+				garbage.  Such tables would always be
+				in ROW_FORMAT=REDUNDANT.  Pretend that
+				all such tables are non-temporary.  That is,
+				do not suppress error printouts about
+				temporary tables not being found. */
+				is_temp = FALSE;
+			}
 
-		if (space_id != 0 && !in_crash_recovery) {
+			fil_space_for_table_exists_in_mem(
+				space_id, name, is_temp, TRUE, !is_temp);
+		} else {
 			/* It is a normal database startup: create the space
 			object and check that the .ibd file exists. */
 
@@ -894,31 +914,6 @@ err_exit:
 				(ulong) flags);
 			goto err_exit;
 		}
-
-		if (fil_space_for_table_exists_in_mem(space, name, FALSE,
-						      FALSE, FALSE)) {
-			/* Ok; (if we did a crash recovery then the tablespace
-			can already be in the memory cache) */
-		} else {
-			/* In >= 4.1.9, InnoDB scans the data dictionary also
-			at a normal mysqld startup. It is an error if the
-			space object does not exist in memory. */
-
-			ut_print_timestamp(stderr);
-			fprintf(stderr,
-				"  InnoDB: error: space object of table %s,\n"
-				"InnoDB: space id %lu did not exist in memory."
-				" Retrying an open.\n",
-				name, (ulong)space);
-			/* Try to open the tablespace */
-			if (!fil_open_single_table_tablespace(
-				    TRUE, space, flags, name)) {
-				/* We failed to find a sensible tablespace
-				file */
-
-				ibd_file_missing = TRUE;
-			}
-		}
 	} else {
 		flags = 0;
 	}
@@ -928,9 +923,63 @@ err_exit:
 	field = rec_get_nth_field_old(rec, 4, &len);
 	n_cols = mach_read_from_4(field);
 
-	/* The high-order bit of N_COLS is the "compact format" flag. */
+	/* The high-order bit of N_COLS is the "compact format" flag.
+	For tables in that format, MIX_LEN may hold additional flags. */
 	if (n_cols & 0x80000000UL) {
+		ulint	flags2;
+
 		flags |= DICT_TF_COMPACT;
+
+		ut_a(name_of_col_is(sys_tables, sys_index, 7, "MIX_LEN"));
+		field = rec_get_nth_field_old(rec, 7, &len);
+
+		flags2 = mach_read_from_4(field);
+
+		if (flags2 & (~0 << (DICT_TF2_BITS - DICT_TF2_SHIFT))) {
+			ut_print_timestamp(stderr);
+			fputs("  InnoDB: Warning: table ", stderr);
+			ut_print_filename(stderr, name);
+			fprintf(stderr, "\n"
+				"InnoDB: in InnoDB data dictionary"
+				" has unknown flags %lx.\n",
+				(ulong) flags2);
+
+			flags2 &= ~(~0 << (DICT_TF2_BITS - DICT_TF2_SHIFT));
+		}
+
+		flags |= flags2 << DICT_TF2_SHIFT;
+	}
+
+	/* See if the tablespace is available. */
+	if (space == 0) {
+		/* The system tablespace is always available. */
+	} else if (!fil_space_for_table_exists_in_mem(
+			   space, name,
+			   (flags >> DICT_TF2_SHIFT) & DICT_TF2_TEMPORARY,
+			   FALSE, FALSE)) {
+
+		if ((flags >> DICT_TF2_SHIFT) & DICT_TF2_TEMPORARY) {
+			/* Do not bother to retry opening temporary tables. */
+			ibd_file_missing = TRUE;
+		} else {
+			ut_print_timestamp(stderr);
+			fprintf(stderr,
+				"  InnoDB: error: space object of table");
+			ut_print_filename(stderr, name);
+			fprintf(stderr, ",\n"
+				"InnoDB: space id %lu did not exist in memory."
+				" Retrying an open.\n",
+				(ulong) space);
+			/* Try to open the tablespace */
+			if (!fil_open_single_table_tablespace(
+				    TRUE, space,
+				    flags & ~(~0 << DICT_TF_BITS), name)) {
+				/* We failed to find a sensible
+				tablespace file */
+
+				ibd_file_missing = TRUE;
+			}
+		}
 	}
 
 	table = dict_mem_table_create(name, space, n_cols & ~0x80000000UL,
diff --git a/storage/innodb_plugin/dict/dict0mem.c b/storage/innodb_plugin/dict/dict0mem.c
index 6458cbab92d..66b4b43f296 100644
--- a/storage/innodb_plugin/dict/dict0mem.c
+++ b/storage/innodb_plugin/dict/dict0mem.c
@@ -59,7 +59,7 @@ dict_mem_table_create(
 	mem_heap_t*	heap;
 
 	ut_ad(name);
-	ut_a(!(flags & (~0 << DICT_TF_BITS)));
+	ut_a(!(flags & (~0 << DICT_TF2_BITS)));
 
 	heap = mem_heap_create(DICT_HEAP_SIZE);
 
diff --git a/storage/innodb_plugin/fil/fil0fil.c b/storage/innodb_plugin/fil/fil0fil.c
index 112a0e27d50..b2593765af3 100644
--- a/storage/innodb_plugin/fil/fil0fil.c
+++ b/storage/innodb_plugin/fil/fil0fil.c
@@ -1101,6 +1101,7 @@ fil_space_create(
 	ROW_FORMAT=REDUNDANT (table->flags == 0).  For any other
 	format, the tablespace flags should equal table->flags. */
 	ut_a(flags != DICT_TF_COMPACT);
+	ut_a(!(flags & (~0UL << DICT_TF_BITS)));
 
 try_again:
 	/*printf(
@@ -2586,6 +2587,7 @@ fil_create_new_single_table_tablespace(
 	ROW_FORMAT=REDUNDANT (table->flags == 0).  For any other
 	format, the tablespace flags should equal table->flags. */
 	ut_a(flags != DICT_TF_COMPACT);
+	ut_a(!(flags & (~0UL << DICT_TF_BITS)));
 
 	path = fil_make_ibd_name(tablename, is_temp);
 
@@ -2958,8 +2960,10 @@ fil_open_single_table_tablespace(
 	/* The tablespace flags (FSP_SPACE_FLAGS) should be 0 for
 	ROW_FORMAT=COMPACT (table->flags == DICT_TF_COMPACT) and
 	ROW_FORMAT=REDUNDANT (table->flags == 0).  For any other
-	format, the tablespace flags should equal table->flags. */
+	format, the tablespace flags should be equal to
+	table->flags & ~(~0 << DICT_TF_BITS). */
 	ut_a(flags != DICT_TF_COMPACT);
+	ut_a(!(flags & (~0UL << DICT_TF_BITS)));
 
 	file = os_file_create_simple_no_error_handling(
 		filepath, OS_FILE_OPEN, OS_FILE_READ_ONLY, &success);
@@ -3011,7 +3015,8 @@ fil_open_single_table_tablespace(
 
 	ut_free(buf2);
 
-	if (UNIV_UNLIKELY(space_id != id || space_flags != flags)) {
+	if (UNIV_UNLIKELY(space_id != id
+			  || space_flags != (flags & ~(~0 << DICT_TF_BITS)))) {
 		ut_print_timestamp(stderr);
 
 		fputs("  InnoDB: Error: tablespace id and flags in file ",
diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 47b8203091c..f67334ac455 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -6333,6 +6333,10 @@ ha_innobase::create(
 		goto cleanup;
 	}
 
+	if (create_info->options & HA_LEX_CREATE_TMP_TABLE) {
+		flags |= DICT_TF2_TEMPORARY << DICT_TF2_SHIFT;
+	}
+
 	error = create_table_def(trx, form, norm_name,
 		create_info->options & HA_LEX_CREATE_TMP_TABLE ? name2 : NULL,
 		flags);
diff --git a/storage/innodb_plugin/include/dict0mem.h b/storage/innodb_plugin/include/dict0mem.h
index 2d001111938..9996fb59a75 100644
--- a/storage/innodb_plugin/include/dict0mem.h
+++ b/storage/innodb_plugin/include/dict0mem.h
@@ -80,21 +80,39 @@ combination of types */
 /** File format */
 /* @{ */
 #define DICT_TF_FORMAT_SHIFT		5	/* file format */
-#define DICT_TF_FORMAT_MASK		(127 << DICT_TF_FORMAT_SHIFT)
+#define DICT_TF_FORMAT_MASK		\
+((~(~0 << (DICT_TF_BITS - DICT_TF_FORMAT_SHIFT))) << DICT_TF_FORMAT_SHIFT)
 #define DICT_TF_FORMAT_51		0	/*!< InnoDB/MySQL up to 5.1 */
 #define DICT_TF_FORMAT_ZIP		1	/*!< InnoDB plugin for 5.1:
 						compressed tables,
 						new BLOB treatment */
 /** Maximum supported file format */
 #define DICT_TF_FORMAT_MAX		DICT_TF_FORMAT_ZIP
-
+/* @} */
 #define DICT_TF_BITS			6	/*!< number of flag bits */
 #if (1 << (DICT_TF_BITS - DICT_TF_FORMAT_SHIFT)) <= DICT_TF_FORMAT_MAX
 # error "DICT_TF_BITS is insufficient for DICT_TF_FORMAT_MAX"
 #endif
 /* @} */
+
+/** @brief Additional table flags.
+
+These flags will be stored in SYS_TABLES.MIX_LEN.  All unused flags
+will be written as 0.  The column may contain garbage for tables
+created with old versions of InnoDB that only implemented
+ROW_FORMAT=REDUNDANT. */
+/* @{ */
+#define DICT_TF2_SHIFT			DICT_TF_BITS
+						/*!< Shift value for
+						table->flags. */
+#define DICT_TF2_TEMPORARY		1	/*!< TRUE for tables from
+						CREATE TEMPORARY TABLE. */
+#define DICT_TF2_BITS			(DICT_TF2_SHIFT + 1)
+						/*!< Total number of bits
+						in table->flags. */
 /* @} */
 
+
 /**********************************************************************//**
 Creates a table memory object.
 @return	own: table object */
@@ -374,7 +392,7 @@ struct dict_table_struct{
 	unsigned	space:32;
 				/*!< space where the clustered index of the
 				table is placed */
-	unsigned	flags:DICT_TF_BITS;/*!< DICT_TF_COMPACT, ... */
+	unsigned	flags:DICT_TF2_BITS;/*!< DICT_TF_COMPACT, ... */
 	unsigned	ibd_file_missing:1;
 				/*!< TRUE if this is in a single-table
 				tablespace and the .ibd file is missing; then
diff --git a/storage/innodb_plugin/row/row0mysql.c b/storage/innodb_plugin/row/row0mysql.c
index 181c39de881..7a43d0f3b92 100644
--- a/storage/innodb_plugin/row/row0mysql.c
+++ b/storage/innodb_plugin/row/row0mysql.c
@@ -3264,7 +3264,7 @@ check_next_foreign:
 
 		ut_error;
 	} else {
-		ibool		is_path;
+		ibool		is_temp;
 		const char*	name_or_path;
 		mem_heap_t*	heap;
 
@@ -3277,12 +3277,13 @@ check_next_foreign:
 		space_id = table->space;
 
 		if (table->dir_path_of_temp_table != NULL) {
-			is_path = TRUE;
 			name_or_path = mem_heap_strdup(
 				heap, table->dir_path_of_temp_table);
+			is_temp = TRUE;
 		} else {
-			is_path = FALSE;
 			name_or_path = name;
+			is_temp = (table->flags >> DICT_TF2_SHIFT)
+				& DICT_TF2_TEMPORARY;
 		}
 
 		dict_table_remove_from_cache(table);
@@ -3302,8 +3303,8 @@ check_next_foreign:
 		if (err == DB_SUCCESS && space_id > 0) {
 			if (!fil_space_for_table_exists_in_mem(space_id,
 							       name_or_path,
-							       is_path,
-							       FALSE, TRUE)) {
+							       is_temp, FALSE,
+							       !is_temp)) {
 				err = DB_SUCCESS;
 
 				fprintf(stderr,

From 4dccc823e3555fd903a1a947c8c88bb4c46e9088 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 14:50:16 +0400
Subject: [PATCH 003/116] Applying InnoDB snapshot

Detailed revision comments:

r6263 | vasil | 2009-12-01 14:49:05 +0200 (Tue, 01 Dec 2009) | 4 lines
branches/zip: Increment version number from 1.0.6 to 1.0.7

1.0.6 has been released
---
 storage/innodb_plugin/include/univ.i | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/include/univ.i b/storage/innodb_plugin/include/univ.i
index 2081e136590..a083620ab6d 100644
--- a/storage/innodb_plugin/include/univ.i
+++ b/storage/innodb_plugin/include/univ.i
@@ -46,7 +46,7 @@ Created 1/20/1994 Heikki Tuuri
 
 #define INNODB_VERSION_MAJOR	1
 #define INNODB_VERSION_MINOR	0
-#define INNODB_VERSION_BUGFIX	6
+#define INNODB_VERSION_BUGFIX	7
 
 /* The following is the InnoDB version as shown in
 SELECT plugin_version FROM information_schema.plugins;

From 22e1dde05195764fd70946bbf957e13611093437 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 14:51:49 +0400
Subject: [PATCH 004/116] Applying InnoDB snapshot

Detailed revision comments:

r6264 | vasil | 2009-12-01 16:19:44 +0200 (Tue, 01 Dec 2009) | 1 line
branches/zip: Add ChangeLog entry for the release of 1.0.6.
---
 storage/innodb_plugin/ChangeLog | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index ac66ff10963..34e5fc426d0 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -8,6 +8,10 @@
 	Bug#41609 Crash recovery does not work for InnoDB temporary tables.
 	The actual crash recovery bug was corrected on 2009-04-29.
 
+2009-11-27	The InnoDB Team
+
+	InnoDB Plugin 1.0.6 released
+
 2009-11-20	The InnoDB Team
 
 	* handler/ha_innodb.cc:

From e20fd747f453cdd9bdd0721fcd7d3216910fdd47 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 14:52:21 +0400
Subject: [PATCH 005/116] Applying InnoDB snapshot

Detailed revision comments:

r6269 | marko | 2009-12-02 11:35:22 +0200 (Wed, 02 Dec 2009) | 2 lines
branches/zip: innobase_start_or_create_for_mysql(): UNIV_IBUF_DEBUG
should not break crash recovery, but UNIV_IBUF_COUNT_DEBUG will.
---
 storage/innodb_plugin/srv/srv0start.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/srv/srv0start.c b/storage/innodb_plugin/srv/srv0start.c
index d5f6120ca31..9d6169f7aaf 100644
--- a/storage/innodb_plugin/srv/srv0start.c
+++ b/storage/innodb_plugin/srv/srv0start.c
@@ -1074,7 +1074,11 @@ innobase_start_or_create_for_mysql(void)
 #ifdef UNIV_IBUF_DEBUG
 	fprintf(stderr,
 		"InnoDB: !!!!!!!! UNIV_IBUF_DEBUG switched on !!!!!!!!!\n"
-		"InnoDB: Crash recovery will fail with UNIV_IBUF_DEBUG\n");
+# ifdef UNIV_IBUF_COUNT_DEBUG
+		"InnoDB: !!!!!!!! UNIV_IBUF_COUNT_DEBUG switched on !!!!!!!!!\n"
+		"InnoDB: Crash recovery will fail with UNIV_IBUF_COUNT_DEBUG\n"
+# endif
+		);
 #endif
 
 #ifdef UNIV_SYNC_DEBUG

From 4a21f29cf0604fa5c366f31ab2411304422190f3 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 14:53:07 +0400
Subject: [PATCH 006/116] Applying InnoDB snapshot

Detailed revision comments:

r6270 | marko | 2009-12-02 11:36:47 +0200 (Wed, 02 Dec 2009) | 1 line
branches/zip: innobase_start_or_create_for_mysql(): Log the zlib version.
---
 storage/innodb_plugin/srv/srv0start.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/srv/srv0start.c b/storage/innodb_plugin/srv/srv0start.c
index 9d6169f7aaf..a0505f43852 100644
--- a/storage/innodb_plugin/srv/srv0start.c
+++ b/storage/innodb_plugin/srv/srv0start.c
@@ -105,6 +105,7 @@ Created 2/16/1996 Heikki Tuuri
 # include "btr0pcur.h"
 # include "thr0loc.h"
 # include "os0sync.h" /* for INNODB_RW_LOCKS_USE_ATOMICS */
+# include "zlib.h" /* for ZLIB_VERSION */
 
 /** Log sequence number immediately after startup */
 UNIV_INTERN ib_uint64_t	srv_start_lsn;
@@ -1105,7 +1106,15 @@ innobase_start_or_create_for_mysql(void)
 			"InnoDB: The InnoDB memory heap is disabled\n");
 	}
 
-	fprintf(stderr, "InnoDB: %s\n", IB_ATOMICS_STARTUP_MSG);
+	fputs("InnoDB: " IB_ATOMICS_STARTUP_MSG
+	      "\nInnoDB: Compressed tables use zlib " ZLIB_VERSION
+#ifdef UNIV_ZIP_DEBUG
+	      " with validation"
+#endif /* UNIV_ZIP_DEBUG */
+#ifdef UNIV_ZIP_COPY
+	      " and extra copying"
+#endif /* UNIV_ZIP_COPY */
+	      "\n" , stderr);
 
 	/* Since InnoDB does not currently clean up all its internal data
 	structures in MySQL Embedded Server Library server_end(), we

From 926d2c4bc0a0d075d798f083dc5e114881387f63 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 14:58:37 +0400
Subject: [PATCH 007/116] Applying InnoDB snapshot

Detailed revision comments:

r6271 | marko | 2009-12-02 11:43:49 +0200 (Wed, 02 Dec 2009) | 2 lines
branches/zip: ChangeLog: Document that since r6270, the zlib version number
will be displayed at start-up.
---
 storage/innodb_plugin/ChangeLog      | 8 ++++++++
 storage/innodb_plugin/include/univ.i | 2 +-
 2 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 34e5fc426d0..dddf37e2334 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,11 @@
+2009-12-02	The InnoDB Team
+
+	* srv/srv0start.c: Display the zlib version number at startup.
+	InnoDB compressed tables use zlib, and the implementation depends
+	on the zlib function compressBound(), whose definition was slightly
+	changed in zlib version 1.2.3.1 in 2006.  MySQL bundles zlib 1.2.3
+	from 2005, but some installations use a more recent zlib.
+
 2009-11-30	The InnoDB Team
 
 	* dict/dict0crea.c, dict/dict0mem.c, dict/dict0load.c,
diff --git a/storage/innodb_plugin/include/univ.i b/storage/innodb_plugin/include/univ.i
index a083620ab6d..939711ac110 100644
--- a/storage/innodb_plugin/include/univ.i
+++ b/storage/innodb_plugin/include/univ.i
@@ -204,11 +204,11 @@ operations (very slow); also UNIV_DEBUG must be defined */
 						in sync0sync.c */
 #define UNIV_BTR_PRINT				/* enable functions for
 						printing B-trees */
+#endif
 #define UNIV_ZIP_DEBUG				/* extensive consistency checks
 						for compressed pages */
 #define UNIV_ZIP_COPY				/* call page_zip_copy_recs()
 						more often */
-#endif
 
 #define UNIV_BTR_DEBUG				/* check B-tree links */
 #define UNIV_LIGHT_MEM_DEBUG			/* light memory debugging */

From fcb1ddcf1f189399cd3a0f39d7e3af0ae69e0e8a Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:00:09 +0400
Subject: [PATCH 008/116] Applying InnoDB snapshot

Detailed revision comments:

r6272 | marko | 2009-12-02 11:46:05 +0200 (Wed, 02 Dec 2009) | 1 line
branches/zip: Revert changes that were accidentally committed in r6271.
---
 storage/innodb_plugin/include/univ.i | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/include/univ.i b/storage/innodb_plugin/include/univ.i
index 939711ac110..a083620ab6d 100644
--- a/storage/innodb_plugin/include/univ.i
+++ b/storage/innodb_plugin/include/univ.i
@@ -204,11 +204,11 @@ operations (very slow); also UNIV_DEBUG must be defined */
 						in sync0sync.c */
 #define UNIV_BTR_PRINT				/* enable functions for
 						printing B-trees */
-#endif
 #define UNIV_ZIP_DEBUG				/* extensive consistency checks
 						for compressed pages */
 #define UNIV_ZIP_COPY				/* call page_zip_copy_recs()
 						more often */
+#endif
 
 #define UNIV_BTR_DEBUG				/* check B-tree links */
 #define UNIV_LIGHT_MEM_DEBUG			/* light memory debugging */

From d678a4bdf42ad72331e76ffc1d8ee77c48893ce9 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:00:30 +0400
Subject: [PATCH 009/116] Applying InnoDB snapshot

Detailed revision comments:

r6274 | marko | 2009-12-03 14:47:12 +0200 (Thu, 03 Dec 2009) | 6 lines
branches/zip: dict_table_check_for_dup_indexes(): Assert that the
data dictionary mutex is being held while table->indexes is accessed.
This is already the case.

Currently, only dict_table_get_next_index() and dict_table_get_first_index()
are being invoked without holding dict_sys->mutex.
---
 storage/innodb_plugin/dict/dict0dict.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/storage/innodb_plugin/dict/dict0dict.c b/storage/innodb_plugin/dict/dict0dict.c
index 2e524a5a2e3..4c62e8de748 100644
--- a/storage/innodb_plugin/dict/dict0dict.c
+++ b/storage/innodb_plugin/dict/dict0dict.c
@@ -4775,6 +4775,8 @@ dict_table_check_for_dup_indexes(
 	const dict_index_t*	index1;
 	const dict_index_t*	index2;
 
+	ut_ad(mutex_own(&dict_sys->mutex));
+
 	/* The primary index _must_ exist */
 	ut_a(UT_LIST_GET_LEN(table->indexes) > 0);
 

From 971085422ec11981ba43a5237509ab28181f14aa Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:01:13 +0400
Subject: [PATCH 010/116] Applying InnoDB snapshot

Detailed revision comments:

r6275 | pekka | 2009-12-03 18:32:47 +0200 (Thu, 03 Dec 2009) | 10 lines
branches/zip: Minor changes which allow build with UNIV_HOTBACKUP
defined to succeed:

include/trx0sys.h: Allow Hot Backup build to see some
                   TRX_SYS_DOUBLEWRITE_... macros.
trx/trx0sys.c:     Exclude trx_sys_close() function from Hot Backup build.
log/log0recv.[ch]: Exclude recv_sys_var_init() function from Hot Backup build.

This change should not affect !UNIV_HOTBACKUP build.
---
 storage/innodb_plugin/include/log0recv.h | 2 ++
 storage/innodb_plugin/include/trx0sys.h  | 4 +++-
 storage/innodb_plugin/log/log0recv.c     | 2 ++
 storage/innodb_plugin/trx/trx0sys.c      | 2 ++
 4 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/include/log0recv.h b/storage/innodb_plugin/include/log0recv.h
index a3d2bd050f5..35576bb579d 100644
--- a/storage/innodb_plugin/include/log0recv.h
+++ b/storage/innodb_plugin/include/log0recv.h
@@ -258,12 +258,14 @@ void
 recv_sys_init(
 /*==========*/
 	ulint	available_memory);	/*!< in: available memory in bytes */
+#ifndef UNIV_HOTBACKUP
 /********************************************************//**
 Reset the state of the recovery system variables. */
 UNIV_INTERN
 void
 recv_sys_var_init(void);
 /*===================*/
+#endif /* !UNIV_HOTBACKUP */
 /*******************************************************************//**
 Empties the hash table of stored log records, applying them to appropriate
 pages. */
diff --git a/storage/innodb_plugin/include/trx0sys.h b/storage/innodb_plugin/include/trx0sys.h
index a53296a06d9..cbb89689748 100644
--- a/storage/innodb_plugin/include/trx0sys.h
+++ b/storage/innodb_plugin/include/trx0sys.h
@@ -333,12 +333,14 @@ UNIV_INTERN
 void
 trx_sys_file_format_tag_init(void);
 /*==============================*/
+#ifndef UNIV_HOTBACKUP
 /*****************************************************************//**
 Shutdown/Close the transaction system. */
 UNIV_INTERN
 void
 trx_sys_close(void);
 /*===============*/
+#endif /* !UNIV_HOTBACKUP */
 /*****************************************************************//**
 Get the name representation of the file format from its id.
 @return	pointer to the name */
@@ -495,7 +497,6 @@ this contains the same fields as TRX_SYS_MYSQL_LOG_INFO below */
 						within that file */
 #define TRX_SYS_MYSQL_LOG_NAME		12	/*!< MySQL log file name */
 
-#ifndef UNIV_HOTBACKUP
 /** Doublewrite buffer */
 /* @{ */
 /** The offset of the doublewrite buffer header on the trx system header page */
@@ -547,6 +548,7 @@ FIL_PAGE_ARCH_LOG_NO_OR_SPACE_NO. */
 #define TRX_SYS_DOUBLEWRITE_BLOCK_SIZE	FSP_EXTENT_SIZE
 /* @} */
 
+#ifndef UNIV_HOTBACKUP
 /** File format tag */
 /* @{ */
 /** The offset of the file format tag on the trx system header page
diff --git a/storage/innodb_plugin/log/log0recv.c b/storage/innodb_plugin/log/log0recv.c
index ddbc71d4b71..0aff9fcd9bf 100644
--- a/storage/innodb_plugin/log/log0recv.c
+++ b/storage/innodb_plugin/log/log0recv.c
@@ -239,6 +239,7 @@ recv_sys_mem_free(void)
 	}
 }
 
+#ifndef UNIV_HOTBACKUP
 /************************************************************
 Reset the state of the recovery system variables. */
 UNIV_INTERN
@@ -278,6 +279,7 @@ recv_sys_var_init(void)
 
 	recv_max_page_lsn = 0;
 }
+#endif /* !UNIV_HOTBACKUP */
 
 /************************************************************
 Inits the recovery system for a recovery operation. */
diff --git a/storage/innodb_plugin/trx/trx0sys.c b/storage/innodb_plugin/trx/trx0sys.c
index 79e5af1c677..253619545af 100644
--- a/storage/innodb_plugin/trx/trx0sys.c
+++ b/storage/innodb_plugin/trx/trx0sys.c
@@ -1535,6 +1535,7 @@ trx_sys_file_format_id_to_name(
 
 #endif /* !UNIV_HOTBACKUP */
 
+#ifndef UNIV_HOTBACKUP
 /*********************************************************************
 Shutdown/Close the transaction system. */
 UNIV_INTERN
@@ -1611,3 +1612,4 @@ trx_sys_close(void)
 	trx_sys = NULL;
 	mutex_exit(&kernel_mutex);
 }
+#endif /* !UNIV_HOTBACKUP */

From 058a71f2ba4cca63fcaf99dd0114e8f9008f7c25 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:01:56 +0400
Subject: [PATCH 011/116] Applying InnoDB snapshot

Detailed revision comments:

r6277 | marko | 2009-12-08 11:13:36 +0200 (Tue, 08 Dec 2009) | 1 line
branches/zip: fsp0fsp.c: Add some missing in/out and const qualifiers.
---
 storage/innodb_plugin/fsp/fsp0fsp.c | 32 ++++++++++++++---------------
 1 file changed, 16 insertions(+), 16 deletions(-)

diff --git a/storage/innodb_plugin/fsp/fsp0fsp.c b/storage/innodb_plugin/fsp/fsp0fsp.c
index 3cc4318fc06..9b53e5585be 100644
--- a/storage/innodb_plugin/fsp/fsp0fsp.c
+++ b/storage/innodb_plugin/fsp/fsp0fsp.c
@@ -386,11 +386,11 @@ UNIV_INLINE
 ibool
 xdes_get_bit(
 /*=========*/
-	xdes_t*	descr,	/*!< in: descriptor */
-	ulint	bit,	/*!< in: XDES_FREE_BIT or XDES_CLEAN_BIT */
-	ulint	offset,	/*!< in: page offset within extent:
-			0 ... FSP_EXTENT_SIZE - 1 */
-	mtr_t*	mtr)	/*!< in: mtr */
+	const xdes_t*	descr,	/*!< in: descriptor */
+	ulint		bit,	/*!< in: XDES_FREE_BIT or XDES_CLEAN_BIT */
+	ulint		offset,	/*!< in: page offset within extent:
+				0 ... FSP_EXTENT_SIZE - 1 */
+	mtr_t*		mtr)	/*!< in: mtr */
 {
 	ulint	index;
 	ulint	byte_index;
@@ -527,8 +527,8 @@ UNIV_INLINE
 ulint
 xdes_get_n_used(
 /*============*/
-	xdes_t*	descr,	/*!< in: descriptor */
-	mtr_t*	mtr)	/*!< in: mtr */
+	const xdes_t*	descr,	/*!< in: descriptor */
+	mtr_t*		mtr)	/*!< in: mtr */
 {
 	ulint	i;
 	ulint	count	= 0;
@@ -551,8 +551,8 @@ UNIV_INLINE
 ibool
 xdes_is_free(
 /*=========*/
-	xdes_t*	descr,	/*!< in: descriptor */
-	mtr_t*	mtr)	/*!< in: mtr */
+	const xdes_t*	descr,	/*!< in: descriptor */
+	mtr_t*		mtr)	/*!< in: mtr */
 {
 	if (0 == xdes_get_n_used(descr, mtr)) {
 
@@ -569,8 +569,8 @@ UNIV_INLINE
 ibool
 xdes_is_full(
 /*=========*/
-	xdes_t*	descr,	/*!< in: descriptor */
-	mtr_t*	mtr)	/*!< in: mtr */
+	const xdes_t*	descr,	/*!< in: descriptor */
+	mtr_t*		mtr)	/*!< in: mtr */
 {
 	if (FSP_EXTENT_SIZE == xdes_get_n_used(descr, mtr)) {
 
@@ -586,7 +586,7 @@ UNIV_INLINE
 void
 xdes_set_state(
 /*===========*/
-	xdes_t*	descr,	/*!< in: descriptor */
+	xdes_t*	descr,	/*!< in/out: descriptor */
 	ulint	state,	/*!< in: state to set */
 	mtr_t*	mtr)	/*!< in: mtr handle */
 {
@@ -605,8 +605,8 @@ UNIV_INLINE
 ulint
 xdes_get_state(
 /*===========*/
-	xdes_t*	descr,	/*!< in: descriptor */
-	mtr_t*	mtr)	/*!< in: mtr handle */
+	const xdes_t*	descr,	/*!< in: descriptor */
+	mtr_t*		mtr)	/*!< in: mtr handle */
 {
 	ulint	state;
 
@@ -705,7 +705,7 @@ UNIV_INLINE
 xdes_t*
 xdes_get_descriptor_with_space_hdr(
 /*===============================*/
-	fsp_header_t*	sp_header,/*!< in: space header, x-latched */
+	fsp_header_t*	sp_header,/*!< in/out: space header, x-latched */
 	ulint		space,	/*!< in: space id */
 	ulint		offset,	/*!< in: page offset;
 				if equal to the free limit,
@@ -1342,7 +1342,7 @@ fsp_fill_free_list(
 					descriptor page and ibuf bitmap page;
 					then we do not allocate more extents */
 	ulint		space,		/*!< in: space */
-	fsp_header_t*	header,		/*!< in: space header */
+	fsp_header_t*	header,		/*!< in/out: space header */
 	mtr_t*		mtr)		/*!< in: mtr */
 {
 	ulint	limit;

From 2a187a515397d5d3800767a749a398d3ac774550 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:03:27 +0400
Subject: [PATCH 012/116] Applying InnoDB snapshot

Detailed revision comments:

r6285 | marko | 2009-12-09 09:24:50 +0200 (Wed, 09 Dec 2009) | 13 lines
branches/zip: row_sel_fetch_columns(): Remove redundant code that was
accidentally added in r1591, which introduced dfield_t::ext in order
to make the merge sort of fast index creation support externally
stored columns,

Initially, I tried to allocate the bit for dfield_t::ext from
dfield_t::len by making the length 31 bits and mapping UNIV_SQL_NULL
to something that would fit in it.  Then I decided that it would be
too risky.  The redundant check was part of the mapping.  The
condition may have been dfield_is_null() initially.

This redundant code was noticed by Sergey Petrunya on the MySQL
internals list.
r6288 | marko | 2009-12-09 09:51:00 +0200 (Wed, 09 Dec 2009) | 15 lines
branches/zip: row_upd_copy_columns(): Remove redundant code that was
accidentally added in r1591, which introduced dfield_t::ext in order
to make the merge sort of fast index creation support externally
stored columns.

Initially, I tried to allocate the bit for dfield_t::ext from
dfield_t::len by making the length 31 bits and mapping UNIV_SQL_NULL
to something that would fit in it.  Then I decided that it would be
too risky.  The redundant check was part of the mapping.  The
condition may have been dfield_is_null() initially.

This is similar to the redundant code in row_sel_fetch_columns() that
was noticed by Sergey Petrunya on the MySQL internals list and removed
in r6285.  As far as I can tell, there are no redundant UNIV_SQL_NULL
assignments remaining after this change.
---
 storage/innodb_plugin/row/row0sel.c | 4 ----
 storage/innodb_plugin/row/row0upd.c | 3 ---
 2 files changed, 7 deletions(-)

diff --git a/storage/innodb_plugin/row/row0sel.c b/storage/innodb_plugin/row/row0sel.c
index 3ef9726588e..5f66b9be159 100644
--- a/storage/innodb_plugin/row/row0sel.c
+++ b/storage/innodb_plugin/row/row0sel.c
@@ -431,10 +431,6 @@ row_sel_fetch_columns(
 				data = rec_get_nth_field(rec, offsets,
 							 field_no, &len);
 
-				if (len == UNIV_SQL_NULL) {
-					len = UNIV_SQL_NULL;
-				}
-
 				needs_copy = column->copy_val;
 			}
 
diff --git a/storage/innodb_plugin/row/row0upd.c b/storage/innodb_plugin/row/row0upd.c
index 58dfd43ead9..95d1d00aeef 100644
--- a/storage/innodb_plugin/row/row0upd.c
+++ b/storage/innodb_plugin/row/row0upd.c
@@ -1344,9 +1344,6 @@ row_upd_copy_columns(
 		data = rec_get_nth_field(rec, offsets,
 					 column->field_nos[SYM_CLUST_FIELD_NO],
 					 &len);
-		if (len == UNIV_SQL_NULL) {
-			len = UNIV_SQL_NULL;
-		}
 		eval_node_copy_and_alloc_val(column, data, len);
 
 		column = UT_LIST_GET_NEXT(col_var_list, column);

From f3ca08d2c14af42509203034082494171d461f23 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:07:37 +0400
Subject: [PATCH 013/116] Applying InnoDB snapshot

Detailed revision comments:

r6305 | marko | 2009-12-14 13:03:57 +0200 (Mon, 14 Dec 2009) | 2 lines
branches/zip: row_undo_mod_del_unmark_sec_and_undo_update(): Add a missing
const qualifier.
---
 storage/innodb_plugin/row/row0umod.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/row/row0umod.c b/storage/innodb_plugin/row/row0umod.c
index 6be475d8c78..fb7fd460769 100644
--- a/storage/innodb_plugin/row/row0umod.c
+++ b/storage/innodb_plugin/row/row0umod.c
@@ -438,7 +438,7 @@ row_undo_mod_del_unmark_sec_and_undo_update(
 				BTR_MODIFY_TREE */
 	que_thr_t*	thr,	/*!< in: query thread */
 	dict_index_t*	index,	/*!< in: index */
-	dtuple_t*	entry)	/*!< in: index entry */
+	const dtuple_t*	entry)	/*!< in: index entry */
 {
 	mem_heap_t*	heap;
 	btr_pcur_t	pcur;

From bf7191c1fe6d1b19b570801073a523a7e887dbb5 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:08:06 +0400
Subject: [PATCH 014/116] Applying InnoDB snapshot

Detailed revision comments:

r6309 | marko | 2009-12-15 14:05:50 +0200 (Tue, 15 Dec 2009) | 3 lines
branches/zip: lock_rec_insert_check_and_lock(): Avoid casting away constness.
Use page_rec_get_next_const() instead. This silences a gcc 4.2.4 warning.
Reported by Sunny Bains.
---
 storage/innodb_plugin/lock/lock0lock.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/lock/lock0lock.c b/storage/innodb_plugin/lock/lock0lock.c
index 1fce8002bdf..8dfeb064863 100644
--- a/storage/innodb_plugin/lock/lock0lock.c
+++ b/storage/innodb_plugin/lock/lock0lock.c
@@ -4918,7 +4918,7 @@ lock_rec_insert_check_and_lock(
 	}
 
 	trx = thr_get_trx(thr);
-	next_rec = page_rec_get_next((rec_t*) rec);
+	next_rec = page_rec_get_next_const(rec);
 	next_rec_heap_no = page_rec_get_heap_no(next_rec);
 
 	lock_mutex_enter_kernel();

From 5065980ee6be686c45fddb735af1967224017f4b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:08:47 +0400
Subject: [PATCH 015/116] Applying InnoDB snapshot

Detailed revision comments:

r6312 | marko | 2009-12-16 10:10:36 +0200 (Wed, 16 Dec 2009) | 6 lines
branches/zip: fil_close(): Add #ifndef UNIV_HOTBACKUP around a debug
assertion on mutex.magic_n.  InnoDB Hot Backup is a single-threaded
program and does not contain mutexes.  This change allows InnoDB Hot
Backup to be compiled with UNIV_DEBUG.

Suggested by Michael Izioumtchenko.
---
 storage/innodb_plugin/fil/fil0fil.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/storage/innodb_plugin/fil/fil0fil.c b/storage/innodb_plugin/fil/fil0fil.c
index b2593765af3..8819113f833 100644
--- a/storage/innodb_plugin/fil/fil0fil.c
+++ b/storage/innodb_plugin/fil/fil0fil.c
@@ -4786,8 +4786,10 @@ void
 fil_close(void)
 /*===========*/
 {
+#ifndef UNIV_HOTBACKUP
 	/* The mutex should already have been freed. */
 	ut_ad(fil_system->mutex.magic_n == 0);
+#endif /* !UNIV_HOTBACKUP */
 
 	hash_table_free(fil_system->spaces);
 

From 357e5c0e295886187fa6a52e9f0260804bf94271 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:09:15 +0400
Subject: [PATCH 016/116] Applying InnoDB snapshot

Detailed revision comments:

r6321 | marko | 2009-12-16 16:16:33 +0200 (Wed, 16 Dec 2009) | 4 lines
branches/zip: row_merge_drop_temp_indexes(): Revert a hack to
transaction isolation level that was made unnecessary by r5826 (Issue #337).
When this function is called, any active data dictionary transaction
should have been rolled back.
---
 storage/innodb_plugin/row/row0merge.c | 9 ---------
 1 file changed, 9 deletions(-)

diff --git a/storage/innodb_plugin/row/row0merge.c b/storage/innodb_plugin/row/row0merge.c
index 25f041c0885..a9bcb7df55c 100644
--- a/storage/innodb_plugin/row/row0merge.c
+++ b/storage/innodb_plugin/row/row0merge.c
@@ -2062,15 +2062,6 @@ row_merge_drop_temp_indexes(void)
 	trx->op_info = "dropping partially created indexes";
 	row_mysql_lock_data_dictionary(trx);
 
-	/* Incomplete transactions may be holding some locks on the
-	data dictionary tables.  However, they should never have been
-	able to lock the records corresponding to the partially
-	created indexes that we are attempting to delete, because the
-	table was locked when the indexes were being created.  We will
-	drop the partially created indexes before the rollback of
-	incomplete transactions is initiated.  Thus, this should not
-	interfere with the incomplete transactions. */
-	trx->isolation_level = TRX_ISO_READ_UNCOMMITTED;
 	err = que_eval_sql(NULL, drop_temp_indexes, FALSE, trx);
 	ut_a(err == DB_SUCCESS);
 

From 0b33b0c26619597855a87f4bb03307ace63a0a3b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:09:37 +0400
Subject: [PATCH 017/116] Applying InnoDB snapshot

Detailed revision comments:

r6345 | marko | 2009-12-21 10:46:14 +0200 (Mon, 21 Dec 2009) | 7 lines
branches/zip: recv_scan_log_recs(): Non-functional change: Replace a
debug assertion ut_ad(len > 0) with ut_ad(len >= OS_FILE_LOG_BLOCK_SIZE).

This change is only for readability, for Issue #428.  Another
assertion on len being an integer multiple of OS_FILE_LOG_BLOCK_SIZE
already ensured together with the old ut_ad(len > 0) that actually len
must be at least OS_FILE_LOG_BLOCK_SIZE.
---
 storage/innodb_plugin/log/log0recv.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/log/log0recv.c b/storage/innodb_plugin/log/log0recv.c
index 0aff9fcd9bf..b060c9e6e24 100644
--- a/storage/innodb_plugin/log/log0recv.c
+++ b/storage/innodb_plugin/log/log0recv.c
@@ -2558,7 +2558,7 @@ recv_scan_log_recs(
 
 	ut_ad(start_lsn % OS_FILE_LOG_BLOCK_SIZE == 0);
 	ut_ad(len % OS_FILE_LOG_BLOCK_SIZE == 0);
-	ut_ad(len > 0);
+	ut_ad(len >= OS_FILE_LOG_BLOCK_SIZE);
 	ut_a(store_to_hash <= TRUE);
 
 	finished = FALSE;
@@ -3264,6 +3264,8 @@ recv_recovery_from_checkpoint_finish(void)
 
 	/* Drop partially created indexes. */
 	row_merge_drop_temp_indexes();
+	/* Drop temporary tables. */
+	row_mysql_drop_temp_tables();
 
 #ifdef UNIV_SYNC_DEBUG
 	/* Wait for a while so that created threads have time to suspend

From acd6a81b855a7f93ece728c375d7a5020bd753b0 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:10:01 +0400
Subject: [PATCH 018/116] Applying InnoDB snapshot

Detailed revision comments:

r6346 | marko | 2009-12-21 12:03:25 +0200 (Mon, 21 Dec 2009) | 2 lines
branches/zip: recv_recovery_from_checkpoint_finish():
Revert a change that was accidentally committed in r6345.
---
 storage/innodb_plugin/log/log0recv.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/storage/innodb_plugin/log/log0recv.c b/storage/innodb_plugin/log/log0recv.c
index b060c9e6e24..d33381e495d 100644
--- a/storage/innodb_plugin/log/log0recv.c
+++ b/storage/innodb_plugin/log/log0recv.c
@@ -3264,8 +3264,6 @@ recv_recovery_from_checkpoint_finish(void)
 
 	/* Drop partially created indexes. */
 	row_merge_drop_temp_indexes();
-	/* Drop temporary tables. */
-	row_mysql_drop_temp_tables();
 
 #ifdef UNIV_SYNC_DEBUG
 	/* Wait for a while so that created threads have time to suspend

From c58543b3588cb9ae3d482cc34c69b8e2d47db5dd Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:17:10 +0400
Subject: [PATCH 019/116] Applying InnoDB snapshot

Detailed revision comments:

r6348 | marko | 2009-12-22 11:04:34 +0200 (Tue, 22 Dec 2009) | 37 lines
branches/zip: Merge a change from MySQL:
r6351 | marko | 2009-12-22 11:11:18 +0200 (Tue, 22 Dec 2009) | 1 line
branches/zip: Remove an obsolete declaration of LOCK_thread_count.
r6352 | marko | 2009-12-22 12:33:01 +0200 (Tue, 22 Dec 2009) | 104 lines
branches/zip: Merge revisions 6206:6350 from branches/5.1,
except r6347, r6349, r6350 which were committed separately
to both branches, and r6310, which was backported from zip to 5.1.

  ------------------------------------------------------------------------
  r6206 | jyang | 2009-11-20 09:38:43 +0200 (Fri, 20 Nov 2009) | 3 lines
  Changed paths:
     M /branches/5.1/handler/ha_innodb.cc

  branches/5.1: Non-functional change, fix formatting.
  ------------------------------------------------------------------------
  r6230 | sunny | 2009-11-24 23:52:43 +0200 (Tue, 24 Nov 2009) | 3 lines
  Changed paths:
     M /branches/5.1/mysql-test/innodb-autoinc.result

  branches/5.1: Fix autoinc failing test results.
  (this should be skipped when merging 5.1 into zip)
  ------------------------------------------------------------------------
  r6231 | sunny | 2009-11-25 10:26:27 +0200 (Wed, 25 Nov 2009) | 7 lines
  Changed paths:
     M /branches/5.1/mysql-test/innodb-autoinc.result
     M /branches/5.1/mysql-test/innodb-autoinc.test
     M /branches/5.1/row/row0sel.c

  branches/5.1: Fix BUG#49032 - auto_increment field does not initialize to last value in InnoDB Storage Engine.

  We use the appropriate function to read the column value for non-integer
  autoinc column types, namely float and double.

  rb://208. Approved by Marko.
  ------------------------------------------------------------------------
  r6232 | sunny | 2009-11-25 10:27:39 +0200 (Wed, 25 Nov 2009) | 2 lines
  Changed paths:
     M /branches/5.1/row/row0sel.c

  branches/5.1: This is an interim fix, fix white space errors.
  ------------------------------------------------------------------------
  r6233 | sunny | 2009-11-25 10:28:35 +0200 (Wed, 25 Nov 2009) | 2 lines
  Changed paths:
     M /branches/5.1/include/mach0data.h
     M /branches/5.1/include/mach0data.ic
     M /branches/5.1/mysql-test/innodb-autoinc.result
     M /branches/5.1/mysql-test/innodb-autoinc.test
     M /branches/5.1/row/row0sel.c

  branches/5.1: This is an interim fix, fix tests and make read float/double arg const.
  ------------------------------------------------------------------------
  r6234 | sunny | 2009-11-25 10:29:03 +0200 (Wed, 25 Nov 2009) | 2 lines
  Changed paths:
     M /branches/5.1/row/row0sel.c

  branches/5.1: This is an interim fix, fix whitepsace issues.
  ------------------------------------------------------------------------
  r6235 | sunny | 2009-11-26 01:14:42 +0200 (Thu, 26 Nov 2009) | 9 lines
  Changed paths:
     M /branches/5.1/handler/ha_innodb.cc
     M /branches/5.1/mysql-test/innodb-autoinc.result
     M /branches/5.1/mysql-test/innodb-autoinc.test

  branches/5.1: Fix Bug#47720 - REPLACE INTO Autoincrement column with negative values.

  This bug is similiar to the negative autoinc filter patch from earlier,
  with the additional handling of filtering out the negative column values
  set explicitly by the user.

  rb://184
  Approved by Heikki.
  ------------------------------------------------------------------------
  r6242 | vasil | 2009-11-27 22:07:12 +0200 (Fri, 27 Nov 2009) | 4 lines
  Changed paths:
     M /branches/5.1/export.sh

  branches/5.1:

  Minor changes to support plugin snapshots.
  ------------------------------------------------------------------------
  r6306 | calvin | 2009-12-14 15:12:46 +0200 (Mon, 14 Dec 2009) | 5 lines
  Changed paths:
     M /branches/5.1/mysql-test/innodb-autoinc.result
     M /branches/5.1/mysql-test/innodb-autoinc.test

  branches/5.1: fix bug#49267: innodb-autoinc.test fails on windows
  because of different case mode

  There is no change to the InnoDB code, only to fix test case by
  changing "T1" to "t1".
  ------------------------------------------------------------------------
  r6324 | jyang | 2009-12-17 06:54:24 +0200 (Thu, 17 Dec 2009) | 8 lines
  Changed paths:
     M /branches/5.1/handler/ha_innodb.cc
     M /branches/5.1/include/lock0lock.h
     M /branches/5.1/include/srv0srv.h
     M /branches/5.1/lock/lock0lock.c
     M /branches/5.1/log/log0log.c
     M /branches/5.1/srv/srv0srv.c
     M /branches/5.1/srv/srv0start.c

  branches/5.1: Fix bug #47814 - Diagnostics are frequently not
  printed after a long lock wait in InnoDB. Separate out the
  lock wait timeout check thread from monitor information
  printing thread.

  rb://200 Approved by Marko.
  ------------------------------------------------------------------------
r6364 | marko | 2009-12-26 21:06:31 +0200 (Sat, 26 Dec 2009) | 4 lines
branches/zip: ibuf_bitmap_get_map_page():
Define a wrapper macro that passes __FILE__, __LINE__ of the caller
to buf_page_get_gen().
This will ease the diagnosis of the likes of Issue #135.
---
 storage/innodb_plugin/handler/ha_innodb.cc |  33 ++--
 storage/innodb_plugin/ibuf/ibuf0ibuf.c     |  37 +++--
 storage/innodb_plugin/include/lock0lock.h  |  11 +-
 storage/innodb_plugin/include/srv0srv.h    |  24 ++-
 storage/innodb_plugin/lock/lock0lock.c     |  21 ++-
 storage/innodb_plugin/log/log0log.c        |   2 +-
 storage/innodb_plugin/row/row0sel.c        |  26 ++-
 storage/innodb_plugin/srv/srv0srv.c        | 181 ++++++++++++++++-----
 storage/innodb_plugin/srv/srv0start.c      |  15 +-
 9 files changed, 257 insertions(+), 93 deletions(-)

diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index f67334ac455..9c214bab3de 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -4463,24 +4463,29 @@ no_commit:
 			update the table upper limit. Note: last_value
 			will be 0 if get_auto_increment() was not called.*/
 
-			if (auto_inc <= col_max_value
-			    && auto_inc >= prebuilt->autoinc_last_value) {
+			if (auto_inc >= prebuilt->autoinc_last_value) {
 set_max_autoinc:
-				ut_a(prebuilt->autoinc_increment > 0);
+				/* This should filter out the negative
+				values set explicitly by the user. */
+				if (auto_inc <= col_max_value) {
+					ut_a(prebuilt->autoinc_increment > 0);
 
-				ulonglong	need;
-				ulonglong	offset;
+					ulonglong	need;
+					ulonglong	offset;
 
-				offset = prebuilt->autoinc_offset;
-				need = prebuilt->autoinc_increment;
+					offset = prebuilt->autoinc_offset;
+					need = prebuilt->autoinc_increment;
 
-				auto_inc = innobase_next_autoinc(
-					auto_inc, need, offset, col_max_value);
+					auto_inc = innobase_next_autoinc(
+						auto_inc,
+						need, offset, col_max_value);
 
-				err = innobase_set_max_autoinc(auto_inc);
+					err = innobase_set_max_autoinc(
+						auto_inc);
 
-				if (err != DB_SUCCESS) {
-					error = err;
+					if (err != DB_SUCCESS) {
+						error = err;
+					}
 				}
 			}
 			break;
@@ -8191,8 +8196,8 @@ innodb_show_status(
 
 	mutex_enter(&srv_monitor_file_mutex);
 	rewind(srv_monitor_file);
-	srv_printf_innodb_monitor(srv_monitor_file,
-				&trx_list_start, &trx_list_end);
+	srv_printf_innodb_monitor(srv_monitor_file, FALSE,
+				  &trx_list_start, &trx_list_end);
 	flen = ftell(srv_monitor_file);
 	os_file_set_eof(srv_monitor_file);
 
diff --git a/storage/innodb_plugin/ibuf/ibuf0ibuf.c b/storage/innodb_plugin/ibuf/ibuf0ibuf.c
index 08986fac0ef..5e9b4b27611 100644
--- a/storage/innodb_plugin/ibuf/ibuf0ibuf.c
+++ b/storage/innodb_plugin/ibuf/ibuf0ibuf.c
@@ -730,24 +730,41 @@ page containing the descriptor bits for the file page; the bitmap page
 is x-latched */
 static
 page_t*
-ibuf_bitmap_get_map_page(
-/*=====================*/
-	ulint	space,	/*!< in: space id of the file page */
-	ulint	page_no,/*!< in: page number of the file page */
-	ulint	zip_size,/*!< in: compressed page size in bytes;
-			0 for uncompressed pages */
-	mtr_t*	mtr)	/*!< in: mtr */
+ibuf_bitmap_get_map_page_func(
+/*==========================*/
+	ulint		space,	/*!< in: space id of the file page */
+	ulint		page_no,/*!< in: page number of the file page */
+	ulint		zip_size,/*!< in: compressed page size in bytes;
+				0 for uncompressed pages */
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line where called */
+	mtr_t*		mtr)	/*!< in: mtr */
 {
 	buf_block_t*	block;
 
-	block = buf_page_get(space, zip_size,
-			     ibuf_bitmap_page_no_calc(zip_size, page_no),
-			     RW_X_LATCH, mtr);
+	block = buf_page_get_gen(space, zip_size,
+				 ibuf_bitmap_page_no_calc(zip_size, page_no),
+				 RW_X_LATCH, NULL, BUF_GET,
+				 file, line, mtr);
 	buf_block_dbg_add_level(block, SYNC_IBUF_BITMAP);
 
 	return(buf_block_get_frame(block));
 }
 
+/********************************************************************//**
+Gets the ibuf bitmap page where the bits describing a given file page are
+stored.
+@return bitmap page where the file page is mapped, that is, the bitmap
+page containing the descriptor bits for the file page; the bitmap page
+is x-latched
+@param space	in: space id of the file page
+@param page_no	in: page number of the file page
+@param zip_size	in: compressed page size in bytes; 0 for uncompressed pages
+@param mtr	in: mini-transaction */
+#define ibuf_bitmap_get_map_page(space, page_no, zip_size, mtr)		\
+	ibuf_bitmap_get_map_page_func(space, page_no, zip_size,		\
+				      __FILE__, __LINE__, mtr)
+
 /************************************************************************//**
 Sets the free bits of the page in the ibuf bitmap. This is done in a separate
 mini-transaction, hence this operation does not restrict further work to only
diff --git a/storage/innodb_plugin/include/lock0lock.h b/storage/innodb_plugin/include/lock0lock.h
index 82e4c9bd976..7d76cbe3c75 100644
--- a/storage/innodb_plugin/include/lock0lock.h
+++ b/storage/innodb_plugin/include/lock0lock.h
@@ -613,13 +613,16 @@ lock_rec_print(
 	FILE*		file,	/*!< in: file where to print */
 	const lock_t*	lock);	/*!< in: record type lock */
 /*********************************************************************//**
-Prints info of locks for all transactions. */
+Prints info of locks for all transactions.
+@return FALSE if not able to obtain kernel mutex
+and exits without printing info */
 UNIV_INTERN
-void
+ibool
 lock_print_info_summary(
 /*====================*/
-	FILE*	file);	/*!< in: file where to print */
-/*********************************************************************//**
+	FILE*	file,	/*!< in: file where to print */
+	ibool   nowait);/*!< in: whether to wait for the kernel mutex */
+/*************************************************************************
 Prints info of locks for each transaction. */
 UNIV_INTERN
 void
diff --git a/storage/innodb_plugin/include/srv0srv.h b/storage/innodb_plugin/include/srv0srv.h
index 228c9f6600a..9001d0988cf 100644
--- a/storage/innodb_plugin/include/srv0srv.h
+++ b/storage/innodb_plugin/include/srv0srv.h
@@ -227,7 +227,8 @@ extern ibool	srv_print_innodb_tablespace_monitor;
 extern ibool	srv_print_verbose_log;
 extern ibool	srv_print_innodb_table_monitor;
 
-extern ibool	srv_lock_timeout_and_monitor_active;
+extern ibool	srv_lock_timeout_active;
+extern ibool	srv_monitor_active;
 extern ibool	srv_error_monitor_active;
 
 extern ulong	srv_n_spin_wait_rounds;
@@ -540,15 +541,23 @@ srv_release_mysql_thread_if_suspended(
 				MySQL OS thread	 */
 /*********************************************************************//**
 A thread which wakes up threads whose lock wait may have lasted too long.
-This also prints the info output by various InnoDB monitors.
 @return	a dummy parameter */
 UNIV_INTERN
 os_thread_ret_t
-srv_lock_timeout_and_monitor_thread(
-/*================================*/
+srv_lock_timeout_thread(
+/*====================*/
 	void*	arg);	/*!< in: a dummy parameter required by
 			os_thread_create */
 /*********************************************************************//**
+A thread which prints the info output by various InnoDB monitors.
+@return	a dummy parameter */
+UNIV_INTERN
+os_thread_ret_t
+srv_monitor_thread(
+/*===============*/
+	void*	arg);	/*!< in: a dummy parameter required by
+			os_thread_create */
+/*************************************************************************
 A thread which prints warnings about semaphore waits which have lasted
 too long. These can be used to track bugs which cause hangs.
 @return	a dummy parameter */
@@ -559,12 +568,15 @@ srv_error_monitor_thread(
 	void*	arg);	/*!< in: a dummy parameter required by
 			os_thread_create */
 /******************************************************************//**
-Outputs to a file the output of the InnoDB Monitor. */
+Outputs to a file the output of the InnoDB Monitor.
+@return FALSE if not all information printed
+due to failure to obtain necessary mutex */
 UNIV_INTERN
-void
+ibool
 srv_printf_innodb_monitor(
 /*======================*/
 	FILE*	file,		/*!< in: output stream */
+	ibool	nowait,		/*!< in: whether to wait for kernel mutex */
 	ulint*	trx_start,	/*!< out: file position of the start of
 				the list of active transactions */
 	ulint*	trx_end);	/*!< out: file position of the end of
diff --git a/storage/innodb_plugin/lock/lock0lock.c b/storage/innodb_plugin/lock/lock0lock.c
index 8dfeb064863..3db7c2b36bc 100644
--- a/storage/innodb_plugin/lock/lock0lock.c
+++ b/storage/innodb_plugin/lock/lock0lock.c
@@ -4317,14 +4317,26 @@ lock_get_n_rec_locks(void)
 #endif /* PRINT_NUM_OF_LOCK_STRUCTS */
 
 /*********************************************************************//**
-Prints info of locks for all transactions. */
+Prints info of locks for all transactions.
+@return FALSE if not able to obtain kernel mutex
+and exits without printing info */
 UNIV_INTERN
-void
+ibool
 lock_print_info_summary(
 /*====================*/
-	FILE*	file)	/*!< in: file where to print */
+	FILE*	file,	/*!< in: file where to print */
+	ibool   nowait)	/*!< in: whether to wait for the kernel mutex */
 {
-	lock_mutex_enter_kernel();
+	/* if nowait is FALSE, wait on the kernel mutex,
+	otherwise return immediately if fail to obtain the
+	mutex. */
+	if (!nowait) {
+		lock_mutex_enter_kernel();
+	} else if (mutex_enter_nowait(&kernel_mutex)) {
+		fputs("FAIL TO OBTAIN KERNEL MUTEX, "
+		      "SKIP LOCK INFO PRINTING\n", file);
+		return(FALSE);
+	}
 
 	if (lock_deadlock_found) {
 		fputs("------------------------\n"
@@ -4356,6 +4368,7 @@ lock_print_info_summary(
 		"Total number of lock structs in row lock hash table %lu\n",
 		(ulong) lock_get_n_rec_locks());
 #endif /* PRINT_NUM_OF_LOCK_STRUCTS */
+	return(TRUE);
 }
 
 /*********************************************************************//**
diff --git a/storage/innodb_plugin/log/log0log.c b/storage/innodb_plugin/log/log0log.c
index d5b696074b3..86c9f9b4130 100644
--- a/storage/innodb_plugin/log/log0log.c
+++ b/storage/innodb_plugin/log/log0log.c
@@ -3095,7 +3095,7 @@ loop:
 
 	if (srv_fast_shutdown < 2
 	   && (srv_error_monitor_active
-	      || srv_lock_timeout_and_monitor_active)) {
+	      || srv_lock_timeout_active || srv_monitor_active)) {
 
 		mutex_exit(&kernel_mutex);
 
diff --git a/storage/innodb_plugin/row/row0sel.c b/storage/innodb_plugin/row/row0sel.c
index 5f66b9be159..23cd97f6826 100644
--- a/storage/innodb_plugin/row/row0sel.c
+++ b/storage/innodb_plugin/row/row0sel.c
@@ -4612,6 +4612,7 @@ row_search_autoinc_read_column(
 	dict_index_t*	index,		/*!< in: index to read from */
 	const rec_t*	rec,		/*!< in: current rec */
 	ulint		col_no,		/*!< in: column number */
+	ulint		mtype,		/*!< in: column main type */
 	ibool		unsigned_type)	/*!< in: signed or unsigned flag */
 {
 	ulint		len;
@@ -4628,10 +4629,26 @@ row_search_autoinc_read_column(
 	data = rec_get_nth_field(rec, offsets, col_no, &len);
 
 	ut_a(len != UNIV_SQL_NULL);
-	ut_a(len <= sizeof value);
 
-	/* we assume AUTOINC value cannot be negative */
-	value = mach_read_int_type(data, len, unsigned_type);
+	switch (mtype) {
+	case DATA_INT:
+		ut_a(len <= sizeof value);
+		value = mach_read_int_type(data, len, unsigned_type);
+		break;
+
+	case DATA_FLOAT:
+		ut_a(len == sizeof(float));
+		value = mach_float_read(data);
+		break;
+
+	case DATA_DOUBLE:
+		ut_a(len == sizeof(double));
+		value = mach_double_read(data);
+		break;
+
+	default:
+		ut_error;
+	}
 
 	if (UNIV_LIKELY_NULL(heap)) {
 		mem_heap_free(heap);
@@ -4717,7 +4734,8 @@ row_search_max_autoinc(
 					dfield->col->prtype & DATA_UNSIGNED);
 
 				*value = row_search_autoinc_read_column(
-					index, rec, i, unsigned_type);
+					index, rec, i,
+					dfield->col->mtype, unsigned_type);
 			}
 		}
 
diff --git a/storage/innodb_plugin/srv/srv0srv.c b/storage/innodb_plugin/srv/srv0srv.c
index 639da1ed2f3..7252929d3e2 100644
--- a/storage/innodb_plugin/srv/srv0srv.c
+++ b/storage/innodb_plugin/srv/srv0srv.c
@@ -119,7 +119,8 @@ UNIV_INTERN ulint	srv_fatal_semaphore_wait_threshold = 600;
 in microseconds, in order to reduce the lagging of the purge thread. */
 UNIV_INTERN ulint	srv_dml_needed_delay = 0;
 
-UNIV_INTERN ibool	srv_lock_timeout_and_monitor_active = FALSE;
+UNIV_INTERN ibool	srv_lock_timeout_active = FALSE;
+UNIV_INTERN ibool	srv_monitor_active = FALSE;
 UNIV_INTERN ibool	srv_error_monitor_active = FALSE;
 
 UNIV_INTERN const char*	srv_main_thread_op_info = "";
@@ -188,7 +189,17 @@ UNIV_INTERN ulong	srv_flush_log_at_trx_commit = 1;
 the checkpoints. */
 UNIV_INTERN char	srv_adaptive_flushing	= TRUE;
 
-/* The sort order table of the MySQL latin1_swedish_ci character set
+/** Maximum number of times allowed to conditionally acquire
+mutex before switching to blocking wait on the mutex */
+#define MAX_MUTEX_NOWAIT	20
+
+/** Check whether the number of failed nonblocking mutex
+acquisition attempts exceeds maximum allowed value. If so,
+srv_printf_innodb_monitor() will request mutex acquisition
+with mutex_enter(), which will wait until it gets the mutex. */
+#define MUTEX_NOWAIT(mutex_skipped)	((mutex_skipped) < MAX_MUTEX_NOWAIT)
+
+/** The sort order table of the MySQL latin1_swedish_ci character set
 collation */
 UNIV_INTERN const byte*	srv_latin1_ordering;
 
@@ -1683,12 +1694,15 @@ srv_refresh_innodb_monitor_stats(void)
 }
 
 /******************************************************************//**
-Outputs to a file the output of the InnoDB Monitor. */
+Outputs to a file the output of the InnoDB Monitor.
+@return FALSE if not all information printed
+due to failure to obtain necessary mutex */
 UNIV_INTERN
-void
+ibool
 srv_printf_innodb_monitor(
 /*======================*/
 	FILE*	file,		/*!< in: output stream */
+	ibool	nowait,		/*!< in: whether to wait for kernel mutex */
 	ulint*	trx_start,	/*!< out: file position of the start of
 				the list of active transactions */
 	ulint*	trx_end)	/*!< out: file position of the end of
@@ -1697,6 +1711,7 @@ srv_printf_innodb_monitor(
 	double	time_elapsed;
 	time_t	current_time;
 	ulint	n_reserved;
+	ibool	ret;
 
 	mutex_enter(&srv_innodb_monitor_mutex);
 
@@ -1746,24 +1761,31 @@ srv_printf_innodb_monitor(
 
 	mutex_exit(&dict_foreign_err_mutex);
 
-	lock_print_info_summary(file);
-	if (trx_start) {
-		long	t = ftell(file);
-		if (t < 0) {
-			*trx_start = ULINT_UNDEFINED;
-		} else {
-			*trx_start = (ulint) t;
-		}
-	}
-	lock_print_info_all_transactions(file);
-	if (trx_end) {
-		long	t = ftell(file);
-		if (t < 0) {
-			*trx_end = ULINT_UNDEFINED;
-		} else {
-			*trx_end = (ulint) t;
+	/* Only if lock_print_info_summary proceeds correctly,
+	before we call the lock_print_info_all_transactions
+	to print all the lock information. */
+	ret = lock_print_info_summary(file, nowait);
+
+	if (ret) {
+		if (trx_start) {
+			long	t = ftell(file);
+			if (t < 0) {
+				*trx_start = ULINT_UNDEFINED;
+			} else {
+				*trx_start = (ulint) t;
+			}
+		}
+		lock_print_info_all_transactions(file);
+		if (trx_end) {
+			long	t = ftell(file);
+			if (t < 0) {
+				*trx_end = ULINT_UNDEFINED;
+			} else {
+				*trx_end = (ulint) t;
+			}
 		}
 	}
+
 	fputs("--------\n"
 	      "FILE I/O\n"
 	      "--------\n", file);
@@ -1861,6 +1883,8 @@ srv_printf_innodb_monitor(
 	      "============================\n", file);
 	mutex_exit(&srv_innodb_monitor_mutex);
 	fflush(file);
+
+	return(ret);
 }
 
 /******************************************************************//**
@@ -1948,26 +1972,23 @@ srv_export_innodb_status(void)
 }
 
 /*********************************************************************//**
-A thread which wakes up threads whose lock wait may have lasted too long.
-This also prints the info output by various InnoDB monitors.
+A thread which prints the info output by various InnoDB monitors.
 @return	a dummy parameter */
 UNIV_INTERN
 os_thread_ret_t
-srv_lock_timeout_and_monitor_thread(
-/*================================*/
+srv_monitor_thread(
+/*===============*/
 	void*	arg __attribute__((unused)))
 			/*!< in: a dummy parameter required by
 			os_thread_create */
 {
-	srv_slot_t*	slot;
 	double		time_elapsed;
 	time_t		current_time;
 	time_t		last_table_monitor_time;
 	time_t		last_tablespace_monitor_time;
 	time_t		last_monitor_time;
-	ibool		some_waits;
-	double		wait_time;
-	ulint		i;
+	ulint		mutex_skipped;
+	ibool		last_srv_print_monitor;
 
 #ifdef UNIV_DEBUG_THREAD_CREATION
 	fprintf(stderr, "Lock timeout thread starts, id %lu\n",
@@ -1978,13 +1999,15 @@ srv_lock_timeout_and_monitor_thread(
 	last_table_monitor_time = time(NULL);
 	last_tablespace_monitor_time = time(NULL);
 	last_monitor_time = time(NULL);
+	mutex_skipped = 0;
+	last_srv_print_monitor = srv_print_innodb_monitor;
 loop:
-	srv_lock_timeout_and_monitor_active = TRUE;
+	srv_monitor_active = TRUE;
 
-	/* When someone is waiting for a lock, we wake up every second
-	and check if a timeout has passed for a lock wait */
+	/* Wake up every 5 seconds to see if we need to print
+	monitor information. */
 
-	os_thread_sleep(1000000);
+	os_thread_sleep(5000000);
 
 	current_time = time(NULL);
 
@@ -1994,14 +2017,40 @@ loop:
 		last_monitor_time = time(NULL);
 
 		if (srv_print_innodb_monitor) {
-			srv_printf_innodb_monitor(stderr, NULL, NULL);
+			/* Reset mutex_skipped counter everytime
+			srv_print_innodb_monitor changes. This is to
+			ensure we will not be blocked by kernel_mutex
+			for short duration information printing,
+			such as requested by sync_array_print_long_waits() */
+			if (!last_srv_print_monitor) {
+				mutex_skipped = 0;
+				last_srv_print_monitor = TRUE;
+			}
+
+			if (!srv_printf_innodb_monitor(stderr,
+						MUTEX_NOWAIT(mutex_skipped),
+						NULL, NULL)) {
+				mutex_skipped++;
+			} else {
+				/* Reset the counter */
+				mutex_skipped = 0;
+			}
+		} else {
+			last_srv_print_monitor = FALSE;
 		}
 
+
 		if (srv_innodb_status) {
 			mutex_enter(&srv_monitor_file_mutex);
 			rewind(srv_monitor_file);
-			srv_printf_innodb_monitor(srv_monitor_file, NULL,
-						  NULL);
+			if (!srv_printf_innodb_monitor(srv_monitor_file,
+						MUTEX_NOWAIT(mutex_skipped),
+						NULL, NULL)) {
+				mutex_skipped++;
+			} else {
+				mutex_skipped = 0;
+			}
+
 			os_file_set_eof(srv_monitor_file);
 			mutex_exit(&srv_monitor_file_mutex);
 		}
@@ -2054,6 +2103,56 @@ loop:
 		}
 	}
 
+	if (srv_shutdown_state >= SRV_SHUTDOWN_CLEANUP) {
+		goto exit_func;
+	}
+
+	if (srv_print_innodb_monitor
+	    || srv_print_innodb_lock_monitor
+	    || srv_print_innodb_tablespace_monitor
+	    || srv_print_innodb_table_monitor) {
+		goto loop;
+	}
+
+	srv_monitor_active = FALSE;
+
+	goto loop;
+
+exit_func:
+	srv_monitor_active = FALSE;
+
+	/* We count the number of threads in os_thread_exit(). A created
+	thread should always use that to exit and not use return() to exit. */
+
+	os_thread_exit(NULL);
+
+	OS_THREAD_DUMMY_RETURN;
+}
+
+/*********************************************************************//**
+A thread which wakes up threads whose lock wait may have lasted too long.
+@return	a dummy parameter */
+UNIV_INTERN
+os_thread_ret_t
+srv_lock_timeout_thread(
+/*====================*/
+	void*	arg __attribute__((unused)))
+			/* in: a dummy parameter required by
+			os_thread_create */
+{
+	srv_slot_t*	slot;
+	ibool		some_waits;
+	double		wait_time;
+	ulint		i;
+
+loop:
+	/* When someone is waiting for a lock, we wake up every second
+	and check if a timeout has passed for a lock wait */
+
+	os_thread_sleep(1000000);
+
+	srv_lock_timeout_active = TRUE;
+
 	mutex_enter(&kernel_mutex);
 
 	some_waits = FALSE;
@@ -2104,17 +2203,11 @@ loop:
 		goto exit_func;
 	}
 
-	if (some_waits || srv_print_innodb_monitor
-	    || srv_print_innodb_lock_monitor
-	    || srv_print_innodb_tablespace_monitor
-	    || srv_print_innodb_table_monitor) {
+	if (some_waits) {
 		goto loop;
 	}
 
-	/* No one was waiting for a lock and no monitor was active:
-	suspend this thread */
-
-	srv_lock_timeout_and_monitor_active = FALSE;
+	srv_lock_timeout_active = FALSE;
 
 #if 0
 	/* The following synchronisation is disabled, since
@@ -2124,7 +2217,7 @@ loop:
 	goto loop;
 
 exit_func:
-	srv_lock_timeout_and_monitor_active = FALSE;
+	srv_lock_timeout_active = FALSE;
 
 	/* We count the number of threads in os_thread_exit(). A created
 	thread should always use that to exit and not use return() to exit. */
diff --git a/storage/innodb_plugin/srv/srv0start.c b/storage/innodb_plugin/srv/srv0start.c
index a0505f43852..cab193615df 100644
--- a/storage/innodb_plugin/srv/srv0start.c
+++ b/storage/innodb_plugin/srv/srv0start.c
@@ -144,9 +144,9 @@ static mutex_t		ios_mutex;
 static ulint		ios;
 
 /** io_handler_thread parameters for thread identification */
-static ulint		n[SRV_MAX_N_IO_THREADS + 5];
+static ulint		n[SRV_MAX_N_IO_THREADS + 6];
 /** io_handler_thread identifiers */
-static os_thread_id_t	thread_ids[SRV_MAX_N_IO_THREADS + 5];
+static os_thread_id_t	thread_ids[SRV_MAX_N_IO_THREADS + 6];
 
 /** We use this mutex to test the return value of pthread_mutex_trylock
    on successful locking. HP-UX does NOT return 0, though Linux et al do. */
@@ -1671,15 +1671,18 @@ innobase_start_or_create_for_mysql(void)
 	/* fprintf(stderr, "Max allowed record size %lu\n",
 	page_get_free_space_of_empty() / 2); */
 
-	/* Create the thread which watches the timeouts for lock waits
-	and prints InnoDB monitor info */
-
-	os_thread_create(&srv_lock_timeout_and_monitor_thread, NULL,
+	/* Create the thread which watches the timeouts for lock waits */
+	os_thread_create(&srv_lock_timeout_thread, NULL,
 			 thread_ids + 2 + SRV_MAX_N_IO_THREADS);
 
 	/* Create the thread which warns of long semaphore waits */
 	os_thread_create(&srv_error_monitor_thread, NULL,
 			 thread_ids + 3 + SRV_MAX_N_IO_THREADS);
+
+	/* Create the thread which prints InnoDB monitor info */
+	os_thread_create(&srv_monitor_thread, NULL,
+			 thread_ids + 4 + SRV_MAX_N_IO_THREADS);
+
 	srv_is_being_started = FALSE;
 
 	if (trx_doublewrite == NULL) {

From ba5d13fdfd94c6d77bec8af94314083504166b5b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:19:18 +0400
Subject: [PATCH 020/116] Applying InnoDB snapshot

Detailed revision comments:

r6367 | marko | 2009-12-28 15:39:19 +0200 (Mon, 28 Dec 2009) | 2 lines
branches/zip: dict_index_add_to_cache(): Always free the index object,
also when returning DB_CORRUPTION.
---
 storage/innodb_plugin/dict/dict0dict.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/storage/innodb_plugin/dict/dict0dict.c b/storage/innodb_plugin/dict/dict0dict.c
index 4c62e8de748..ca129c29d20 100644
--- a/storage/innodb_plugin/dict/dict0dict.c
+++ b/storage/innodb_plugin/dict/dict0dict.c
@@ -1460,6 +1460,7 @@ dict_index_add_to_cache(
 
 	if (!dict_index_find_cols(table, index)) {
 
+		dict_mem_index_free(index);
 		return(DB_CORRUPTION);
 	}
 

From 68afad078a8a8d2aa3c9d1e4298e670678b0dbc7 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:19:38 +0400
Subject: [PATCH 021/116] Applying InnoDB snapshot

Detailed revision comments:

r6425 | marko | 2010-01-12 13:47:11 +0200 (Tue, 12 Jan 2010) | 45 lines
branches/zip: Merge revisions 6350:6424 from branches/5.1:

  ------------------------------------------------------------------------
  r6421 | jyang | 2010-01-12 07:59:16 +0200 (Tue, 12 Jan 2010) | 8 lines
  Changed paths:
     M /branches/5.1/row/row0mysql.c

  branches/5.1: Fix bug #49238: Creating/Dropping a temporary table
  while at 1023 transactions will cause assert. Handle possible
  DB_TOO_MANY_CONCURRENT_TRXS when deleting metadata in
  row_drop_table_for_mysql().

  rb://220, approved by Marko
  ------------------------------------------------------------------------
  r6422 | marko | 2010-01-12 11:34:27 +0200 (Tue, 12 Jan 2010) | 3 lines
  Changed paths:
     M /branches/5.1/handler/ha_innodb.cc
     M /branches/5.1/handler/ha_innodb.h

  branches/5.1: Non-functional change:
  Make innobase_get_int_col_max_value() a static function.
  It does not access any fields of class ha_innobase.
  ------------------------------------------------------------------------
  r6424 | marko | 2010-01-12 12:22:19 +0200 (Tue, 12 Jan 2010) | 16 lines
  Changed paths:
     M /branches/5.1/handler/ha_innodb.cc
     M /branches/5.1/handler/ha_innodb.h

  branches/5.1: In innobase_initialize_autoinc(), do not attempt to read
  the maximum auto-increment value from the table if
  innodb_force_recovery is set to at least 4, so that writes are
  disabled. (Bug #46193)

  innobase_get_int_col_max_value(): Move the function definition before
  ha_innobase::innobase_initialize_autoinc(), because that function now
  calls this function.

  ha_innobase::innobase_initialize_autoinc(): Change the return type to
  void.  Do not attempt to read the maximum auto-increment value from
  the table if innodb_force_recovery is set to at least 4.  Issue
  ER_AUTOINC_READ_FAILED to the client when the auto-increment value
  cannot be read.

  rb://144 by Sunny, revised by Marko
  ------------------------------------------------------------------------
---
 storage/innodb_plugin/ChangeLog            |  15 +-
 storage/innodb_plugin/handler/ha_innodb.cc | 239 ++++++++++++---------
 storage/innodb_plugin/handler/ha_innodb.h  |   3 +-
 storage/innodb_plugin/row/row0mysql.c      |  32 ++-
 4 files changed, 171 insertions(+), 118 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index dddf37e2334..334c8b5a8cc 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,6 +1,19 @@
+2010-01-12	The InnoDB Team
+
+	* handler/ha_innodb.cc, handler/ha_innodb.h:
+	Fix Bug #46193 crash when accessing tables after enabling
+	innodb_force_recovery option
+
+2010-01-12	The InnoDB Team
+
+	* row/row0mysql.c:
+	Fix Bug#49238 Creating/Dropping a temporary table while at 1023
+	transactions will cause assert.
+
 2009-12-02	The InnoDB Team
 
-	* srv/srv0start.c: Display the zlib version number at startup.
+	* srv/srv0start.c:
+	Display the zlib version number at startup.
 	InnoDB compressed tables use zlib, and the implementation depends
 	on the zlib function compressBound(), whose definition was slightly
 	changed in zlib version 1.2.3.1 in 2006.  MySQL bundles zlib 1.2.3
diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 9c214bab3de..95b0d331076 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -3006,59 +3006,150 @@ normalize_table_name(
 }
 
 /********************************************************************//**
+Get the upper limit of the MySQL integral and floating-point type.
+@return maximum allowed value for the field */
+static
+ulonglong
+innobase_get_int_col_max_value(
+/*===========================*/
+	const Field*	field)	/*!< in: MySQL field */
+{
+	ulonglong	max_value = 0;
+
+	switch(field->key_type()) {
+	/* TINY */
+	case HA_KEYTYPE_BINARY:
+		max_value = 0xFFULL;
+		break;
+	case HA_KEYTYPE_INT8:
+		max_value = 0x7FULL;
+		break;
+	/* SHORT */
+	case HA_KEYTYPE_USHORT_INT:
+		max_value = 0xFFFFULL;
+		break;
+	case HA_KEYTYPE_SHORT_INT:
+		max_value = 0x7FFFULL;
+		break;
+	/* MEDIUM */
+	case HA_KEYTYPE_UINT24:
+		max_value = 0xFFFFFFULL;
+		break;
+	case HA_KEYTYPE_INT24:
+		max_value = 0x7FFFFFULL;
+		break;
+	/* LONG */
+	case HA_KEYTYPE_ULONG_INT:
+		max_value = 0xFFFFFFFFULL;
+		break;
+	case HA_KEYTYPE_LONG_INT:
+		max_value = 0x7FFFFFFFULL;
+		break;
+	/* BIG */
+	case HA_KEYTYPE_ULONGLONG:
+		max_value = 0xFFFFFFFFFFFFFFFFULL;
+		break;
+	case HA_KEYTYPE_LONGLONG:
+		max_value = 0x7FFFFFFFFFFFFFFFULL;
+		break;
+	case HA_KEYTYPE_FLOAT:
+		/* We use the maximum as per IEEE754-2008 standard, 2^24 */
+		max_value = 0x1000000ULL;
+		break;
+	case HA_KEYTYPE_DOUBLE:
+		/* We use the maximum as per IEEE754-2008 standard, 2^53 */
+		max_value = 0x20000000000000ULL;
+		break;
+	default:
+		ut_error;
+	}
+
+	return(max_value);
+}
+
+/************************************************************************
 Set the autoinc column max value. This should only be called once from
-ha_innobase::open(). Therefore there's no need for a covering lock.
-@return	DB_SUCCESS or error code */
+ha_innobase::open(). Therefore there's no need for a covering lock. */
 UNIV_INTERN
-ulint
+void
 ha_innobase::innobase_initialize_autoinc()
 /*======================================*/
 {
-	dict_index_t*	index;
 	ulonglong	auto_inc;
-	const char*	col_name;
-	ulint		error;
-
-	col_name = table->found_next_number_field->field_name;
-	index = innobase_get_index(table->s->next_number_index);
-
-	/* Execute SELECT MAX(col_name) FROM TABLE; */
-	error = row_search_max_autoinc(index, col_name, &auto_inc);
-
-	switch (error) {
-	case DB_SUCCESS:
-
-		/* At the this stage we don't know the increment
-		or the offset, so use default inrement of 1. */
-		++auto_inc;
-		break;
-
-	case DB_RECORD_NOT_FOUND:
-		ut_print_timestamp(stderr);
-		fprintf(stderr, "  InnoDB: MySQL and InnoDB data "
-			"dictionaries are out of sync.\n"
-			"InnoDB: Unable to find the AUTOINC column %s in the "
-			"InnoDB table %s.\n"
-			"InnoDB: We set the next AUTOINC column value to the "
-			"maximum possible value,\n"
-			"InnoDB: in effect disabling the AUTOINC next value "
-			"generation.\n"
-			"InnoDB: You can either set the next AUTOINC value "
-			"explicitly using ALTER TABLE\n"
-			"InnoDB: or fix the data dictionary by recreating "
-			"the table.\n",
-			col_name, index->table->name);
+	const Field*	field = table->found_next_number_field;
 
+	if (field != NULL) {
+		auto_inc = innobase_get_int_col_max_value(field);
+	} else {
+		/* We have no idea what's been passed in to us as the
+		autoinc column. We set it to the MAX_INT of our table
+		autoinc type. */
 		auto_inc = 0xFFFFFFFFFFFFFFFFULL;
-		break;
 
-	default:
-		return(error);
+		ut_print_timestamp(stderr);
+		fprintf(stderr, "  InnoDB: Unable to determine the AUTOINC "
+				"column name\n");
+	}
+
+	if (srv_force_recovery >= SRV_FORCE_NO_IBUF_MERGE) {
+		/* If the recovery level is set so high that writes
+		are disabled we force the AUTOINC counter to the MAX
+		value effectively disabling writes to the table.
+		Secondly, we avoid reading the table in case the read
+		results in failure due to a corrupted table/index.
+
+		We will not return an error to the client, so that the
+		tables can be dumped with minimal hassle.  If an error
+		were returned in this case, the first attempt to read
+		the table would fail and subsequent SELECTs would succeed. */
+	} else if (field == NULL) {
+		my_error(ER_AUTOINC_READ_FAILED, MYF(0));
+	} else {
+		dict_index_t*	index;
+		const char*	col_name;
+		ulonglong	read_auto_inc;
+		ulint		err;
+
+		update_thd(ha_thd());
+		col_name = field->field_name;
+		index = innobase_get_index(table->s->next_number_index);
+
+		/* Execute SELECT MAX(col_name) FROM TABLE; */
+		err = row_search_max_autoinc(index, col_name, &read_auto_inc);
+
+		switch (err) {
+		case DB_SUCCESS:
+			/* At the this stage we do not know the increment
+			or the offset, so use a default increment of 1. */
+			auto_inc = read_auto_inc + 1;
+			break;
+
+		case DB_RECORD_NOT_FOUND:
+			ut_print_timestamp(stderr);
+			fprintf(stderr, "  InnoDB: MySQL and InnoDB data "
+				"dictionaries are out of sync.\n"
+				"InnoDB: Unable to find the AUTOINC column "
+				"%s in the InnoDB table %s.\n"
+				"InnoDB: We set the next AUTOINC column "
+				"value to the maximum possible value,\n"
+				"InnoDB: in effect disabling the AUTOINC "
+				"next value generation.\n"
+				"InnoDB: You can either set the next "
+				"AUTOINC value explicitly using ALTER TABLE\n"
+				"InnoDB: or fix the data dictionary by "
+				"recreating the table.\n",
+				col_name, index->table->name);
+
+			my_error(ER_AUTOINC_READ_FAILED, MYF(0));
+			break;
+		default:
+			/* row_search_max_autoinc() should only return
+			one of DB_SUCCESS or DB_RECORD_NOT_FOUND. */
+			ut_error;
+		}
 	}
 
 	dict_table_autoinc_initialize(prebuilt->table, auto_inc);
-
-	return(DB_SUCCESS);
 }
 
 /*****************************************************************//**
@@ -3265,8 +3356,6 @@ retry:
 
 	/* Only if the table has an AUTOINC column. */
 	if (prebuilt->table != NULL && table->found_next_number_field != NULL) {
-		ulint	error;
-
 		dict_table_autoinc_lock(prebuilt->table);
 
 		/* Since a table can already be "open" in InnoDB's internal
@@ -3275,8 +3364,7 @@ retry:
 		autoinc value from a previous MySQL open. */
 		if (dict_table_autoinc_read(prebuilt->table) == 0) {
 
-			error = innobase_initialize_autoinc();
-			ut_a(error == DB_SUCCESS);
+			innobase_initialize_autoinc();
 		}
 
 		dict_table_autoinc_unlock(prebuilt->table);
@@ -4092,67 +4180,6 @@ skip_field:
 	}
 }
 
-/********************************************************************//**
-Get the upper limit of the MySQL integral and floating-point type. */
-UNIV_INTERN
-ulonglong
-ha_innobase::innobase_get_int_col_max_value(
-/*========================================*/
-	const Field*	field)
-{
-	ulonglong	max_value = 0;
-
-	switch(field->key_type()) {
-	/* TINY */
-	case HA_KEYTYPE_BINARY:
-		max_value = 0xFFULL;
-		break;
-	case HA_KEYTYPE_INT8:
-		max_value = 0x7FULL;
-		break;
-	/* SHORT */
-	case HA_KEYTYPE_USHORT_INT:
-		max_value = 0xFFFFULL;
-		break;
-	case HA_KEYTYPE_SHORT_INT:
-		max_value = 0x7FFFULL;
-		break;
-	/* MEDIUM */
-	case HA_KEYTYPE_UINT24:
-		max_value = 0xFFFFFFULL;
-		break;
-	case HA_KEYTYPE_INT24:
-		max_value = 0x7FFFFFULL;
-		break;
-	/* LONG */
-	case HA_KEYTYPE_ULONG_INT:
-		max_value = 0xFFFFFFFFULL;
-		break;
-	case HA_KEYTYPE_LONG_INT:
-		max_value = 0x7FFFFFFFULL;
-		break;
-	/* BIG */
-	case HA_KEYTYPE_ULONGLONG:
-		max_value = 0xFFFFFFFFFFFFFFFFULL;
-		break;
-	case HA_KEYTYPE_LONGLONG:
-		max_value = 0x7FFFFFFFFFFFFFFFULL;
-		break;
-	case HA_KEYTYPE_FLOAT:
-		/* We use the maximum as per IEEE754-2008 standard, 2^24 */
-		max_value = 0x1000000ULL;
-		break;
-	case HA_KEYTYPE_DOUBLE:
-		/* We use the maximum as per IEEE754-2008 standard, 2^53 */
-		max_value = 0x20000000000000ULL;
-		break;
-	default:
-		ut_error;
-	}
-
-	return(max_value);
-}
-
 /********************************************************************//**
 This special handling is really to overcome the limitations of MySQL's
 binlogging. We need to eliminate the non-determinism that will arise in
diff --git a/storage/innodb_plugin/handler/ha_innodb.h b/storage/innodb_plugin/handler/ha_innodb.h
index 31e88ed8530..0e366a1eb2c 100644
--- a/storage/innodb_plugin/handler/ha_innodb.h
+++ b/storage/innodb_plugin/handler/ha_innodb.h
@@ -91,9 +91,8 @@ class ha_innobase: public handler
 	ulint innobase_reset_autoinc(ulonglong auto_inc);
 	ulint innobase_get_autoinc(ulonglong* value);
 	ulint innobase_update_autoinc(ulonglong	auto_inc);
-	ulint innobase_initialize_autoinc();
+	void innobase_initialize_autoinc();
 	dict_index_t* innobase_get_index(uint keynr);
- 	ulonglong innobase_get_int_col_max_value(const Field* field);
 
 	/* Init values for the class: */
  public:
diff --git a/storage/innodb_plugin/row/row0mysql.c b/storage/innodb_plugin/row/row0mysql.c
index 7a43d0f3b92..e9fd12e9747 100644
--- a/storage/innodb_plugin/row/row0mysql.c
+++ b/storage/innodb_plugin/row/row0mysql.c
@@ -3255,19 +3255,13 @@ check_next_foreign:
 			   "END;\n"
 			   , FALSE, trx);
 
-	if (err != DB_SUCCESS) {
-		ut_a(err == DB_OUT_OF_FILE_SPACE);
-
-		err = DB_MUST_GET_MORE_FILE_SPACE;
-
-		row_mysql_handle_errors(&err, trx, NULL, NULL);
-
-		ut_error;
-	} else {
+	switch (err) {
 		ibool		is_temp;
 		const char*	name_or_path;
 		mem_heap_t*	heap;
 
+	case DB_SUCCESS:
+
 		heap = mem_heap_create(200);
 
 		/* Clone the name, in case it has been allocated
@@ -3333,7 +3327,27 @@ check_next_foreign:
 		}
 
 		mem_heap_free(heap);
+		break;
+
+	case DB_TOO_MANY_CONCURRENT_TRXS:
+		/* Cannot even find a free slot for the
+		the undo log. We can directly exit here
+		and return the DB_TOO_MANY_CONCURRENT_TRXS
+		error. */
+		break;
+
+	case DB_OUT_OF_FILE_SPACE:
+		err = DB_MUST_GET_MORE_FILE_SPACE;
+
+		row_mysql_handle_errors(&err, trx, NULL, NULL);
+
+		/* Fall through to raise error */
+
+	default:
+		/* No other possible error returns */
+		ut_error;
 	}
+
 funct_exit:
 
 	if (locked_dictionary) {

From e64b2cdd2343c4de7f0f44419a9b3f93e5ecb667 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:20:10 +0400
Subject: [PATCH 022/116] Applying InnoDB snapshot

Detailed revision comments:

r6426 | marko | 2010-01-12 15:36:14 +0200 (Tue, 12 Jan 2010) | 2 lines
branches/zip: row_sel_sec_rec_is_for_clust_rec(): Document the return value
more accurately.
---
 storage/innodb_plugin/row/row0sel.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/row/row0sel.c b/storage/innodb_plugin/row/row0sel.c
index 23cd97f6826..aacd96ea65e 100644
--- a/storage/innodb_plugin/row/row0sel.c
+++ b/storage/innodb_plugin/row/row0sel.c
@@ -132,7 +132,8 @@ index record.
 NOTE: the comparison is NOT done as a binary comparison, but character
 fields are compared with collation!
 @return TRUE if the secondary record is equal to the corresponding
-fields in the clustered record, when compared with collation */
+fields in the clustered record, when compared with collation;
+FALSE if not equal or if the clustered record has been marked for deletion */
 static
 ibool
 row_sel_sec_rec_is_for_clust_rec(

From 2b39ae46a91be7f6aa7b80e36c6cd2271366f599 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:20:36 +0400
Subject: [PATCH 023/116] Applying InnoDB snapshot

Detailed revision comments:

r6433 | marko | 2010-01-13 13:19:00 +0200 (Wed, 13 Jan 2010) | 2 lines
branches/zip: dict_sys_tables_get_flags(), dict_create_sys_*_tuple():
Add some const qualifiers and comments.
---
 storage/innodb_plugin/dict/dict0crea.c | 102 ++++++++++++++-----------
 storage/innodb_plugin/dict/dict0load.c |   2 +-
 2 files changed, 57 insertions(+), 47 deletions(-)

diff --git a/storage/innodb_plugin/dict/dict0crea.c b/storage/innodb_plugin/dict/dict0crea.c
index b0341e5eeab..5bbc2d17ddd 100644
--- a/storage/innodb_plugin/dict/dict0crea.c
+++ b/storage/innodb_plugin/dict/dict0crea.c
@@ -51,16 +51,18 @@ static
 dtuple_t*
 dict_create_sys_tables_tuple(
 /*=========================*/
-	dict_table_t*	table,	/*!< in: table */
-	mem_heap_t*	heap)	/*!< in: memory heap from which the memory for
-				the built tuple is allocated */
+	const dict_table_t*	table,	/*!< in: table */
+	mem_heap_t*		heap)	/*!< in: memory heap from
+					which the memory for the built
+					tuple is allocated */
 {
 	dict_table_t*	sys_tables;
 	dtuple_t*	entry;
 	dfield_t*	dfield;
 	byte*		ptr;
 
-	ut_ad(table && heap);
+	ut_ad(table);
+	ut_ad(heap);
 
 	sys_tables = dict_sys->sys_tables;
 
@@ -69,18 +71,18 @@ dict_create_sys_tables_tuple(
 	dict_table_copy_types(entry, sys_tables);
 
 	/* 0: NAME -----------------------------*/
-	dfield = dtuple_get_nth_field(entry, 0);
+	dfield = dtuple_get_nth_field(entry, 0/*NAME*/);
 
 	dfield_set_data(dfield, table->name, ut_strlen(table->name));
 	/* 3: ID -------------------------------*/
-	dfield = dtuple_get_nth_field(entry, 1);
+	dfield = dtuple_get_nth_field(entry, 1/*ID*/);
 
 	ptr = mem_heap_alloc(heap, 8);
 	mach_write_to_8(ptr, table->id);
 
 	dfield_set_data(dfield, ptr, 8);
 	/* 4: N_COLS ---------------------------*/
-	dfield = dtuple_get_nth_field(entry, 2);
+	dfield = dtuple_get_nth_field(entry, 2/*N_COLS*/);
 
 #if DICT_TF_COMPACT != 1
 #error
@@ -91,7 +93,7 @@ dict_create_sys_tables_tuple(
 			| ((table->flags & DICT_TF_COMPACT) << 31));
 	dfield_set_data(dfield, ptr, 4);
 	/* 5: TYPE -----------------------------*/
-	dfield = dtuple_get_nth_field(entry, 3);
+	dfield = dtuple_get_nth_field(entry, 3/*TYPE*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	if (table->flags & (~DICT_TF_COMPACT & ~(~0 << DICT_TF_BITS))) {
@@ -107,25 +109,25 @@ dict_create_sys_tables_tuple(
 
 	dfield_set_data(dfield, ptr, 4);
 	/* 6: MIX_ID (obsolete) ---------------------------*/
-	dfield = dtuple_get_nth_field(entry, 4);
+	dfield = dtuple_get_nth_field(entry, 4/*MIX_ID*/);
 
 	ptr = mem_heap_zalloc(heap, 8);
 
 	dfield_set_data(dfield, ptr, 8);
 	/* 7: MIX_LEN (additional flags) --------------------------*/
 
-	dfield = dtuple_get_nth_field(entry, 5);
+	dfield = dtuple_get_nth_field(entry, 5/*MIX_LEN*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, table->flags >> DICT_TF2_SHIFT);
 
 	dfield_set_data(dfield, ptr, 4);
 	/* 8: CLUSTER_NAME ---------------------*/
-	dfield = dtuple_get_nth_field(entry, 6);
+	dfield = dtuple_get_nth_field(entry, 6/*CLUSTER_NAME*/);
 	dfield_set_null(dfield); /* not supported */
 
 	/* 9: SPACE ----------------------------*/
-	dfield = dtuple_get_nth_field(entry, 7);
+	dfield = dtuple_get_nth_field(entry, 7/*SPACE*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, table->space);
@@ -144,19 +146,21 @@ static
 dtuple_t*
 dict_create_sys_columns_tuple(
 /*==========================*/
-	dict_table_t*	table,	/*!< in: table */
-	ulint		i,	/*!< in: column number */
-	mem_heap_t*	heap)	/*!< in: memory heap from which the memory for
-				the built tuple is allocated */
+	const dict_table_t*	table,	/*!< in: table */
+	ulint			i,	/*!< in: column number */
+	mem_heap_t*		heap)	/*!< in: memory heap from
+					which the memory for the built
+					tuple is allocated */
 {
 	dict_table_t*		sys_columns;
 	dtuple_t*		entry;
 	const dict_col_t*	column;
 	dfield_t*		dfield;
 	byte*			ptr;
-	const char*	col_name;
+	const char*		col_name;
 
-	ut_ad(table && heap);
+	ut_ad(table);
+	ut_ad(heap);
 
 	column = dict_table_get_nth_col(table, i);
 
@@ -167,47 +171,47 @@ dict_create_sys_columns_tuple(
 	dict_table_copy_types(entry, sys_columns);
 
 	/* 0: TABLE_ID -----------------------*/
-	dfield = dtuple_get_nth_field(entry, 0);
+	dfield = dtuple_get_nth_field(entry, 0/*TABLE_ID*/);
 
 	ptr = mem_heap_alloc(heap, 8);
 	mach_write_to_8(ptr, table->id);
 
 	dfield_set_data(dfield, ptr, 8);
 	/* 1: POS ----------------------------*/
-	dfield = dtuple_get_nth_field(entry, 1);
+	dfield = dtuple_get_nth_field(entry, 1/*POS*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, i);
 
 	dfield_set_data(dfield, ptr, 4);
 	/* 4: NAME ---------------------------*/
-	dfield = dtuple_get_nth_field(entry, 2);
+	dfield = dtuple_get_nth_field(entry, 2/*NAME*/);
 
 	col_name = dict_table_get_col_name(table, i);
 	dfield_set_data(dfield, col_name, ut_strlen(col_name));
 	/* 5: MTYPE --------------------------*/
-	dfield = dtuple_get_nth_field(entry, 3);
+	dfield = dtuple_get_nth_field(entry, 3/*MTYPE*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, column->mtype);
 
 	dfield_set_data(dfield, ptr, 4);
 	/* 6: PRTYPE -------------------------*/
-	dfield = dtuple_get_nth_field(entry, 4);
+	dfield = dtuple_get_nth_field(entry, 4/*PRTYPE*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, column->prtype);
 
 	dfield_set_data(dfield, ptr, 4);
 	/* 7: LEN ----------------------------*/
-	dfield = dtuple_get_nth_field(entry, 5);
+	dfield = dtuple_get_nth_field(entry, 5/*LEN*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, column->len);
 
 	dfield_set_data(dfield, ptr, 4);
 	/* 8: PREC ---------------------------*/
-	dfield = dtuple_get_nth_field(entry, 6);
+	dfield = dtuple_get_nth_field(entry, 6/*PREC*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, 0/* unused */);
@@ -325,9 +329,10 @@ static
 dtuple_t*
 dict_create_sys_indexes_tuple(
 /*==========================*/
-	dict_index_t*	index,	/*!< in: index */
-	mem_heap_t*	heap)	/*!< in: memory heap from which the memory for
-				the built tuple is allocated */
+	const dict_index_t*	index,	/*!< in: index */
+	mem_heap_t*		heap)	/*!< in: memory heap from
+					which the memory for the built
+					tuple is allocated */
 {
 	dict_table_t*	sys_indexes;
 	dict_table_t*	table;
@@ -336,7 +341,8 @@ dict_create_sys_indexes_tuple(
 	byte*		ptr;
 
 	ut_ad(mutex_own(&(dict_sys->mutex)));
-	ut_ad(index && heap);
+	ut_ad(index);
+	ut_ad(heap);
 
 	sys_indexes = dict_sys->sys_indexes;
 
@@ -347,32 +353,32 @@ dict_create_sys_indexes_tuple(
 	dict_table_copy_types(entry, sys_indexes);
 
 	/* 0: TABLE_ID -----------------------*/
-	dfield = dtuple_get_nth_field(entry, 0);
+	dfield = dtuple_get_nth_field(entry, 0/*TABLE_ID*/);
 
 	ptr = mem_heap_alloc(heap, 8);
 	mach_write_to_8(ptr, table->id);
 
 	dfield_set_data(dfield, ptr, 8);
 	/* 1: ID ----------------------------*/
-	dfield = dtuple_get_nth_field(entry, 1);
+	dfield = dtuple_get_nth_field(entry, 1/*ID*/);
 
 	ptr = mem_heap_alloc(heap, 8);
 	mach_write_to_8(ptr, index->id);
 
 	dfield_set_data(dfield, ptr, 8);
 	/* 4: NAME --------------------------*/
-	dfield = dtuple_get_nth_field(entry, 2);
+	dfield = dtuple_get_nth_field(entry, 2/*NAME*/);
 
 	dfield_set_data(dfield, index->name, ut_strlen(index->name));
 	/* 5: N_FIELDS ----------------------*/
-	dfield = dtuple_get_nth_field(entry, 3);
+	dfield = dtuple_get_nth_field(entry, 3/*N_FIELDS*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, index->n_fields);
 
 	dfield_set_data(dfield, ptr, 4);
 	/* 6: TYPE --------------------------*/
-	dfield = dtuple_get_nth_field(entry, 4);
+	dfield = dtuple_get_nth_field(entry, 4/*TYPE*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, index->type);
@@ -384,7 +390,7 @@ dict_create_sys_indexes_tuple(
 #error "DICT_SYS_INDEXES_SPACE_NO_FIELD != 7"
 #endif
 
-	dfield = dtuple_get_nth_field(entry, 5);
+	dfield = dtuple_get_nth_field(entry, 5/*SPACE*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, index->space);
@@ -396,7 +402,7 @@ dict_create_sys_indexes_tuple(
 #error "DICT_SYS_INDEXES_PAGE_NO_FIELD != 8"
 #endif
 
-	dfield = dtuple_get_nth_field(entry, 6);
+	dfield = dtuple_get_nth_field(entry, 6/*PAGE_NO*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 	mach_write_to_4(ptr, FIL_NULL);
@@ -415,10 +421,11 @@ static
 dtuple_t*
 dict_create_sys_fields_tuple(
 /*=========================*/
-	dict_index_t*	index,	/*!< in: index */
-	ulint		i,	/*!< in: field number */
-	mem_heap_t*	heap)	/*!< in: memory heap from which the memory for
-				the built tuple is allocated */
+	const dict_index_t*	index,	/*!< in: index */
+	ulint			i,	/*!< in: field number */
+	mem_heap_t*		heap)	/*!< in: memory heap from
+					which the memory for the built
+					tuple is allocated */
 {
 	dict_table_t*	sys_fields;
 	dtuple_t*	entry;
@@ -428,7 +435,8 @@ dict_create_sys_fields_tuple(
 	ibool		index_contains_column_prefix_field	= FALSE;
 	ulint		j;
 
-	ut_ad(index && heap);
+	ut_ad(index);
+	ut_ad(heap);
 
 	for (j = 0; j < index->n_fields; j++) {
 		if (dict_index_get_nth_field(index, j)->prefix_len > 0) {
@@ -446,7 +454,7 @@ dict_create_sys_fields_tuple(
 	dict_table_copy_types(entry, sys_fields);
 
 	/* 0: INDEX_ID -----------------------*/
-	dfield = dtuple_get_nth_field(entry, 0);
+	dfield = dtuple_get_nth_field(entry, 0/*INDEX_ID*/);
 
 	ptr = mem_heap_alloc(heap, 8);
 	mach_write_to_8(ptr, index->id);
@@ -454,7 +462,7 @@ dict_create_sys_fields_tuple(
 	dfield_set_data(dfield, ptr, 8);
 	/* 1: POS + PREFIX LENGTH ----------------------------*/
 
-	dfield = dtuple_get_nth_field(entry, 1);
+	dfield = dtuple_get_nth_field(entry, 1/*POS*/);
 
 	ptr = mem_heap_alloc(heap, 4);
 
@@ -474,7 +482,7 @@ dict_create_sys_fields_tuple(
 
 	dfield_set_data(dfield, ptr, 4);
 	/* 4: COL_NAME -------------------------*/
-	dfield = dtuple_get_nth_field(entry, 2);
+	dfield = dtuple_get_nth_field(entry, 2/*COL_NAME*/);
 
 	dfield_set_data(dfield, field->name,
 			ut_strlen(field->name));
@@ -605,6 +613,7 @@ dict_create_index_tree_step(
 	dict_table_t*	sys_indexes;
 	dict_table_t*	table;
 	dtuple_t*	search_tuple;
+	ulint		zip_size;
 	btr_pcur_t	pcur;
 	mtr_t		mtr;
 
@@ -629,8 +638,9 @@ dict_create_index_tree_step(
 
 	btr_pcur_move_to_next_user_rec(&pcur, &mtr);
 
-	node->page_no = btr_create(index->type, index->space,
-				   dict_table_zip_size(index->table),
+	zip_size = dict_table_zip_size(index->table);
+
+	node->page_no = btr_create(index->type, index->space, zip_size,
 				   index->id, index, &mtr);
 	/* printf("Created a new index tree in space %lu root page %lu\n",
 	index->space, index->page_no); */
diff --git a/storage/innodb_plugin/dict/dict0load.c b/storage/innodb_plugin/dict/dict0load.c
index 2867125e39d..0c72a2e8f81 100644
--- a/storage/innodb_plugin/dict/dict0load.c
+++ b/storage/innodb_plugin/dict/dict0load.c
@@ -260,7 +260,7 @@ dict_sys_tables_get_flags(
 		return(0);
 	}
 
-	field = rec_get_nth_field_old(rec, 4, &len);
+	field = rec_get_nth_field_old(rec, 4/*N_COLS*/, &len);
 	n_cols = mach_read_from_4(field);
 
 	if (UNIV_UNLIKELY(!(n_cols & 0x80000000UL))) {

From 0d3662f5984d0846d2fc553a51384bf4e8d6cd78 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:20:54 +0400
Subject: [PATCH 024/116] Applying InnoDB snapshot

Detailed revision comments:

r6445 | marko | 2010-01-13 17:15:29 +0200 (Wed, 13 Jan 2010) | 3 lines
branches/zip: buf_pool_drop_hash_index(): Check block->page.state
before checking block->is_hashed, because the latter may be uninitialized
right after server startup.
---
 storage/innodb_plugin/ChangeLog     | 7 +++++++
 storage/innodb_plugin/buf/buf0buf.c | 4 +++-
 2 files changed, 10 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 334c8b5a8cc..d5748c825b6 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,10 @@
+2010-01-13	The InnoDB Team
+
+	* buf/buf0buf.c:
+	When disabling the adaptive hash index, check the block state
+	before checking block->is_hashed, because the latter may be
+	uninitialized right after server startup.
+
 2010-01-12	The InnoDB Team
 
 	* handler/ha_innodb.cc, handler/ha_innodb.h:
diff --git a/storage/innodb_plugin/buf/buf0buf.c b/storage/innodb_plugin/buf/buf0buf.c
index 111d396fbc5..549d1a31942 100644
--- a/storage/innodb_plugin/buf/buf0buf.c
+++ b/storage/innodb_plugin/buf/buf0buf.c
@@ -1058,7 +1058,9 @@ buf_pool_drop_hash_index(void)
 				when we have an x-latch on btr_search_latch;
 				see the comment in buf0buf.h */
 
-				if (!block->is_hashed) {
+				if (buf_block_get_state(block)
+				    != BUF_BLOCK_FILE_PAGE
+				    || !block->is_hashed) {
 					continue;
 				}
 

From 4888c5e016fef9f4b10b569770d8eec1f2ab4d9d Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:21:13 +0400
Subject: [PATCH 025/116] Applying InnoDB snapshot

Detailed revision comments:

r6446 | marko | 2010-01-13 17:20:10 +0200 (Wed, 13 Jan 2010) | 3 lines
branches/zip: Treat mem_hash_mutex specially in mutex_free(),
and explicitly free mem_hash_mutex in mem_close().
This fixes the breakage of UNIV_MEM_DEBUG that was filed as Issue #434.
---
 storage/innodb_plugin/include/mem0dbg.h  |  7 +++++++
 storage/innodb_plugin/include/mem0dbg.ic |  3 ---
 storage/innodb_plugin/mem/mem0dbg.c      |  4 ++++
 storage/innodb_plugin/sync/sync0sync.c   | 19 ++++++++++++++++++-
 4 files changed, 29 insertions(+), 4 deletions(-)

diff --git a/storage/innodb_plugin/include/mem0dbg.h b/storage/innodb_plugin/include/mem0dbg.h
index a064af5c678..8ddf4a13cba 100644
--- a/storage/innodb_plugin/include/mem0dbg.h
+++ b/storage/innodb_plugin/include/mem0dbg.h
@@ -28,6 +28,13 @@ Created 6/9/1994 Heikki Tuuri
 check fields whose sizes are given below */
 
 #ifdef UNIV_MEM_DEBUG
+# ifndef UNIV_HOTBACKUP
+/* The mutex which protects in the debug version the hash table
+containing the list of live memory heaps, and also the global
+variables in mem0dbg.c. */
+extern mutex_t	mem_hash_mutex;
+# endif /* !UNIV_HOTBACKUP */
+
 #define MEM_FIELD_HEADER_SIZE	ut_calc_align(2 * sizeof(ulint),\
 						UNIV_MEM_ALIGNMENT)
 #define MEM_FIELD_TRAILER_SIZE	sizeof(ulint)
diff --git a/storage/innodb_plugin/include/mem0dbg.ic b/storage/innodb_plugin/include/mem0dbg.ic
index cb9245411dc..9c6e5a78263 100644
--- a/storage/innodb_plugin/include/mem0dbg.ic
+++ b/storage/innodb_plugin/include/mem0dbg.ic
@@ -25,9 +25,6 @@ Created 6/8/1994 Heikki Tuuri
 *************************************************************************/
 
 #ifdef UNIV_MEM_DEBUG
-# ifndef UNIV_HOTBACKUP
-extern mutex_t	mem_hash_mutex;
-# endif /* !UNIV_HOTBACKUP */
 extern ulint	mem_current_allocated_memory;
 
 /******************************************************************//**
diff --git a/storage/innodb_plugin/mem/mem0dbg.c b/storage/innodb_plugin/mem/mem0dbg.c
index 01eda20ec45..4973ead4213 100644
--- a/storage/innodb_plugin/mem/mem0dbg.c
+++ b/storage/innodb_plugin/mem/mem0dbg.c
@@ -180,6 +180,10 @@ mem_close(void)
 {
 	mem_pool_free(mem_comm_pool);
 	mem_comm_pool = NULL;
+#ifdef UNIV_MEM_DEBUG
+	mutex_free(&mem_hash_mutex);
+	mem_hash_initialized = FALSE;
+#endif /* UNIV_MEM_DEBUG */
 }
 #endif /* !UNIV_HOTBACKUP */
 
diff --git a/storage/innodb_plugin/sync/sync0sync.c b/storage/innodb_plugin/sync/sync0sync.c
index 569fc6328c4..f1527907ef6 100644
--- a/storage/innodb_plugin/sync/sync0sync.c
+++ b/storage/innodb_plugin/sync/sync0sync.c
@@ -315,6 +315,15 @@ mutex_free(
 	ut_a(mutex_get_lock_word(mutex) == 0);
 	ut_a(mutex_get_waiters(mutex) == 0);
 
+#ifdef UNIV_MEM_DEBUG
+	if (mutex == &mem_hash_mutex) {
+		ut_ad(UT_LIST_GET_LEN(mutex_list) == 1);
+		ut_ad(UT_LIST_GET_FIRST(mutex_list) == &mem_hash_mutex);
+		UT_LIST_REMOVE(list, mutex_list, mutex);
+		goto func_exit;
+	}
+#endif /* UNIV_MEM_DEBUG */
+
 	if (mutex != &mutex_list_mutex
 #ifdef UNIV_SYNC_DEBUG
 	    && mutex != &sync_thread_mutex
@@ -336,7 +345,9 @@ mutex_free(
 	}
 
 	os_event_free(mutex->event);
-
+#ifdef UNIV_MEM_DEBUG
+func_exit:
+#endif /* UNIV_MEM_DEBUG */
 #if !defined(HAVE_ATOMIC_BUILTINS)
 	os_fast_mutex_free(&(mutex->os_fast_mutex));
 #endif
@@ -1370,6 +1381,12 @@ sync_close(void)
 	mutex = UT_LIST_GET_FIRST(mutex_list);
 
 	while (mutex) {
+#ifdef UNIV_MEM_DEBUG
+		if (mutex == &mem_hash_mutex) {
+			mutex = UT_LIST_GET_NEXT(list, mutex);
+			continue;
+		}
+#endif /* UNIV_MEM_DEBUG */
 		mutex_free(mutex);
 		mutex = UT_LIST_GET_FIRST(mutex_list);
 	}

From 55cc8f1f1164b4eddfb0d02fdff6cd057454e481 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:21:36 +0400
Subject: [PATCH 026/116] Applying InnoDB snapshot

Detailed revision comments:

r6447 | marko | 2010-01-13 17:43:44 +0200 (Wed, 13 Jan 2010) | 5 lines
branches/zip: row_sel_get_clust_rec_for_mysql(): On the READ UNCOMMITTED
isolation level, do not attempt to access a clustered index record
that has been marked for deletion.  This fixes Issue #433.

Approved by Heikki over the IM.
---
 storage/innodb_plugin/ChangeLog     | 8 ++++++++
 storage/innodb_plugin/row/row0sel.c | 1 +
 2 files changed, 9 insertions(+)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index d5748c825b6..c0ad21cd132 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,11 @@
+2010-01-13	The InnoDB Team
+
+	* row/row0sel.c:
+	On the READ UNCOMMITTED isolation level, do not attempt to access
+	a clustered index record that has been marked for deletion.  The
+	built-in InnoDB in MySQL 5.1 and earlier would attempt to retrieve
+	a previous version of the record in this case.
+
 2010-01-13	The InnoDB Team
 
 	* buf/buf0buf.c:
diff --git a/storage/innodb_plugin/row/row0sel.c b/storage/innodb_plugin/row/row0sel.c
index aacd96ea65e..e14f29d8d64 100644
--- a/storage/innodb_plugin/row/row0sel.c
+++ b/storage/innodb_plugin/row/row0sel.c
@@ -2978,6 +2978,7 @@ row_sel_get_clust_rec_for_mysql(
 
 		if (clust_rec
 		    && (old_vers
+			|| trx->isolation_level <= TRX_ISO_READ_UNCOMMITTED
 			|| rec_get_deleted_flag(rec, dict_table_is_comp(
 							sec_index->table)))
 		    && !row_sel_sec_rec_is_for_clust_rec(

From 169256184e24f495e4d1632101f03bd91e570f31 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:22:07 +0400
Subject: [PATCH 027/116] Applying InnoDB snapshot

Detailed revision comments:

r6449 | marko | 2010-01-13 22:38:53 +0200 (Wed, 13 Jan 2010) | 18 lines
branches/zip: lock_rec_validate_page(): Only validate the record
queues when the thread is not holding a space->latch.

When UNIV_DEBUG is defined while UNIV_SYNC_DEBUG is not,
latching order violations will still occur and deadlocks will be possible.

sync_thread_levels_nonempty_gen(): Renamed from
sync_thread_levels_empty_gen().  Return the violating latch or NULL
instead of FALSE or TRUE, except that there will be a ut_error before
the non-NULL return.

sync_thread_levels_empty_gen(): A macro that negates the return value of
sync_thread_levels_nonempty_gen().

sync_thread_levels_contains(): New function, based on
sync_thread_levels_nonempty_gen().

This should fix Issue #441.
---
 storage/innodb_plugin/include/sync0sync.h | 21 +++++--
 storage/innodb_plugin/lock/lock0lock.c    |  7 +++
 storage/innodb_plugin/sync/sync0sync.c    | 68 ++++++++++++++++++++---
 3 files changed, 82 insertions(+), 14 deletions(-)

diff --git a/storage/innodb_plugin/include/sync0sync.h b/storage/innodb_plugin/include/sync0sync.h
index df990823cc4..f8d86eb1c84 100644
--- a/storage/innodb_plugin/include/sync0sync.h
+++ b/storage/innodb_plugin/include/sync0sync.h
@@ -238,16 +238,27 @@ ibool
 sync_thread_levels_empty(void);
 /*==========================*/
 /******************************************************************//**
-Checks that the level array for the current thread is empty.
-@return	TRUE if empty except the exceptions specified below */
+Checks if the level array for the current thread contains a
+mutex or rw-latch at the specified level.
+@return	a matching latch, or NULL if not found */
 UNIV_INTERN
-ibool
-sync_thread_levels_empty_gen(
-/*=========================*/
+void*
+sync_thread_levels_contains(
+/*========================*/
+	ulint	level);			/*!< in: latching order level
+					(SYNC_DICT, ...)*/
+/******************************************************************//**
+Checks if the level array for the current thread is empty.
+@return	a latch, or NULL if empty except the exceptions specified below */
+UNIV_INTERN
+void*
+sync_thread_levels_nonempty_gen(
+/*============================*/
 	ibool	dict_mutex_allowed);	/*!< in: TRUE if dictionary mutex is
 					allowed to be owned by the thread,
 					also purge_is_running mutex is
 					allowed */
+#define sync_thread_levels_empty_gen(d) (!sync_thread_levels_nonempty_gen(d))
 /******************************************************************//**
 Gets the debug information for a reserved mutex. */
 UNIV_INTERN
diff --git a/storage/innodb_plugin/lock/lock0lock.c b/storage/innodb_plugin/lock/lock0lock.c
index 3db7c2b36bc..0fa67d31716 100644
--- a/storage/innodb_plugin/lock/lock0lock.c
+++ b/storage/innodb_plugin/lock/lock0lock.c
@@ -4766,6 +4766,13 @@ loop:
 	     || lock->trx->conc_state == TRX_PREPARED
 	     || lock->trx->conc_state == TRX_COMMITTED_IN_MEMORY);
 
+# ifdef UNIV_SYNC_DEBUG
+	/* Only validate the record queues when this thread is not
+	holding a space->latch.  Deadlocks are possible due to
+	latching order violation when UNIV_DEBUG is defined while
+	UNIV_SYNC_DEBUG is not. */
+	if (!sync_thread_levels_contains(SYNC_FSP))
+# endif /* UNIV_SYNC_DEBUG */
 	for (i = nth_bit; i < lock_rec_get_n_bits(lock); i++) {
 
 		if (i == 1 || lock_rec_get_nth_bit(lock, i)) {
diff --git a/storage/innodb_plugin/sync/sync0sync.c b/storage/innodb_plugin/sync/sync0sync.c
index f1527907ef6..669e0667485 100644
--- a/storage/innodb_plugin/sync/sync0sync.c
+++ b/storage/innodb_plugin/sync/sync0sync.c
@@ -958,12 +958,62 @@ sync_thread_levels_contain(
 }
 
 /******************************************************************//**
-Checks that the level array for the current thread is empty.
-@return	TRUE if empty except the exceptions specified below */
+Checks if the level array for the current thread contains a
+mutex or rw-latch at the specified level.
+@return	a matching latch, or NULL if not found */
 UNIV_INTERN
-ibool
-sync_thread_levels_empty_gen(
-/*=========================*/
+void*
+sync_thread_levels_contains(
+/*========================*/
+	ulint	level)			/*!< in: latching order level
+					(SYNC_DICT, ...)*/
+{
+	sync_level_t*	arr;
+	sync_thread_t*	thread_slot;
+	sync_level_t*	slot;
+	ulint		i;
+
+	if (!sync_order_checks_on) {
+
+		return(NULL);
+	}
+
+	mutex_enter(&sync_thread_mutex);
+
+	thread_slot = sync_thread_level_arrays_find_slot();
+
+	if (thread_slot == NULL) {
+
+		mutex_exit(&sync_thread_mutex);
+
+		return(NULL);
+	}
+
+	arr = thread_slot->levels;
+
+	for (i = 0; i < SYNC_THREAD_N_LEVELS; i++) {
+
+		slot = sync_thread_levels_get_nth(arr, i);
+
+		if (slot->latch != NULL && slot->level == level) {
+
+			mutex_exit(&sync_thread_mutex);
+			return(slot->latch);
+		}
+	}
+
+	mutex_exit(&sync_thread_mutex);
+
+	return(NULL);
+}
+
+/******************************************************************//**
+Checks that the level array for the current thread is empty.
+@return	a latch, or NULL if empty except the exceptions specified below */
+UNIV_INTERN
+void*
+sync_thread_levels_nonempty_gen(
+/*============================*/
 	ibool	dict_mutex_allowed)	/*!< in: TRUE if dictionary mutex is
 					allowed to be owned by the thread,
 					also purge_is_running mutex is
@@ -976,7 +1026,7 @@ sync_thread_levels_empty_gen(
 
 	if (!sync_order_checks_on) {
 
-		return(TRUE);
+		return(NULL);
 	}
 
 	mutex_enter(&sync_thread_mutex);
@@ -987,7 +1037,7 @@ sync_thread_levels_empty_gen(
 
 		mutex_exit(&sync_thread_mutex);
 
-		return(TRUE);
+		return(NULL);
 	}
 
 	arr = thread_slot->levels;
@@ -1004,13 +1054,13 @@ sync_thread_levels_empty_gen(
 			mutex_exit(&sync_thread_mutex);
 			ut_error;
 
-			return(FALSE);
+			return(slot->latch);
 		}
 	}
 
 	mutex_exit(&sync_thread_mutex);
 
-	return(TRUE);
+	return(NULL);
 }
 
 /******************************************************************//**

From 535b4ccb06856e435f6af24ed330aafe432ccb4e Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:22:39 +0400
Subject: [PATCH 028/116] Applying InnoDB snapshot

Detailed revision comments:

r6463 | marko | 2010-01-14 15:43:37 +0200 (Thu, 14 Jan 2010) | 5 lines
branches/zip: page_copy_rec_list_end(), page_copy_rec_list_start():
Update PAGE_MAX_TRX_ID before attempting to compress the page.  This
fixes Issue #382 (a debug assertion failure in page_zip_reorganize())
and reduces the generated redo log.  There was no bug or crash in
non-debug builds.
---
 storage/innodb_plugin/page/page0page.c | 34 +++++++++++++++-----------
 1 file changed, 20 insertions(+), 14 deletions(-)

diff --git a/storage/innodb_plugin/page/page0page.c b/storage/innodb_plugin/page/page0page.c
index ab2ba60570e..17c40170e14 100644
--- a/storage/innodb_plugin/page/page0page.c
+++ b/storage/innodb_plugin/page/page0page.c
@@ -658,6 +658,14 @@ page_copy_rec_list_end(
 						index, mtr);
 	}
 
+	/* Update PAGE_MAX_TRX_ID on the uncompressed page.
+	Modifications will be redo logged and copied to the compressed
+	page in page_zip_compress() or page_zip_reorganize() below. */
+	if (dict_index_is_sec_or_ibuf(index) && page_is_leaf(page)) {
+		page_update_max_trx_id(new_block, NULL,
+				       page_get_max_trx_id(page), mtr);
+	}
+
 	if (UNIV_LIKELY_NULL(new_page_zip)) {
 		mtr_set_log_mode(mtr, log_mode);
 
@@ -696,15 +704,10 @@ page_copy_rec_list_end(
 		}
 	}
 
-	/* Update the lock table, MAX_TRX_ID, and possible hash index */
+	/* Update the lock table and possible hash index */
 
 	lock_move_rec_list_end(new_block, block, rec);
 
-	if (dict_index_is_sec_or_ibuf(index) && page_is_leaf(page)) {
-		page_update_max_trx_id(new_block, new_page_zip,
-				       page_get_max_trx_id(page), mtr);
-	}
-
 	btr_search_move_or_delete_hash_entries(new_block, block, index);
 
 	return(ret);
@@ -772,6 +775,16 @@ page_copy_rec_list_start(
 		mem_heap_free(heap);
 	}
 
+	/* Update PAGE_MAX_TRX_ID on the uncompressed page.
+	Modifications will be redo logged and copied to the compressed
+	page in page_zip_compress() or page_zip_reorganize() below. */
+	if (dict_index_is_sec_or_ibuf(index)
+	    && page_is_leaf(page_align(rec))) {
+		page_update_max_trx_id(new_block, NULL,
+				       page_get_max_trx_id(page_align(rec)),
+				       mtr);
+	}
+
 	if (UNIV_LIKELY_NULL(new_page_zip)) {
 		mtr_set_log_mode(mtr, log_mode);
 
@@ -809,14 +822,7 @@ page_copy_rec_list_start(
 		}
 	}
 
-	/* Update MAX_TRX_ID, the lock table, and possible hash index */
-
-	if (dict_index_is_sec_or_ibuf(index)
-	    && page_is_leaf(page_align(rec))) {
-		page_update_max_trx_id(new_block, new_page_zip,
-				       page_get_max_trx_id(page_align(rec)),
-				       mtr);
-	}
+	/* Update the lock table and possible hash index */
 
 	lock_move_rec_list_start(new_block, block, rec, ret);
 

From 5cd0fae437ede398cecec5d997e3863079bdbf56 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:23:04 +0400
Subject: [PATCH 029/116] Applying InnoDB snapshot

Detailed revision comments:

r6467 | inaam | 2010-01-14 20:46:00 +0200 (Thu, 14 Jan 2010) | 10 lines
branches/zip rb://226

log_sys->written_to_all_lsn does not accurately represent the LSN
upto which write and flush has taken place. Under a race condition
it can fall behind log_sys->flushed_to_disk_lsn which is accurate.
Besides written_to_all_lsn is redundant as currently InnoDB supports
only one log group.

Approved by: Heikki
---
 storage/innodb_plugin/include/log0log.h | 12 +++++++++++-
 storage/innodb_plugin/log/log0log.c     |  2 +-
 2 files changed, 12 insertions(+), 2 deletions(-)

diff --git a/storage/innodb_plugin/include/log0log.h b/storage/innodb_plugin/include/log0log.h
index 135aeb69e2d..233714eb63a 100644
--- a/storage/innodb_plugin/include/log0log.h
+++ b/storage/innodb_plugin/include/log0log.h
@@ -825,7 +825,17 @@ struct log_struct{
 					written to some log group; for this to
 					be advanced, it is enough that the
 					write i/o has been completed for all
-					log groups */
+					log groups.
+					Note that since InnoDB currently
+					has only one log group therefore
+					this value is redundant. Also it
+					is possible that this value
+					falls behind the
+					flushed_to_disk_lsn transiently.
+					It is appropriate to use either
+					flushed_to_disk_lsn or
+					write_lsn which are always
+					up-to-date and accurate. */
 	ib_uint64_t	write_lsn;	/*!< end lsn for the current running
 					write */
 	ulint		write_end_offset;/*!< the data in buffer has
diff --git a/storage/innodb_plugin/log/log0log.c b/storage/innodb_plugin/log/log0log.c
index 86c9f9b4130..063581055e9 100644
--- a/storage/innodb_plugin/log/log0log.c
+++ b/storage/innodb_plugin/log/log0log.c
@@ -2013,7 +2013,7 @@ log_checkpoint(
 		return(TRUE);
 	}
 
-	ut_ad(log_sys->written_to_all_lsn >= oldest_lsn);
+	ut_ad(log_sys->flushed_to_disk_lsn >= oldest_lsn);
 
 	if (log_sys->n_pending_checkpoint_writes > 0) {
 		/* A checkpoint write is running */

From 33bac9249fc19c371bacd68baef6e78dee032b7b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:25:06 +0400
Subject: [PATCH 030/116] Applying InnoDB snapshot

Detailed revision comments:

r6472 | calvin | 2010-01-16 01:53:47 +0200 (Sat, 16 Jan 2010) | 12 lines
branches/zip: Merge revisions 6425:6471 from branches/5.1
to pick up the first part fix of bug49396.

    ------------------------------------------------------------------------
    r6471 | calvin | 2010-01-15 17:43:27 -0600 (Fri, 15 Jan 2010) | 4 lines

    branches/5.1: fix bug#49396: main.innodb test fails in embedded mode

    Change replace_result by using $MYSQLD_DATADIR. Tested in both embedded
    mode and normal server mode.
    ------------------------------------------------------------------------

r6473 | calvin | 2010-01-16 01:58:16 +0200 (Sat, 16 Jan 2010) | 6 lines
branches/zip: fix bug#49396: innodb.innodb-index test fails in
embedded mode

This is 2nd part of the fix for bug#49396. The 1st part is
innodb.test. Tested in both embedded mode and normal server mode.
---
 mysql-test/suite/innodb/t/innodb-index.test | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/mysql-test/suite/innodb/t/innodb-index.test b/mysql-test/suite/innodb/t/innodb-index.test
index b5dd2e037e7..09d6e2f8e3f 100644
--- a/mysql-test/suite/innodb/t/innodb-index.test
+++ b/mysql-test/suite/innodb/t/innodb-index.test
@@ -1,5 +1,7 @@
 -- source include/have_innodb.inc
 
+let $MYSQLD_DATADIR= `select @@datadir`;
+
 let $innodb_file_format_check_orig=`select @@innodb_file_format_check`;
 
 create table t1(a int not null, b int, c char(10) not null, d varchar(20)) engine = innodb;
@@ -146,7 +148,9 @@ delete from t1;
 --error ER_CANT_DROP_FIELD_OR_KEY
 drop index dc on t4;
 # there is no foreign key dc on t3
---replace_regex /'\.\/test\/#sql2-[0-9a-f-]*'/'#sql2-temporary'/
+--replace_regex /'[^']*test\/#sql2-[0-9a-f-]*'/'#sql2-temporary'/
+# Embedded server doesn't chdir to data directory
+--replace_result $MYSQLD_DATADIR ./ master-data/ ''
 --error ER_ERROR_ON_RENAME
 alter table t3 drop foreign key dc;
 alter table t4 drop foreign key dc;

From 2eeef6b03632bae035c1f4d49b8ef34ae92c9676 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:25:39 +0400
Subject: [PATCH 031/116] Applying InnoDB snapshot

Detailed revision comments:

r6498 | marko | 2010-01-21 11:22:52 +0200 (Thu, 21 Jan 2010) | 15 lines
branches/zip: buf_page_get_gen(): Obey recv_no_ibuf_operations
and do not call ibuf_merge_or_delete_for_page() in crash recovery,
before the redo log has been applied.
This could cure some hard-to-repeat, hard-to-explain bugs
related to secondary indexes.

A possible recipe to repeat the bug:

1. update a secondary index leaf page on a compressed table
2. evict the page from the buffer pool while it is still dirty
3. ibuf_insert() something for the page
4. crash
5. crash recovery; ibuf merge would be done too early,
before applying redo log to the sec index page or the ibuf pages
---
 storage/innodb_plugin/ChangeLog     | 6 ++++++
 storage/innodb_plugin/buf/buf0buf.c | 2 +-
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index c0ad21cd132..45f99504b67 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,9 @@
+2010-01-21	The InnoDB Team
+
+	* buf/buf0buf.c:
+	Do not merge buffered inserts to compressed pages before
+	the redo log has been applied in crash recovery.
+
 2010-01-13	The InnoDB Team
 
 	* row/row0sel.c:
diff --git a/storage/innodb_plugin/buf/buf0buf.c b/storage/innodb_plugin/buf/buf0buf.c
index 549d1a31942..fd5b07330cb 100644
--- a/storage/innodb_plugin/buf/buf0buf.c
+++ b/storage/innodb_plugin/buf/buf0buf.c
@@ -2239,7 +2239,7 @@ wait_until_unfixed:
 		while not holding buf_pool_mutex or block->mutex. */
 		success = buf_zip_decompress(block, srv_use_checksums);
 
-		if (UNIV_LIKELY(success)) {
+		if (UNIV_LIKELY(success && !recv_no_ibuf_operations)) {
 			ibuf_merge_or_delete_for_page(block, space, offset,
 						      zip_size, TRUE);
 		}

From 6a4c518f0de71cc0eebcf5c92bbb335f9bfebd57 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:26:39 +0400
Subject: [PATCH 032/116] Applying InnoDB snapshot, fixes BUG#41609.

Detailed revision comments:

r6521 | marko | 2010-01-27 10:49:01 +0200 (Wed, 27 Jan 2010) | 17 lines
branches/zip: Drop temporary tables at startup.
This addresses the third aspect of Bug #41609.

row_mysql_drop_temp_tables(): New function, to drop all temporary
tables. These can be distinguished by the least significant bit of
MIX_LEN. However, we will skip ROW_FORMAT=REDUNDANT tables, because in
the records for those tables, that bit may be garbage.

recv_recovery_from_checkpoint_finish(): Invoke
row_mysql_drop_temp_tables().  Normally, if the .frm files for the
temporary tables exist at startup, MySQL will ask InnoDB to drop the
temporary tables.  However, if the files are deleted, for instance, by
the boot scripts of the operating system, the tables would remain in
the InnoDB data dictionary unless someone digs them up by
innodb_table_monitor and creates .frm files for dropping the tables.

rb://221 approved by Sunny Bains.
---
 storage/innodb_plugin/ChangeLog           |  7 ++
 storage/innodb_plugin/include/row0mysql.h |  6 ++
 storage/innodb_plugin/log/log0recv.c      |  2 +
 storage/innodb_plugin/row/row0mysql.c     | 93 +++++++++++++++++++++++
 4 files changed, 108 insertions(+)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 45f99504b67..85bcaf9dcf7 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,10 @@
+2010-01-27	The InnoDB Team
+
+	* include/row0mysql.h, log/log0recv.c, row/row0mysql.c:
+	Drop temporary tables at startup.
+	This addresses the third aspect of
+	Bug#41609 Crash recovery does not work for InnoDB temporary tables.
+
 2010-01-21	The InnoDB Team
 
 	* buf/buf0buf.c:
diff --git a/storage/innodb_plugin/include/row0mysql.h b/storage/innodb_plugin/include/row0mysql.h
index b05241f00f8..28cd528bd25 100644
--- a/storage/innodb_plugin/include/row0mysql.h
+++ b/storage/innodb_plugin/include/row0mysql.h
@@ -451,6 +451,12 @@ row_drop_table_for_mysql(
 	const char*	name,	/*!< in: table name */
 	trx_t*		trx,	/*!< in: transaction handle */
 	ibool		drop_db);/*!< in: TRUE=dropping whole database */
+/*********************************************************************//**
+Drop all temporary tables during crash recovery. */
+UNIV_INTERN
+void
+row_mysql_drop_temp_tables(void);
+/*============================*/
 
 /*********************************************************************//**
 Discards the tablespace of a table which stored in an .ibd file. Discarding
diff --git a/storage/innodb_plugin/log/log0recv.c b/storage/innodb_plugin/log/log0recv.c
index d33381e495d..b060c9e6e24 100644
--- a/storage/innodb_plugin/log/log0recv.c
+++ b/storage/innodb_plugin/log/log0recv.c
@@ -3264,6 +3264,8 @@ recv_recovery_from_checkpoint_finish(void)
 
 	/* Drop partially created indexes. */
 	row_merge_drop_temp_indexes();
+	/* Drop temporary tables. */
+	row_mysql_drop_temp_tables();
 
 #ifdef UNIV_SYNC_DEBUG
 	/* Wait for a while so that created threads have time to suspend
diff --git a/storage/innodb_plugin/row/row0mysql.c b/storage/innodb_plugin/row/row0mysql.c
index e9fd12e9747..e7a44d10b61 100644
--- a/storage/innodb_plugin/row/row0mysql.c
+++ b/storage/innodb_plugin/row/row0mysql.c
@@ -3363,6 +3363,99 @@ funct_exit:
 	return((int) err);
 }
 
+/*********************************************************************//**
+Drop all temporary tables during crash recovery. */
+UNIV_INTERN
+void
+row_mysql_drop_temp_tables(void)
+/*============================*/
+{
+	trx_t*	trx;
+	ulint	err;
+
+	trx = trx_allocate_for_background();
+	trx->op_info = "dropping temporary tables";
+	row_mysql_lock_data_dictionary(trx);
+
+	err = que_eval_sql(
+		NULL,
+		"PROCEDURE DROP_TEMP_TABLES_PROC () IS\n"
+		"table_name CHAR;\n"
+		"table_id CHAR;\n"
+		"foreign_id CHAR;\n"
+		"index_id CHAR;\n"
+		"DECLARE CURSOR c IS SELECT NAME,ID FROM SYS_TABLES\n"
+		"WHERE N_COLS > 2147483647\n"
+		/* N_COLS>>31 is set unless ROW_FORMAT=REDUNDANT,
+		and MIX_LEN may be garbage for those tables */
+		"AND MIX_LEN=(MIX_LEN/2*2+1);\n"
+		/* MIX_LEN & 1 is set for temporary tables */
+#if DICT_TF2_TEMPORARY != 1
+# error "DICT_TF2_TEMPORARY != 1"
+#endif
+		"BEGIN\n"
+		"OPEN c;\n"
+		"WHILE 1=1 LOOP\n"
+		"	FETCH c INTO table_name, table_id;\n"
+		"	IF (SQL % NOTFOUND) THEN\n"
+		"		EXIT;\n"
+		"	END IF;\n"
+		"	WHILE 1=1 LOOP\n"
+		"		SELECT ID INTO index_id\n"
+		"		FROM SYS_INDEXES\n"
+		"		WHERE TABLE_ID = table_id\n"
+		"		LOCK IN SHARE MODE;\n"
+		"		IF (SQL % NOTFOUND) THEN\n"
+		"			EXIT;\n"
+		"		END IF;\n"
+
+		/* Do not drop tables for which there exist
+		foreign key constraints. */
+		"		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;"
+		"		IF NOT (SQL % NOTFOUND) THEN\n"
+		"			EXIT;\n"
+		"		END IF;\n"
+
+		"		SELECT ID INTO foreign_id\n"
+		"		FROM SYS_FOREIGN\n"
+		"		WHERE REF_NAME = table_name\n"
+		"		AND TO_BINARY(REF_NAME)\n"
+		"		  = TO_BINARY(table_name)\n;"
+		"		IF NOT (SQL % NOTFOUND) THEN\n"
+		"			EXIT;\n"
+		"		END IF;\n"
+
+		"		DELETE FROM SYS_FIELDS\n"
+		"		WHERE INDEX_ID = index_id;\n"
+		"		DELETE FROM SYS_INDEXES\n"
+		"		WHERE ID = index_id\n"
+		"		AND TABLE_ID = table_id;\n"
+		"	END LOOP;\n"
+		"	DELETE FROM SYS_COLUMNS\n"
+		"	WHERE TABLE_ID = table_id;\n"
+		"	DELETE FROM SYS_TABLES\n"
+		"	WHERE ID = table_id;\n"
+		"END LOOP;\n"
+		"COMMIT WORK;\n"
+		"END;\n"
+		, FALSE, trx);
+
+	if (err != DB_SUCCESS) {
+		ut_print_timestamp(stderr);
+		fprintf(stderr,
+			"  InnoDB: Failed to drop temporary tables:"
+			" error %lu occurred\n",
+			(ulong) err);
+	}
+
+	row_mysql_unlock_data_dictionary(trx);
+	trx_free_for_background(trx);
+}
+
 /*******************************************************************//**
 Drop all foreign keys in a database, see Bug#18942.
 Called at the end of row_drop_database_for_mysql().

From 43a62af1543924b9b8b9c5683c6128e15e68ccf1 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:27:18 +0400
Subject: [PATCH 033/116] Applying InnoDB snapshot, fixes BUG#35077.

Detailed revision comments:

r6525 | marko | 2010-01-28 16:23:15 +0200 (Thu, 28 Jan 2010) | 11 lines
branches/zip: buf_LRU_invalidate_tablespace(): Do not unnecessarily
acquire the block_mutex for every block in the LRU list. Only acquire
it when holding buf_pool_mutex is not sufficient. This should speed up
the function and considerably reduce traffic on the memory bus and
caches.

I noticed this deficiency when working on Issue #157.
This deficiency popped up again in Issue #449 (Bug #35077),
which this fix does not fully address.

rb://78 revision 1 approved by Heikki Tuuri.
---
 storage/innodb_plugin/buf/buf0lru.c | 32 +++++++++++++++++++++--------
 1 file changed, 23 insertions(+), 9 deletions(-)

diff --git a/storage/innodb_plugin/buf/buf0lru.c b/storage/innodb_plugin/buf/buf0lru.c
index 4f19fd13fa5..eeef43de83b 100644
--- a/storage/innodb_plugin/buf/buf0lru.c
+++ b/storage/innodb_plugin/buf/buf0lru.c
@@ -347,20 +347,34 @@ scan_again:
 
 	all_freed = TRUE;
 
+rescan:
 	bpage = UT_LIST_GET_LAST(buf_pool->LRU);
 
 	while (bpage != NULL) {
-		mutex_t*	block_mutex = buf_page_get_mutex(bpage);
 		buf_page_t*	prev_bpage;
 
 		ut_a(buf_page_in_file(bpage));
 
-		mutex_enter(block_mutex);
 		prev_bpage = UT_LIST_GET_PREV(LRU, bpage);
 
-		if (buf_page_get_space(bpage) == id) {
-			if (bpage->buf_fix_count > 0
-			    || buf_page_get_io_fix(bpage) != BUF_IO_NONE) {
+		/* bpage->space and bpage->io_fix are protected by
+		buf_pool_mutex and block_mutex.  It is safe to check
+		them while holding buf_pool_mutex only. */
+
+		if (buf_page_get_space(bpage) != id) {
+			/* Skip this block, as it does not belong to
+			the space that is being invalidated. */
+		} else if (buf_page_get_io_fix(bpage) != BUF_IO_NONE) {
+			/* We cannot remove this page during this scan
+			yet; maybe the system is currently reading it
+			in, or flushing the modifications to the file */
+
+			all_freed = FALSE;
+		} else {
+			mutex_t* block_mutex = buf_page_get_mutex(bpage);
+			mutex_enter(block_mutex);
+
+			if (bpage->buf_fix_count > 0) {
 
 				/* We cannot remove this page during
 				this scan yet; maybe the system is
@@ -423,12 +437,12 @@ scan_again:
 				buf_buddy_free() may have relocated
 				prev_bpage.  Rescan the LRU list. */
 
-				bpage = UT_LIST_GET_LAST(buf_pool->LRU);
-				continue;
+				goto rescan;
 			}
-		}
 next_page:
-		mutex_exit(block_mutex);
+			mutex_exit(block_mutex);
+		}
+
 		bpage = prev_bpage;
 	}
 

From 09353b5909bc9efd6b17514b95f520d9a40a4a4b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:30:11 +0400
Subject: [PATCH 034/116] Applying InnoDB snapshot, fixes BUG#47622.

Detailed revision comments:

r6526 | jyang | 2010-01-28 18:12:40 +0200 (Thu, 28 Jan 2010) | 8 lines
branches/zip: Add index translation table to map mysql index
number to InnoDB index structure directly. Fix Bug #47622:
"the new index is added before the existing ones in MySQL,
but after one in SE".

rb://215, approved by Marko
---
 .../suite/innodb/r/innodb_bug47622.result     |  23 ++
 .../suite/innodb/t/innodb_bug47622.test       |  55 ++++
 storage/innodb_plugin/handler/ha_innodb.cc    | 262 +++++++++++++++++-
 storage/innodb_plugin/handler/ha_innodb.h     |  30 +-
 .../innodb_plugin/handler/handler0alter.cc    |   8 +
 5 files changed, 363 insertions(+), 15 deletions(-)
 create mode 100644 mysql-test/suite/innodb/r/innodb_bug47622.result
 create mode 100644 mysql-test/suite/innodb/t/innodb_bug47622.test

diff --git a/mysql-test/suite/innodb/r/innodb_bug47622.result b/mysql-test/suite/innodb/r/innodb_bug47622.result
new file mode 100644
index 00000000000..f5d13711c52
--- /dev/null
+++ b/mysql-test/suite/innodb/r/innodb_bug47622.result
@@ -0,0 +1,23 @@
+CREATE TABLE bug47622(
+`rule_key` int(11) NOT NULL DEFAULT '0',
+`seq` smallint(6) NOT NULL DEFAULT '0',
+`action` smallint(6) NOT NULL DEFAULT '0',
+`arg_id` smallint(6) DEFAULT NULL,
+`else_ind` TINYINT NOT NULL,
+KEY IDX_A (`arg_id`)
+) ENGINE=InnoDB;
+ALTER TABLE bug47622 ADD UNIQUE IDX_B (rule_key,else_ind,seq,action,arg_id);
+drop index IDX_B on bug47622;
+create index idx on bug47622(seq, arg_id);
+ALTER TABLE bug47622 ADD UNIQUE IDX_X (rule_key,else_ind,seq,action);
+drop table bug47622;
+CREATE TABLE bug47622 (
+`a` int(11) NOT NULL,
+`b` int(11) DEFAULT NULL,
+`c` char(10) DEFAULT NULL,
+`d` varchar(20) DEFAULT NULL,
+PRIMARY KEY (`a`),
+KEY `b` (`b`)
+) ENGINE=InnoDB;
+alter table bug47622 add unique index (c), add index (d);
+drop table bug47622;
diff --git a/mysql-test/suite/innodb/t/innodb_bug47622.test b/mysql-test/suite/innodb/t/innodb_bug47622.test
new file mode 100644
index 00000000000..9cf9d0e531b
--- /dev/null
+++ b/mysql-test/suite/innodb/t/innodb_bug47622.test
@@ -0,0 +1,55 @@
+# This is the test for bug 47622. There could be index
+# metadata sequence mismatch between MySQL and Innodb
+# after creating index through FIC interfaces.
+# We resolve the problem by sync the index sequence
+# up when opening the table.
+
+--source include/have_innodb.inc
+
+connect (a,localhost,root,,);
+connect (b,localhost,root,,);
+
+# Create a table with a non-unique index
+CREATE TABLE bug47622(
+	`rule_key` int(11) NOT NULL DEFAULT '0',
+	`seq` smallint(6) NOT NULL DEFAULT '0',
+	`action` smallint(6) NOT NULL DEFAULT '0',
+	`arg_id` smallint(6) DEFAULT NULL,
+	`else_ind` TINYINT NOT NULL,
+	KEY IDX_A (`arg_id`)
+) ENGINE=InnoDB;
+
+connection a;
+
+# A subsequent creating unique index should not trigger
+# any error message. Unique index would be ranked ahead
+# of regular index.
+ALTER TABLE bug47622 ADD UNIQUE IDX_B (rule_key,else_ind,seq,action,arg_id);
+
+drop index IDX_B on bug47622;
+
+# In another connection, create additional set of normal
+# index and unique index. Again, unique index would be ranked
+# ahead of regular index.
+connection b;
+create index idx on bug47622(seq, arg_id);
+
+ALTER TABLE bug47622 ADD UNIQUE IDX_X (rule_key,else_ind,seq,action);
+
+drop table bug47622;
+
+# Create a table with one Primary key and a non-unique key
+CREATE TABLE bug47622 (
+  `a` int(11) NOT NULL,
+  `b` int(11) DEFAULT NULL,
+  `c` char(10) DEFAULT NULL,
+  `d` varchar(20) DEFAULT NULL,
+  PRIMARY KEY (`a`),
+  KEY `b` (`b`)
+) ENGINE=InnoDB;
+
+# Add two index with one unique and one non-unique.
+# Index sequence is "PRIMARY", "c", "b" and "d"
+alter table bug47622 add unique index (c), add index (d);
+
+drop table bug47622;
diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 95b0d331076..9219bd2acc2 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -3067,6 +3067,208 @@ innobase_get_int_col_max_value(
 	return(max_value);
 }
 
+/*******************************************************************//**
+This function checks whether the index column information
+is consistent between KEY info from mysql and that from innodb index.
+@return TRUE if all column types match. */
+static
+ibool
+innobase_match_index_columns(
+/*=========================*/
+	const KEY*		key_info,	/*!< in: Index info
+						from mysql */
+	const dict_index_t*	index_info)	/*!< in: Index info
+						from Innodb */
+{
+	const KEY_PART_INFO*	key_part;
+	const KEY_PART_INFO*	key_end;
+	const dict_field_t*	innodb_idx_fld;
+	const dict_field_t*	innodb_idx_fld_end;
+
+	DBUG_ENTER("innobase_match_index_columns");
+
+	/* Check whether user defined index column count matches */
+	if (key_info->key_parts != index_info->n_user_defined_cols) {
+		DBUG_RETURN(FALSE);
+	}
+
+	key_part = key_info->key_part;
+	key_end = key_part + key_info->key_parts;
+	innodb_idx_fld = index_info->fields;
+	innodb_idx_fld_end = index_info->fields + index_info->n_fields;
+
+	/* Check each index column's datatype. We do not check
+	column name because there exists case that index
+	column name got modified in mysql but such change does not
+	propagate to InnoDB.
+	One hidden assumption here is that the index column sequences
+	are matched up between those in mysql and Innodb. */
+	for (; key_part != key_end; ++key_part) {
+		ulint	col_type;
+		ibool	is_unsigned;
+		ulint	mtype = innodb_idx_fld->col->mtype;
+
+		/* Need to translate to InnoDB column type before
+		comparison. */
+		col_type = get_innobase_type_from_mysql_type(&is_unsigned,
+							     key_part->field);
+
+		/* Ignore Innodb specific system columns. */
+		while (mtype == DATA_SYS) {
+			innodb_idx_fld++;
+
+			if (innodb_idx_fld >= innodb_idx_fld_end) {
+				DBUG_RETURN(FALSE);
+			}
+		}
+
+		if (col_type != mtype) {
+			/* Column Type mismatches */
+			DBUG_RETURN(FALSE);
+		}
+
+		innodb_idx_fld++;
+	}
+
+	DBUG_RETURN(TRUE);
+}
+
+/*******************************************************************//**
+This function builds a translation table in INNOBASE_SHARE
+structure for fast index location with mysql array number from its
+table->key_info structure. This also provides the necessary translation
+between the key order in mysql key_info and Innodb ib_table->indexes if
+they are not fully matched with each other.
+Note we do not have any mutex protecting the translation table
+building based on the assumption that there is no concurrent
+index creation/drop and DMLs that requires index lookup. All table
+handle will be closed before the index creation/drop.
+@return TRUE if index translation table built successfully */
+static
+ibool
+innobase_build_index_translation(
+/*=============================*/
+	const TABLE*		table,	  /*!< in: table in MySQL data
+					  dictionary */
+	dict_table_t*		ib_table, /*!< in: table in Innodb data
+					  dictionary */
+	INNOBASE_SHARE*		share)	  /*!< in/out: share structure
+					  where index translation table
+					  will be constructed in. */
+{
+	ulint		mysql_num_index;
+	ulint		ib_num_index;
+	dict_index_t**	index_mapping;
+	ibool		ret = TRUE;
+
+	DBUG_ENTER("innobase_build_index_translation");
+
+	mysql_num_index = table->s->keys;
+	ib_num_index = UT_LIST_GET_LEN(ib_table->indexes);
+
+	/* Innodb could have own system indexes. */
+	ut_a(ib_num_index >= mysql_num_index);
+
+	index_mapping = share->idx_trans_tbl.index_mapping;
+
+	/* If index entry count is non-zero, nothing has
+	changed since last update, directly return TRUE */
+	if (share->idx_trans_tbl.index_count) {
+		/* Index entry count should still match mysql_num_index */
+		ut_a(share->idx_trans_tbl.index_count == mysql_num_index);
+		goto func_exit;
+	}
+
+	/* The number of index increased, rebuild the mapping table */
+	if (mysql_num_index > share->idx_trans_tbl.array_size) {
+		index_mapping = (dict_index_t**) my_realloc(index_mapping,
+							mysql_num_index *
+							sizeof(*index_mapping),
+							MYF(MY_ALLOW_ZERO_PTR));
+
+		if (!index_mapping) {
+			ret = FALSE;
+			goto func_exit;
+		}
+
+		share->idx_trans_tbl.array_size = mysql_num_index;
+	}
+
+
+	/* For each index in the mysql key_info array, fetch its
+	corresponding InnoDB index pointer into index_mapping
+	array. */
+	for (ulint count = 0; count < mysql_num_index; count++) {
+
+		/* Fetch index pointers into index_mapping according to mysql
+		index sequence */
+		index_mapping[count] = dict_table_get_index_on_name(
+			ib_table, table->key_info[count].name);
+
+		if (!index_mapping[count]) {
+			sql_print_error("Cannot find index %s in InnoDB "
+					"index dictionary.",
+					table->key_info[count].name);
+			ret = FALSE;
+			goto func_exit;
+		}
+
+		/* Double check fetched index has the same
+		column info as those in mysql key_info. */
+		if (!innobase_match_index_columns(&table->key_info[count],
+					          index_mapping[count])) {
+			sql_print_error("Found index %s whose column info "
+					"does not match that of MySQL.",
+					table->key_info[count].name);
+			ret = FALSE;
+			goto func_exit;
+		}
+	}
+
+	/* Successfully built the translation table */
+	share->idx_trans_tbl.index_count = mysql_num_index;
+
+func_exit:
+	if (!ret) {
+		/* Build translation table failed. */
+		my_free(index_mapping, MYF(MY_ALLOW_ZERO_PTR));
+
+		share->idx_trans_tbl.array_size = 0;
+		share->idx_trans_tbl.index_count = 0;
+		index_mapping = NULL;
+	}
+
+	share->idx_trans_tbl.index_mapping = index_mapping;
+
+	DBUG_RETURN(ret);
+}
+
+/*******************************************************************//**
+This function uses index translation table to quickly locate the
+requested index structure.
+Note we do not have mutex protection for the index translatoin table
+access, it is based on the assumption that there is no concurrent
+translation table rebuild (fter create/drop index) and DMLs that
+require index lookup.
+@return dict_index_t structure for requested index. NULL if
+fail to locate the index structure. */
+static
+dict_index_t*
+innobase_index_lookup(
+/*==================*/
+	INNOBASE_SHARE*	share,	/*!< in: share structure for index
+				translation table. */
+	uint		keynr)	/*!< in: index number for the requested
+				index */
+{
+	if (!share->idx_trans_tbl.index_mapping
+	    || keynr >= share->idx_trans_tbl.index_count) {
+		return(NULL);
+	}
+
+	return(share->idx_trans_tbl.index_mapping[keynr]);
+}
+
 /************************************************************************
 Set the autoinc column max value. This should only be called once from
 ha_innobase::open(). Therefore there's no need for a covering lock. */
@@ -3283,6 +3485,11 @@ retry:
 	primary_key = table->s->primary_key;
 	key_used_on_scan = primary_key;
 
+	if (!innobase_build_index_translation(table, ib_table, share)) {
+		  sql_print_error("Build InnoDB index translation table for"
+				  " Table %s failed", name);
+	}
+
 	/* Allocate a buffer for a 'row reference'. A row reference is
 	a string of bytes of length ref_length which uniquely specifies
 	a row in our table. Note that MySQL may also compare two row
@@ -5198,8 +5405,20 @@ ha_innobase::innobase_get_index(
 	if (keynr != MAX_KEY && table->s->keys > 0) {
 		key = table->key_info + keynr;
 
-		index = dict_table_get_index_on_name(prebuilt->table,
-						     key->name);
+		index = innobase_index_lookup(share, keynr);
+
+		if (index) {
+			ut_a(ut_strcmp(index->name, key->name) == 0);
+		} else {
+			sql_print_error("InnoDB could not find index %s "
+					"key no %u for table %s through "
+					" its index translation table",
+					key ? key->name : "NULL", keynr,
+					prebuilt->table->name);
+
+			index = dict_table_get_index_on_name(prebuilt->table,
+							     key->name);
+		}
 	} else {
 		index = dict_table_get_first_index(prebuilt->table);
 	}
@@ -6890,7 +7109,9 @@ ha_innobase::records_in_range(
 
 	key = table->key_info + active_index;
 
-	index = dict_table_get_index_on_name(prebuilt->table, key->name);
+	index = innobase_get_index(keynr);
+
+	ut_a(ut_strcmp(index->name, key->name) == 0);
 
 	/* MySQL knows about this index and so we must be able to find it.*/
 	ut_a(index);
@@ -7079,6 +7300,7 @@ ha_innobase::info(
 	char		path[FN_REFLEN];
 	os_file_stat_t	stat_info;
 
+
 	DBUG_ENTER("info");
 
 	/* If we are forcing recovery at a high level, we will suppress
@@ -7239,13 +7461,29 @@ ha_innobase::info(
 	}
 
 	if (flag & HA_STATUS_CONST) {
-		index = dict_table_get_first_index(ib_table);
+		/* Verify the number of index in InnoDB and MySQL
+		matches up. If prebuilt->clust_index_was_generated
+		holds, InnoDB defines GEN_CLUST_INDEX internally */
+		ulint	num_innodb_index = UT_LIST_GET_LEN(ib_table->indexes)
+					- prebuilt->clust_index_was_generated;
 
-		if (prebuilt->clust_index_was_generated) {
-			index = dict_table_get_next_index(index);
+		if (table->s->keys != num_innodb_index) {
+			sql_print_error("Table %s contains %lu "
+					"indexes inside InnoDB, which "
+					"is different from the number of "
+					"indexes %u defined in the MySQL ",
+					ib_table->name, num_innodb_index,
+					table->s->keys);
 		}
 
 		for (i = 0; i < table->s->keys; i++) {
+			/* We could get index quickly through internal
+			index mapping with the index translation table.
+			The identity of index (match up index name with
+			that of table->key_info[i]) is already verified in
+			innobase_get_index().  */
+			index = innobase_get_index(i);
+
 			if (index == NULL) {
 				sql_print_error("Table %s contains fewer "
 						"indexes inside InnoDB than "
@@ -7297,8 +7535,6 @@ ha_innobase::info(
 				  rec_per_key >= ~(ulong) 0 ? ~(ulong) 0 :
 				  (ulong) rec_per_key;
 			}
-
-			index = dict_table_get_next_index(index);
 		}
 	}
 
@@ -8462,6 +8698,11 @@ static INNOBASE_SHARE* get_share(const char* table_name)
 			    innobase_open_tables, fold, share);
 
 		thr_lock_init(&share->lock);
+
+		/* Index translation table initialization */
+		share->idx_trans_tbl.index_mapping = NULL;
+		share->idx_trans_tbl.index_count = 0;
+		share->idx_trans_tbl.array_size = 0;
 	}
 
 	share->use_count++;
@@ -8492,6 +8733,11 @@ static void free_share(INNOBASE_SHARE* share)
 		HASH_DELETE(INNOBASE_SHARE, table_name_hash,
 			    innobase_open_tables, fold, share);
 		thr_lock_delete(&share->lock);
+
+		/* Free any memory from index translation table */
+		my_free(share->idx_trans_tbl.index_mapping,
+			MYF(MY_ALLOW_ZERO_PTR));
+
 		my_free(share, MYF(0));
 
 		/* TODO: invoke HASH_MIGRATE if innobase_open_tables
diff --git a/storage/innodb_plugin/handler/ha_innodb.h b/storage/innodb_plugin/handler/ha_innodb.h
index 0e366a1eb2c..5dd9726ae4c 100644
--- a/storage/innodb_plugin/handler/ha_innodb.h
+++ b/storage/innodb_plugin/handler/ha_innodb.h
@@ -27,15 +27,31 @@ Place, Suite 330, Boston, MA 02111-1307 USA
 #pragma interface			/* gcc class implementation */
 #endif
 
+/* Structure defines translation table between mysql index and innodb
+index structures */
+typedef struct innodb_idx_translate_struct {
+	ulint		index_count;	/*!< number of valid index entries
+					in the index_mapping array */
+	ulint		array_size;	/*!< array size of index_mapping */
+	dict_index_t**	index_mapping;	/*!< index pointer array directly
+					maps to index in Innodb from MySQL
+					array index */
+} innodb_idx_translate_t;
+
+
 /** InnoDB table share */
 typedef struct st_innobase_share {
-	THR_LOCK	lock;		/*!< MySQL lock protecting
-					this structure */
-	const char*	table_name;	/*!< InnoDB table name */
-	uint		use_count;	/*!< reference count,
-					incremented in get_share()
-					and decremented in free_share() */
-	void*		table_name_hash;/*!< hash table chain node */
+	THR_LOCK		lock;		/*!< MySQL lock protecting
+						this structure */
+	const char*		table_name;	/*!< InnoDB table name */
+	uint			use_count;	/*!< reference count,
+						incremented in get_share()
+						and decremented in
+						free_share() */
+	void*			table_name_hash;/*!< hash table chain node */
+	innodb_idx_translate_t	idx_trans_tbl;	/*!< index translation
+						table between MySQL and
+						Innodb */
 } INNOBASE_SHARE;
 
 
diff --git a/storage/innodb_plugin/handler/handler0alter.cc b/storage/innodb_plugin/handler/handler0alter.cc
index a5008991400..47999ae37f8 100644
--- a/storage/innodb_plugin/handler/handler0alter.cc
+++ b/storage/innodb_plugin/handler/handler0alter.cc
@@ -764,6 +764,10 @@ err_exit:
 
 	ut_ad(error == DB_SUCCESS);
 
+	/* We will need to rebuild index translation table. Set
+	valid index entry count in the translation table to zero */
+	share->idx_trans_tbl.index_count = 0;
+
 	/* Commit the data dictionary transaction in order to release
 	the table locks on the system tables.  This means that if
 	MySQL crashes while creating a new primary key inside
@@ -1198,6 +1202,10 @@ ha_innobase::final_drop_index(
 		ut_a(!index->to_be_dropped);
 	}
 
+	/* We will need to rebuild index translation table. Set
+	valid index entry count in the translation table to zero */
+	share->idx_trans_tbl.index_count = 0;
+
 #ifdef UNIV_DEBUG
 	dict_table_check_for_dup_indexes(prebuilt->table);
 #endif

From e4bae32de794430a64002ec9b0d64ff5e87f9a3d Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:31:17 +0400
Subject: [PATCH 035/116] Applying InnoDB snapshot

Detailed revision comments:

r6527 | vasil | 2010-01-29 14:39:48 +0200 (Fri, 29 Jan 2010) | 6 lines
branches/zip:

Extend the comment about row_mysql_handle_errors().

Suggested by:	Heikki
---
 storage/innodb_plugin/row/row0mysql.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/row/row0mysql.c b/storage/innodb_plugin/row/row0mysql.c
index e7a44d10b61..cf0a632aee8 100644
--- a/storage/innodb_plugin/row/row0mysql.c
+++ b/storage/innodb_plugin/row/row0mysql.c
@@ -485,7 +485,7 @@ next_column:
 /****************************************************************//**
 Handles user errors and lock waits detected by the database engine.
 @return TRUE if it was a lock wait and we should continue running the
-query thread */
+query thread and in that case the thr is ALREADY in the running state. */
 UNIV_INTERN
 ibool
 row_mysql_handle_errors(

From ed2765d948572687e2add49f80cb2ffcea56fd66 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:31:37 +0400
Subject: [PATCH 036/116] Applying InnoDB snapshot

Detailed revision comments:

r6533 | calvin | 2010-01-29 23:31:59 +0200 (Fri, 29 Jan 2010) | 4 lines
branches/zip: remove duplicated copyright and license info.
---
 storage/innodb_plugin/handler/ha_innodb.cc | 34 +++++-----------------
 storage/innodb_plugin/include/log0log.h    | 17 -----------
 storage/innodb_plugin/include/os0file.h    | 17 -----------
 storage/innodb_plugin/include/srv0srv.h    | 34 +++++-----------------
 storage/innodb_plugin/log/log0log.c        | 17 -----------
 storage/innodb_plugin/os/os0file.c         | 17 -----------
 storage/innodb_plugin/srv/srv0srv.c        | 34 +++++-----------------
 storage/innodb_plugin/srv/srv0start.c      | 34 +++++-----------------
 8 files changed, 32 insertions(+), 172 deletions(-)

diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 9219bd2acc2..9ca2471a68f 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -2,6 +2,7 @@
 
 Copyright (c) 2000, 2009, MySQL AB & Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, 2009 Google Inc.
+Copyright (c) 2009, Percona Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
 Google, Inc. Those modifications are gratefully acknowledged and are described
@@ -9,6 +10,13 @@ briefly in the InnoDB documentation. The contributions by Google are
 incorporated with their permission, and subject to the conditions contained in
 the file COPYING.Google.
 
+Portions of this file contain modifications contributed and copyrighted
+by Percona Inc.. Those modifications are
+gratefully acknowledged and are described briefly in the InnoDB
+documentation. The contributions by Percona Inc. are incorporated with
+their permission, and subject to the conditions contained in the file
+COPYING.Percona.
+
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
 Foundation; version 2 of the License.
@@ -22,32 +30,6 @@ this program; if not, write to the Free Software Foundation, Inc., 59 Temple
 Place, Suite 330, Boston, MA 02111-1307 USA
 
 *****************************************************************************/
-/***********************************************************************
-
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
-Copyright (c) 2009, Percona Inc.
-
-Portions of this file contain modifications contributed and copyrighted
-by Percona Inc.. Those modifications are
-gratefully acknowledged and are described briefly in the InnoDB
-documentation. The contributions by Percona Inc. are incorporated with
-their permission, and subject to the conditions contained in the file
-COPYING.Percona.
-
-This program is free software; you can redistribute it and/or modify it
-under the terms of the GNU General Public License as published by the
-Free Software Foundation; version 2 of the License.
-
-This program is distributed in the hope that it will be useful, but
-WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
-Public License for more details.
-
-You should have received a copy of the GNU General Public License along
-with this program; if not, write to the Free Software Foundation, Inc.,
-59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
-
-***********************************************************************/
 
 /* TODO list for the InnoDB handler in 5.0:
   - Remove the flag trx->active_trans and look at trx->conc_state
diff --git a/storage/innodb_plugin/include/log0log.h b/storage/innodb_plugin/include/log0log.h
index 233714eb63a..7fc24b58e4c 100644
--- a/storage/innodb_plugin/include/log0log.h
+++ b/storage/innodb_plugin/include/log0log.h
@@ -1,22 +1,5 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
-
-This program is free software; you can redistribute it and/or modify it under
-the terms of the GNU General Public License as published by the Free Software
-Foundation; version 2 of the License.
-
-This program is distributed in the hope that it will be useful, but WITHOUT
-ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
-
-*****************************************************************************/
-/*****************************************************************************
-
 Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
 Copyright (c) 2009, Google Inc.
 
diff --git a/storage/innodb_plugin/include/os0file.h b/storage/innodb_plugin/include/os0file.h
index 16568579f31..ad31b37a2a1 100644
--- a/storage/innodb_plugin/include/os0file.h
+++ b/storage/innodb_plugin/include/os0file.h
@@ -1,20 +1,3 @@
-/*****************************************************************************
-
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
-
-This program is free software; you can redistribute it and/or modify it under
-the terms of the GNU General Public License as published by the Free Software
-Foundation; version 2 of the License.
-
-This program is distributed in the hope that it will be useful, but WITHOUT
-ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
-
-*****************************************************************************/
 /***********************************************************************
 
 Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
diff --git a/storage/innodb_plugin/include/srv0srv.h b/storage/innodb_plugin/include/srv0srv.h
index 9001d0988cf..8700cc6d66e 100644
--- a/storage/innodb_plugin/include/srv0srv.h
+++ b/storage/innodb_plugin/include/srv0srv.h
@@ -2,6 +2,7 @@
 
 Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, 2009, Google Inc.
+Copyright (c) 2009, Percona Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
 Google, Inc. Those modifications are gratefully acknowledged and are described
@@ -9,6 +10,13 @@ briefly in the InnoDB documentation. The contributions by Google are
 incorporated with their permission, and subject to the conditions contained in
 the file COPYING.Google.
 
+Portions of this file contain modifications contributed and copyrighted
+by Percona Inc.. Those modifications are
+gratefully acknowledged and are described briefly in the InnoDB
+documentation. The contributions by Percona Inc. are incorporated with
+their permission, and subject to the conditions contained in the file
+COPYING.Percona.
+
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
 Foundation; version 2 of the License.
@@ -22,32 +30,6 @@ this program; if not, write to the Free Software Foundation, Inc., 59 Temple
 Place, Suite 330, Boston, MA 02111-1307 USA
 
 *****************************************************************************/
-/***********************************************************************
-
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
-Copyright (c) 2009, Percona Inc.
-
-Portions of this file contain modifications contributed and copyrighted
-by Percona Inc.. Those modifications are
-gratefully acknowledged and are described briefly in the InnoDB
-documentation. The contributions by Percona Inc. are incorporated with
-their permission, and subject to the conditions contained in the file
-COPYING.Percona.
-
-This program is free software; you can redistribute it and/or modify it
-under the terms of the GNU General Public License as published by the
-Free Software Foundation; version 2 of the License.
-
-This program is distributed in the hope that it will be useful, but
-WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
-Public License for more details.
-
-You should have received a copy of the GNU General Public License along
-with this program; if not, write to the Free Software Foundation, Inc.,
-59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
-
-***********************************************************************/
 
 /**************************************************//**
 @file include/srv0srv.h
diff --git a/storage/innodb_plugin/log/log0log.c b/storage/innodb_plugin/log/log0log.c
index 063581055e9..6ddbcd22f77 100644
--- a/storage/innodb_plugin/log/log0log.c
+++ b/storage/innodb_plugin/log/log0log.c
@@ -1,22 +1,5 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
-
-This program is free software; you can redistribute it and/or modify it under
-the terms of the GNU General Public License as published by the Free Software
-Foundation; version 2 of the License.
-
-This program is distributed in the hope that it will be useful, but WITHOUT
-ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
-
-*****************************************************************************/
-/*****************************************************************************
-
 Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
 Copyright (c) 2009, Google Inc.
 
diff --git a/storage/innodb_plugin/os/os0file.c b/storage/innodb_plugin/os/os0file.c
index 37edad442db..2bb652980b0 100644
--- a/storage/innodb_plugin/os/os0file.c
+++ b/storage/innodb_plugin/os/os0file.c
@@ -1,20 +1,3 @@
-/*****************************************************************************
-
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
-
-This program is free software; you can redistribute it and/or modify it under
-the terms of the GNU General Public License as published by the Free Software
-Foundation; version 2 of the License.
-
-This program is distributed in the hope that it will be useful, but WITHOUT
-ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
-FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License along with
-this program; if not, write to the Free Software Foundation, Inc., 59 Temple
-Place, Suite 330, Boston, MA 02111-1307 USA
-
-*****************************************************************************/
 /***********************************************************************
 
 Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
diff --git a/storage/innodb_plugin/srv/srv0srv.c b/storage/innodb_plugin/srv/srv0srv.c
index 7252929d3e2..23ec3b8db4c 100644
--- a/storage/innodb_plugin/srv/srv0srv.c
+++ b/storage/innodb_plugin/srv/srv0srv.c
@@ -2,6 +2,7 @@
 
 Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, 2009 Google Inc.
+Copyright (c) 2009, Percona Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
 Google, Inc. Those modifications are gratefully acknowledged and are described
@@ -9,6 +10,13 @@ briefly in the InnoDB documentation. The contributions by Google are
 incorporated with their permission, and subject to the conditions contained in
 the file COPYING.Google.
 
+Portions of this file contain modifications contributed and copyrighted
+by Percona Inc.. Those modifications are
+gratefully acknowledged and are described briefly in the InnoDB
+documentation. The contributions by Percona Inc. are incorporated with
+their permission, and subject to the conditions contained in the file
+COPYING.Percona.
+
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
 Foundation; version 2 of the License.
@@ -22,32 +30,6 @@ this program; if not, write to the Free Software Foundation, Inc., 59 Temple
 Place, Suite 330, Boston, MA 02111-1307 USA
 
 *****************************************************************************/
-/***********************************************************************
-
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
-Copyright (c) 2009, Percona Inc.
-
-Portions of this file contain modifications contributed and copyrighted
-by Percona Inc.. Those modifications are
-gratefully acknowledged and are described briefly in the InnoDB
-documentation. The contributions by Percona Inc. are incorporated with
-their permission, and subject to the conditions contained in the file
-COPYING.Percona.
-
-This program is free software; you can redistribute it and/or modify it
-under the terms of the GNU General Public License as published by the
-Free Software Foundation; version 2 of the License.
-
-This program is distributed in the hope that it will be useful, but
-WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
-Public License for more details.
-
-You should have received a copy of the GNU General Public License along
-with this program; if not, write to the Free Software Foundation, Inc.,
-59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
-
-***********************************************************************/
 
 /**************************************************//**
 @file srv/srv0srv.c
diff --git a/storage/innodb_plugin/srv/srv0start.c b/storage/innodb_plugin/srv/srv0start.c
index cab193615df..71e84d1c185 100644
--- a/storage/innodb_plugin/srv/srv0start.c
+++ b/storage/innodb_plugin/srv/srv0start.c
@@ -2,6 +2,7 @@
 
 Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
+Copyright (c) 2009, Percona Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
 Google, Inc. Those modifications are gratefully acknowledged and are described
@@ -9,6 +10,13 @@ briefly in the InnoDB documentation. The contributions by Google are
 incorporated with their permission, and subject to the conditions contained in
 the file COPYING.Google.
 
+Portions of this file contain modifications contributed and copyrighted
+by Percona Inc.. Those modifications are
+gratefully acknowledged and are described briefly in the InnoDB
+documentation. The contributions by Percona Inc. are incorporated with
+their permission, and subject to the conditions contained in the file
+COPYING.Percona.
+
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
 Foundation; version 2 of the License.
@@ -22,32 +30,6 @@ this program; if not, write to the Free Software Foundation, Inc., 59 Temple
 Place, Suite 330, Boston, MA 02111-1307 USA
 
 *****************************************************************************/
-/***********************************************************************
-
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
-Copyright (c) 2009, Percona Inc.
-
-Portions of this file contain modifications contributed and copyrighted
-by Percona Inc.. Those modifications are
-gratefully acknowledged and are described briefly in the InnoDB
-documentation. The contributions by Percona Inc. are incorporated with
-their permission, and subject to the conditions contained in the file
-COPYING.Percona.
-
-This program is free software; you can redistribute it and/or modify it
-under the terms of the GNU General Public License as published by the
-Free Software Foundation; version 2 of the License.
-
-This program is distributed in the hope that it will be useful, but
-WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
-Public License for more details.
-
-You should have received a copy of the GNU General Public License along
-with this program; if not, write to the Free Software Foundation, Inc.,
-59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
-
-***********************************************************************/
 
 /********************************************************************//**
 @file srv/srv0start.c

From b419c7523043983f26de2996d54e698a532b7480 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:32:13 +0400
Subject: [PATCH 037/116] Applying InnoDB snapshot, fixes BUG#49047.

Detailed revision comments:

r6534 | sunny | 2010-01-29 23:42:49 +0200 (Fri, 29 Jan 2010) | 15 lines
branches/zip: Two changes to fix the problem:

1. First scan the joining transaction's locks and check if no other
transaction is waiting for a lock held by the joining transaction.
If no other transaction is waiting then  no deadlock an occur and
we avoid doing an exhaustive search.

2. Change the direction of the lock traversal from backward to forward.
Previously we traversed backward from the lock that has to wait, the function
to that fetched the previous node was very inefficient resulting in O(n^2)
access to the rec lock list.

Fix Bug #49047 InnoDB deadlock detection is CPU intensive with many locks on a single row.

rb://218
---
 storage/innodb_plugin/include/trx0trx.h |   8 +-
 storage/innodb_plugin/lock/lock0lock.c  | 132 ++++++++++++++++++++----
 2 files changed, 117 insertions(+), 23 deletions(-)

diff --git a/storage/innodb_plugin/include/trx0trx.h b/storage/innodb_plugin/include/trx0trx.h
index 5f2c1246f37..e2ad1f2f722 100644
--- a/storage/innodb_plugin/include/trx0trx.h
+++ b/storage/innodb_plugin/include/trx0trx.h
@@ -526,6 +526,9 @@ struct trx_struct{
 					/* 0, RW_S_LATCH, or RW_X_LATCH:
 					the latch mode trx currently holds
 					on dict_operation_lock */
+	unsigned	deadlock_mark:1;/*!< a mark field used in deadlock
+					checking algorithm. Always protected
+					by the kernel_mutex. */
 	time_t		start_time;	/*!< time the trx object was created
 					or the state last time became
 					TRX_ACTIVE */
@@ -640,11 +643,6 @@ struct trx_struct{
 			wait_thrs;	/*!< query threads belonging to this
 					trx that are in the QUE_THR_LOCK_WAIT
 					state */
-	ulint		deadlock_mark;	/*!< a mark field used in deadlock
-					checking algorithm.  This must be
-					in its own machine word, because
-					it can be changed by other
-					threads while holding kernel_mutex. */
 	/*------------------------------*/
 	mem_heap_t*	lock_heap;	/*!< memory heap for the locks of the
 					transaction */
diff --git a/storage/innodb_plugin/lock/lock0lock.c b/storage/innodb_plugin/lock/lock0lock.c
index 0fa67d31716..4552c1e7cde 100644
--- a/storage/innodb_plugin/lock/lock0lock.c
+++ b/storage/innodb_plugin/lock/lock0lock.c
@@ -401,7 +401,7 @@ lock_deadlock_recursive(
 /*====================*/
 	trx_t*	start,		/*!< in: recursion starting point */
 	trx_t*	trx,		/*!< in: a transaction waiting for a lock */
-	lock_t*	wait_lock,	/*!< in: the lock trx is waiting to be granted */
+	lock_t*	wait_lock,	/*!< in:  lock that is waiting to be granted */
 	ulint*	cost,		/*!< in/out: number of calculation steps thus
 				far: if this exceeds LOCK_MAX_N_STEPS_...
 				we return LOCK_VICTIM_IS_START */
@@ -411,7 +411,7 @@ lock_deadlock_recursive(
 
 /*********************************************************************//**
 Gets the nth bit of a record lock.
-@return	TRUE if bit set */
+@return	TRUE if bit set also if i == ULINT_UNDEFINED return FALSE*/
 UNIV_INLINE
 ibool
 lock_rec_get_nth_bit(
@@ -1222,7 +1222,7 @@ lock_rec_get_first_on_page(
 
 /*********************************************************************//**
 Gets the next explicit lock request on a record.
-@return	next lock, NULL if none exists */
+@return	next lock, NULL if none exists or if heap_no == ULINT_UNDEFINED */
 UNIV_INLINE
 lock_t*
 lock_rec_get_next(
@@ -3311,6 +3311,64 @@ retry:
 	return(FALSE);
 }
 
+/********************************************************************//**
+Check that no other transaction is waiting on this transaction's locks.
+@return TRUE if some other transaction is waiting for this lock. */
+static
+ulint
+lock_trx_has_no_waiters(
+/*====================*/
+	const trx_t*	trx)		/*!< in: the transaction to check */
+{
+	const lock_t*	lock;
+
+	ut_ad(mutex_own(&kernel_mutex));
+
+	for (lock = UT_LIST_GET_FIRST(trx->trx_locks);
+	     lock != NULL;
+	     lock = UT_LIST_GET_NEXT(trx_locks, lock)) {
+
+		const lock_t*	wait_lock = lock;
+
+		/* Look for all transactions that could be waiting on this
+		transaction's locks. For that we need to search forward. */
+		if (lock_get_type_low(lock) == LOCK_REC) {
+
+			ulint	heap_no;
+
+			/* It's possible for heap_no to be undefined here.
+			This can happen during lock move from one page to
+			another when we split. */
+
+			heap_no = lock_rec_find_set_bit(lock);
+
+			do {
+				wait_lock = lock_rec_get_next(
+					heap_no, (lock_t*) wait_lock);
+
+				if (wait_lock != NULL
+				    && lock_has_to_wait(wait_lock, lock)) {
+
+					return(FALSE);
+				}
+			} while (wait_lock != NULL);
+		} else {
+			do {
+				wait_lock = UT_LIST_GET_NEXT(
+					un_member.tab_lock.locks, wait_lock);
+
+				if (wait_lock != NULL
+				    && lock_has_to_wait(wait_lock, lock) ) {
+
+					return(FALSE);
+				}
+			} while (wait_lock != NULL);
+		}
+	}
+
+	return(TRUE);
+}
+
 /********************************************************************//**
 Looks recursively for a deadlock.
 @return 0 if no deadlock found, LOCK_VICTIM_IS_START if there was a
@@ -3324,7 +3382,7 @@ lock_deadlock_recursive(
 /*====================*/
 	trx_t*	start,		/*!< in: recursion starting point */
 	trx_t*	trx,		/*!< in: a transaction waiting for a lock */
-	lock_t*	wait_lock,	/*!< in: the lock trx is waiting to be granted */
+	lock_t*	wait_lock,	/*!< in: lock that is waiting to be granted */
 	ulint*	cost,		/*!< in/out: number of calculation steps thus
 				far: if this exceeds LOCK_MAX_N_STEPS_...
 				we return LOCK_VICTIM_IS_START */
@@ -3332,10 +3390,10 @@ lock_deadlock_recursive(
 				LOCK_MAX_DEPTH_IN_DEADLOCK_CHECK, we
 				return LOCK_VICTIM_IS_START */
 {
-	lock_t*	lock;
-	ulint	bit_no		= ULINT_UNDEFINED;
-	trx_t*	lock_trx;
 	ulint	ret;
+	lock_t*	lock;
+	trx_t*	lock_trx;
+	ulint	heap_no		= ULINT_UNDEFINED;
 
 	ut_a(trx);
 	ut_a(start);
@@ -3346,32 +3404,54 @@ lock_deadlock_recursive(
 		/* We have already exhaustively searched the subtree starting
 		from this trx */
 
+		return(0);
+	} else if (lock_trx_has_no_waiters(trx)) {
+		/* If no other transaction is waiting for this transaction
+		to release its locks then no deadlock can occur. */
 		return(0);
 	}
 
 	*cost = *cost + 1;
 
-	lock = wait_lock;
 
 	if (lock_get_type_low(wait_lock) == LOCK_REC) {
+		ulint		space;
+		ulint		page_no;
 
-		bit_no = lock_rec_find_set_bit(wait_lock);
+		heap_no = lock_rec_find_set_bit(wait_lock);
+		ut_a(heap_no != ULINT_UNDEFINED);
 
-		ut_a(bit_no != ULINT_UNDEFINED);
+		space = wait_lock->un_member.rec_lock.space;
+		page_no = wait_lock->un_member.rec_lock.page_no;
+
+		lock = lock_rec_get_first_on_page_addr(space, page_no);
+
+		/* Position the iterator on the first matching record lock. */
+		while (lock != NULL
+		       && lock != wait_lock
+		       && !lock_rec_get_nth_bit(lock, heap_no)) {
+
+			lock = lock_rec_get_next_on_page(lock);
+		}
+
+		if (lock == wait_lock) {
+			lock = NULL;
+		}
+
+		ut_ad(lock == NULL || lock_rec_get_nth_bit(lock, heap_no));
+
+	} else {
+		lock = wait_lock;
 	}
 
 	/* Look at the locks ahead of wait_lock in the lock queue */
 
 	for (;;) {
-		if (lock_get_type_low(lock) & LOCK_TABLE) {
+		/* Get previous table lock. */
+		if (heap_no == ULINT_UNDEFINED) {
 
-			lock = UT_LIST_GET_PREV(un_member.tab_lock.locks,
-						lock);
-		} else {
-			ut_ad(lock_get_type_low(lock) == LOCK_REC);
-			ut_a(bit_no != ULINT_UNDEFINED);
-
-			lock = (lock_t*) lock_rec_get_prev(lock, bit_no);
+			lock = UT_LIST_GET_PREV(
+				un_member.tab_lock.locks, lock);
 		}
 
 		if (lock == NULL) {
@@ -3493,12 +3573,28 @@ lock_deadlock_recursive(
 				ret = lock_deadlock_recursive(
 					start, lock_trx,
 					lock_trx->wait_lock, cost, depth + 1);
+
 				if (ret != 0) {
 
 					return(ret);
 				}
 			}
 		}
+		/* Get the next record lock to check. */
+		if (heap_no != ULINT_UNDEFINED) {
+
+			ut_a(lock != NULL);
+
+			do {
+				lock = lock_rec_get_next_on_page(lock);
+			} while (lock != NULL
+				&& lock != wait_lock
+				&& !lock_rec_get_nth_bit(lock, heap_no));
+
+			if (lock == wait_lock) {
+				lock = NULL;
+			}
+		}
 	}/* end of the 'for (;;)'-loop */
 }
 

From 22864b86b2bdefb411decb8b55fc56cc6aa36065 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:34:53 +0400
Subject: [PATCH 038/116] Applying InnoDB snapshot

Detailed revision comments:

r6539 | marko | 2010-02-01 11:31:12 +0200 (Mon, 01 Feb 2010) | 75 lines
branches/zip: Merge revisions 6471:6538 from branches/5.1:

  ------------------------------------------------------------------------
  r6488 | sunny | 2010-01-21 02:55:08 +0200 (Thu, 21 Jan 2010) | 2 lines
  Changed paths:
     M /branches/5.1/mysql-test/innodb-autoinc.result
     M /branches/5.1/mysql-test/innodb-autoinc.test

  branches/5.1: Factor out test for bug#44030 from innodb-autoinc.test
  into a separate test/result files.
  ------------------------------------------------------------------------
  r6489 | sunny | 2010-01-21 02:57:50 +0200 (Thu, 21 Jan 2010) | 2 lines
  Changed paths:
     A /branches/5.1/mysql-test/innodb-autoinc-44030.result
     A /branches/5.1/mysql-test/innodb-autoinc-44030.test

  branches/5.1: Factor out test for bug#44030 from innodb-autoinc.test
  into a separate test/result files.
  ------------------------------------------------------------------------
  r6492 | sunny | 2010-01-21 09:38:35 +0200 (Thu, 21 Jan 2010) | 1 line
  Changed paths:
     M /branches/5.1/mysql-test/innodb-autoinc-44030.test

  branches/5.1: Add reference to bug#47621 in the comment.
  ------------------------------------------------------------------------
  r6535 | sunny | 2010-01-30 00:08:40 +0200 (Sat, 30 Jan 2010) | 11 lines
  Changed paths:
     M /branches/5.1/handler/ha_innodb.cc

  branches/5.1: Undo the change from r6424. We need to return DB_SUCCESS even
  if we were unable to initialize the tabe autoinc value. This is required for
  the open to succeed. The only condition we currently treat as a hard error
  is if the autoinc field instance passed in by MySQL is NULL.

  Previously if the table autoinc value was 0 and the next value was requested
  we had an assertion that would fail. Change that assertion and treat a value
  of 0 to mean that the autoinc system is unavailable. Generation of next
  value will now return failure.

  rb://237
  ------------------------------------------------------------------------
  r6536 | sunny | 2010-01-30 00:13:42 +0200 (Sat, 30 Jan 2010) | 6 lines
  Changed paths:
     M /branches/5.1/handler/ha_innodb.cc
     M /branches/5.1/mysql-test/innodb-autoinc.result
     M /branches/5.1/mysql-test/innodb-autoinc.test

  branches/5.1: Check *first_value everytime against the column max
  value and  set *first_value to next autoinc if it's > col max value.
  ie.  not rely on what is passed in from MySQL.

  [49497] Error 1467 (ER_AUTOINC_READ_FAILED) on inserting a negative value
  rb://236
  ------------------------------------------------------------------------
  r6537 | sunny | 2010-01-30 00:35:00 +0200 (Sat, 30 Jan 2010) | 2 lines
  Changed paths:
     M /branches/5.1/handler/ha_innodb.cc
     M /branches/5.1/mysql-test/innodb-autoinc.result
     M /branches/5.1/mysql-test/innodb-autoinc.test

  branches/5.1: Undo r6536.
  ------------------------------------------------------------------------
  r6538 | sunny | 2010-01-30 00:43:06 +0200 (Sat, 30 Jan 2010) | 6 lines
  Changed paths:
     M /branches/5.1/handler/ha_innodb.cc
     M /branches/5.1/mysql-test/innodb-autoinc.result
     M /branches/5.1/mysql-test/innodb-autoinc.test

  branches/5.1: Check *first_value every time against the column max
  value and  set *first_value to next autoinc if it's > col max value.
  ie.  not rely on what is passed in from MySQL.

  [49497] Error 1467 (ER_AUTOINC_READ_FAILED) on inserting a negative value
  rb://236
  ------------------------------------------------------------------------
---
 storage/innodb_plugin/handler/ha_innodb.cc | 54 +++++++++++++++-------
 1 file changed, 38 insertions(+), 16 deletions(-)

diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 9ca2471a68f..d2aab9e31ad 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -3266,9 +3266,9 @@ ha_innobase::innobase_initialize_autoinc()
 		auto_inc = innobase_get_int_col_max_value(field);
 	} else {
 		/* We have no idea what's been passed in to us as the
-		autoinc column. We set it to the MAX_INT of our table
-		autoinc type. */
-		auto_inc = 0xFFFFFFFFFFFFFFFFULL;
+		autoinc column. We set it to the 0, effectively disabling
+		updates to the table. */
+		auto_inc = 0;
 
 		ut_print_timestamp(stderr);
 		fprintf(stderr, "  InnoDB: Unable to determine the AUTOINC "
@@ -3277,7 +3277,7 @@ ha_innobase::innobase_initialize_autoinc()
 
 	if (srv_force_recovery >= SRV_FORCE_NO_IBUF_MERGE) {
 		/* If the recovery level is set so high that writes
-		are disabled we force the AUTOINC counter to the MAX
+		are disabled we force the AUTOINC counter to 0
 		value effectively disabling writes to the table.
 		Secondly, we avoid reading the table in case the read
 		results in failure due to a corrupted table/index.
@@ -3286,7 +3286,10 @@ ha_innobase::innobase_initialize_autoinc()
 		tables can be dumped with minimal hassle.  If an error
 		were returned in this case, the first attempt to read
 		the table would fail and subsequent SELECTs would succeed. */
+		auto_inc = 0;
 	} else if (field == NULL) {
+		/* This is a far more serious error, best to avoid
+		opening the table and return failure. */
 		my_error(ER_AUTOINC_READ_FAILED, MYF(0));
 	} else {
 		dict_index_t*	index;
@@ -3315,7 +3318,7 @@ ha_innobase::innobase_initialize_autoinc()
 				"InnoDB: Unable to find the AUTOINC column "
 				"%s in the InnoDB table %s.\n"
 				"InnoDB: We set the next AUTOINC column "
-				"value to the maximum possible value,\n"
+				"value to 0,\n"
 				"InnoDB: in effect disabling the AUTOINC "
 				"next value generation.\n"
 				"InnoDB: You can either set the next "
@@ -3324,7 +3327,13 @@ ha_innobase::innobase_initialize_autoinc()
 				"recreating the table.\n",
 				col_name, index->table->name);
 
-			my_error(ER_AUTOINC_READ_FAILED, MYF(0));
+			/* This will disable the AUTOINC generation. */
+			auto_inc = 0;
+
+			/* We want the open to succeed, so that the user can
+			take corrective action. ie. reads should succeed but
+			updates should fail. */
+			err = DB_SUCCESS;
 			break;
 		default:
 			/* row_search_max_autoinc() should only return
@@ -4594,11 +4603,17 @@ no_commit:
 		prebuilt->autoinc_error = DB_SUCCESS;
 
 		if ((error = update_auto_increment())) {
-
 			/* We don't want to mask autoinc overflow errors. */
-			if (prebuilt->autoinc_error != DB_SUCCESS) {
-				error = (int) prebuilt->autoinc_error;
 
+			/* Handle the case where the AUTOINC sub-system
+			failed during initialization. */
+			if (prebuilt->autoinc_error == DB_UNSUPPORTED) {
+				error_result = ER_AUTOINC_READ_FAILED;
+				/* Set the error message to report too. */
+				my_error(ER_AUTOINC_READ_FAILED, MYF(0));
+				goto func_exit;
+			} else if (prebuilt->autoinc_error != DB_SUCCESS) {
+				error = (int) prebuilt->autoinc_error;
 				goto report_error;
 			}
 
@@ -8954,7 +8969,10 @@ ha_innobase::innobase_get_autoinc(
 		*value = dict_table_autoinc_read(prebuilt->table);
 
 		/* It should have been initialized during open. */
-		ut_a(*value != 0);
+		if (*value == 0) {
+			prebuilt->autoinc_error = DB_UNSUPPORTED;
+			dict_table_autoinc_unlock(prebuilt->table);
+		}
 	}
 
 	return(prebuilt->autoinc_error);
@@ -9034,6 +9052,11 @@ ha_innobase::get_auto_increment(
 	invoking this method. So we are not sure if it's guaranteed to
 	be 0 or not. */
 
+	/* We need the upper limit of the col type to check for
+	whether we update the table autoinc counter or not. */
+	ulonglong	col_max_value = innobase_get_int_col_max_value(
+		table->next_number_field);
+
 	/* Called for the first time ? */
 	if (trx->n_autoinc_rows == 0) {
 
@@ -9050,6 +9073,11 @@ ha_innobase::get_auto_increment(
 	/* Not in the middle of a mult-row INSERT. */
 	} else if (prebuilt->autoinc_last_value == 0) {
 		set_if_bigger(*first_value, autoinc);
+	/* Check for -ve values. */
+	} else if (*first_value > col_max_value && trx->n_autoinc_rows > 0) {
+		/* Set to next logical value. */
+		ut_a(autoinc > trx->n_autoinc_rows);
+		*first_value = (autoinc - trx->n_autoinc_rows) - 1;
 	}
 
 	*nb_reserved_values = trx->n_autoinc_rows;
@@ -9060,12 +9088,6 @@ ha_innobase::get_auto_increment(
 		ulonglong	need;
 		ulonglong	current;
 		ulonglong	next_value;
-		ulonglong	col_max_value;
-
-		/* We need the upper limit of the col type to check for
-		whether we update the table autoinc counter or not. */
-		col_max_value = innobase_get_int_col_max_value(
-			table->next_number_field);
 
 		current = *first_value > col_max_value ? autoinc : *first_value;
 		need = *nb_reserved_values * increment;

From 4f02b2b5a2fd4d5f0a0e9dc70fe8a905e7bbea5f Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:35:30 +0400
Subject: [PATCH 039/116] Applying InnoDB snapshot

Detailed revision comments:

r6540 | marko | 2010-02-01 11:35:13 +0200 (Mon, 01 Feb 2010) | 1 line
branches/zip: ChangeLog: Document the merge of 6471:6538 from branches/5.1.
---
 storage/innodb_plugin/ChangeLog | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 85bcaf9dcf7..81bc0310139 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,12 @@
+2010-02-01	The InnoDB Team
+
+	* handler/ha_innodb.cc, mysql-test/innodb-autoinc.test,
+	mysql-test/innodb-autoinc.result,
+	mysql-test/innodb-autoinc-44030.test,
+	mysql-test/innodb-autoinc-44030.result:
+	Fix Bug#49497 Error 1467 (ER_AUTOINC_READ_FAILED) on inserting
+	a negative value
+
 2010-01-27	The InnoDB Team
 
 	* include/row0mysql.h, log/log0recv.c, row/row0mysql.c:

From bce33719cce5dd9c65724f5008a2565c5c5ba3fb Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:36:02 +0400
Subject: [PATCH 040/116] Applying InnoDB snapshot

Detailed revision comments:

r6546 | jyang | 2010-02-03 11:05:24 +0200 (Wed, 03 Feb 2010) | 9 lines
branches/zip: Relax assertion on the number of index defined
in InnoDB must be comparable with that of MySQL to tolerate
possible dictionary inconsistency. Fix Mantis issue #455,
"UNIV_DEBUG+ assert ha_innodb.cc:3152 ib_num_index >=
mysql_num_index".

rb://248 Approved by Marko.
---
 storage/innodb_plugin/handler/ha_innodb.cc | 41 +++++++++++++++-------
 1 file changed, 29 insertions(+), 12 deletions(-)

diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index d2aab9e31ad..4e59ce567fb 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -3148,11 +3148,17 @@ innobase_build_index_translation(
 	mysql_num_index = table->s->keys;
 	ib_num_index = UT_LIST_GET_LEN(ib_table->indexes);
 
-	/* Innodb could have own system indexes. */
-	ut_a(ib_num_index >= mysql_num_index);
-
 	index_mapping = share->idx_trans_tbl.index_mapping;
 
+	/* If there exists inconsistency between MySQL and InnoDB dictionary
+	(metadata) information, the number of index defined in MySQL
+	could exceed that in InnoDB, do not build index translation
+	table in such case */
+	if (UNIV_UNLIKELY(ib_num_index < mysql_num_index)) {
+		ret = FALSE;
+		goto func_exit;
+	}
+
 	/* If index entry count is non-zero, nothing has
 	changed since last update, directly return TRUE */
 	if (share->idx_trans_tbl.index_count) {
@@ -5407,11 +5413,18 @@ ha_innobase::innobase_get_index(
 		if (index) {
 			ut_a(ut_strcmp(index->name, key->name) == 0);
 		} else {
-			sql_print_error("InnoDB could not find index %s "
-					"key no %u for table %s through "
-					" its index translation table",
-					key ? key->name : "NULL", keynr,
-					prebuilt->table->name);
+			/* Can't find index with keynr in the translation
+			table. Only print message if the index translation
+			table exists */
+			if (share->idx_trans_tbl.index_mapping) {
+				sql_print_error("InnoDB could not find "
+						"index %s key no %u for "
+						"table %s through its "
+						"index translation table",
+						key ? key->name : "NULL",
+						keynr,
+						prebuilt->table->name);
+			}
 
 			index = dict_table_get_index_on_name(prebuilt->table,
 							     key->name);
@@ -7108,10 +7121,13 @@ ha_innobase::records_in_range(
 
 	index = innobase_get_index(keynr);
 
-	ut_a(ut_strcmp(index->name, key->name) == 0);
-
-	/* MySQL knows about this index and so we must be able to find it.*/
-	ut_a(index);
+	/* There exists possibility of not being able to find requested
+	index due to inconsistency between MySQL and InoDB dictionary info.
+	Necessary message should have been printed in innobase_get_index() */
+	if (UNIV_UNLIKELY(!index)) {
+		n_rows = HA_POS_ERROR;
+		goto func_exit;
+	}
 
 	heap = mem_heap_create(2 * (key->key_parts * sizeof(dfield_t)
 				    + sizeof(dtuple_t)));
@@ -7156,6 +7172,7 @@ ha_innobase::records_in_range(
 
 	mem_heap_free(heap);
 
+func_exit:
 	my_free(key_val_buff2, MYF(0));
 
 	prebuilt->trx->op_info = (char*)"";

From 101d6a09cb83d6ca81ea22f3c098d26cdc2f4e1f Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:36:27 +0400
Subject: [PATCH 041/116] Applying InnoDB snapshot

Detailed revision comments:

r6547 | marko | 2010-02-03 14:43:38 +0200 (Wed, 03 Feb 2010) | 14 lines
branches/zip: Clean up CHECK TABLE error handling. (Issue #220)

ha_innobase::change_active_index(): Clean up code formatting.

ha_innobase::check(): Incorporate the code from
row_check_table_for_mysql().  Report errors to the client connection
instead of writing them to the error log.

row_check_table_for_mysql(): Remove.

row_check_index_for_mysql(): Renamed from row_scan_and_check_index().
Let the caller initialize prebuilt, and assume that the index is usable.

rb://178 approved by Sunny Bains
---
 storage/innodb_plugin/ChangeLog            |   5 +
 storage/innodb_plugin/handler/ha_innodb.cc | 152 +++++++++++++++++++--
 storage/innodb_plugin/include/row0mysql.h  |  15 +-
 storage/innodb_plugin/row/row0mysql.c      | 150 ++------------------
 4 files changed, 163 insertions(+), 159 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 81bc0310139..7338bf5cbab 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,8 @@
+2010-02-03	The InnoDB Team
+
+	* handler/ha_innodb.cc, include/row0mysql.h, row/row0mysql.c:
+	Clean up CHECK TABLE error handling.
+
 2010-02-01	The InnoDB Team
 
 	* handler/ha_innodb.cc, mysql-test/innodb-autoinc.test,
diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 4e59ce567fb..44196e131a5 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -5489,7 +5489,7 @@ ha_innobase::change_active_index(
 	dtuple_set_n_fields(prebuilt->search_tuple, prebuilt->index->n_fields);
 
 	dict_index_copy_types(prebuilt->search_tuple, prebuilt->index,
-			prebuilt->index->n_fields);
+			      prebuilt->index->n_fields);
 
 	/* MySQL changes the active index for a handle also during some
 	queries, for example SELECT MAX(a), SUM(a) first retrieves the MAX()
@@ -7626,8 +7626,13 @@ ha_innobase::check(
 	HA_CHECK_OPT*	check_opt)	/*!< in: check options, currently
 					ignored */
 {
-	ulint		ret;
+	dict_index_t*	index;
+	ulint		n_rows;
+	ulint		n_rows_in_table	= ULINT_UNDEFINED;
+	ibool		is_ok		= TRUE;
+	ulint		old_isolation_level;
 
+	DBUG_ENTER("ha_innobase::check");
 	DBUG_ASSERT(thd == ha_thd());
 	ut_a(prebuilt->trx);
 	ut_a(prebuilt->trx->magic_n == TRX_MAGIC_N);
@@ -7640,17 +7645,140 @@ ha_innobase::check(
 		build_template(prebuilt, NULL, table, ROW_MYSQL_WHOLE_ROW);
 	}
 
-	ret = row_check_table_for_mysql(prebuilt);
-
-	switch (ret) {
-	case DB_SUCCESS:
-		return(HA_ADMIN_OK);
-	case DB_INTERRUPTED:
-		my_error(ER_QUERY_INTERRUPTED, MYF(0));
-		return(-1);
-	default:
-		return(HA_ADMIN_CORRUPT);
+	if (prebuilt->table->ibd_file_missing) {
+		sql_print_error("InnoDB: Error:\n"
+			"InnoDB: MySQL is trying to use a table handle"
+			" but the .ibd file for\n"
+			"InnoDB: table %s does not exist.\n"
+			"InnoDB: Have you deleted the .ibd file"
+			" from the database directory under\n"
+			"InnoDB: the MySQL datadir, or have you"
+			" used DISCARD TABLESPACE?\n"
+			"InnoDB: Please refer to\n"
+			"InnoDB: " REFMAN "innodb-troubleshooting.html\n"
+			"InnoDB: how you can resolve the problem.\n",
+			prebuilt->table->name);
+		DBUG_RETURN(HA_ADMIN_CORRUPT);
 	}
+
+	prebuilt->trx->op_info = "checking table";
+
+	old_isolation_level = prebuilt->trx->isolation_level;
+
+	/* We must run the index record counts at an isolation level
+	>= READ COMMITTED, because a dirty read can see a wrong number
+	of records in some index; to play safe, we use always
+	REPEATABLE READ here */
+
+	prebuilt->trx->isolation_level = TRX_ISO_REPEATABLE_READ;
+
+	/* Enlarge the fatal lock wait timeout during CHECK TABLE. */
+	mutex_enter(&kernel_mutex);
+	srv_fatal_semaphore_wait_threshold += 7200; /* 2 hours */
+	mutex_exit(&kernel_mutex);
+
+	for (index = dict_table_get_first_index(prebuilt->table);
+	     index != NULL;
+	     index = dict_table_get_next_index(index)) {
+#if 0
+		fputs("Validating index ", stderr);
+		ut_print_name(stderr, trx, FALSE, index->name);
+		putc('\n', stderr);
+#endif
+
+		if (!btr_validate_index(index, prebuilt->trx)) {
+			is_ok = FALSE;
+			push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
+					    ER_NOT_KEYFILE,
+					    "InnoDB: The B-tree of"
+					    " index '%-.200s' is corrupted.",
+					    index->name);
+			continue;
+		}
+
+		/* Instead of invoking change_active_index(), set up
+		a dummy template for non-locking reads, disabling
+		access to the clustered index. */
+		prebuilt->index = index;
+
+		prebuilt->index_usable = row_merge_is_index_usable(
+			prebuilt->trx, prebuilt->index);
+
+		if (UNIV_UNLIKELY(!prebuilt->index_usable)) {
+			push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
+					    HA_ERR_TABLE_DEF_CHANGED,
+					    "InnoDB: Insufficient history for"
+					    " index '%-.200s'",
+					    index->name);
+			continue;
+		}
+
+		prebuilt->sql_stat_start = TRUE;
+		prebuilt->template_type = ROW_MYSQL_DUMMY_TEMPLATE;
+		prebuilt->n_template = 0;
+		prebuilt->need_to_access_clustered = FALSE;
+
+		dtuple_set_n_fields(prebuilt->search_tuple, 0);
+
+		prebuilt->select_lock_type = LOCK_NONE;
+
+		if (!row_check_index_for_mysql(prebuilt, index, &n_rows)) {
+			push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
+					    ER_NOT_KEYFILE,
+					    "InnoDB: The B-tree of"
+					    " index '%-.200s' is corrupted.",
+					    index->name);
+			is_ok = FALSE;
+		}
+
+		if (thd_killed(user_thd)) {
+			break;
+		}
+
+#if 0
+		fprintf(stderr, "%lu entries in index %s\n", n_rows,
+			index->name);
+#endif
+
+		if (index == dict_table_get_first_index(prebuilt->table)) {
+			n_rows_in_table = n_rows;
+		} else if (n_rows != n_rows_in_table) {
+			push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
+					    ER_NOT_KEYFILE,
+					    "InnoDB: Index '%-.200s'"
+					    " contains %lu entries,"
+					    " should be %lu.",
+					    index->name,
+					    (ulong) n_rows,
+					    (ulong) n_rows_in_table);
+			is_ok = FALSE;
+		}
+	}
+
+	/* Restore the original isolation level */
+	prebuilt->trx->isolation_level = old_isolation_level;
+
+	/* We validate also the whole adaptive hash index for all tables
+	at every CHECK TABLE */
+
+	if (!btr_search_validate()) {
+		push_warning(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
+			     ER_NOT_KEYFILE,
+			     "InnoDB: The adaptive hash index is corrupted.");
+		is_ok = FALSE;
+	}
+
+	/* Restore the fatal lock wait timeout after CHECK TABLE. */
+	mutex_enter(&kernel_mutex);
+	srv_fatal_semaphore_wait_threshold -= 7200; /* 2 hours */
+	mutex_exit(&kernel_mutex);
+
+	prebuilt->trx->op_info = "";
+	if (thd_killed(user_thd)) {
+		my_error(ER_QUERY_INTERRUPTED, MYF(0));
+	}
+
+	DBUG_RETURN(is_ok ? HA_ADMIN_OK : HA_ADMIN_CORRUPT);
 }
 
 /*************************************************************//**
diff --git a/storage/innodb_plugin/include/row0mysql.h b/storage/innodb_plugin/include/row0mysql.h
index 28cd528bd25..03b29fd6538 100644
--- a/storage/innodb_plugin/include/row0mysql.h
+++ b/storage/innodb_plugin/include/row0mysql.h
@@ -500,14 +500,19 @@ row_rename_table_for_mysql(
 	trx_t*		trx,		/*!< in: transaction handle */
 	ibool		commit);	/*!< in: if TRUE then commit trx */
 /*********************************************************************//**
-Checks a table for corruption.
-@return	DB_ERROR or DB_SUCCESS */
+Checks that the index contains entries in an ascending order, unique
+constraint is not broken, and calculates the number of index entries
+in the read view of the current transaction.
+@return	DB_SUCCESS if ok */
 UNIV_INTERN
 ulint
-row_check_table_for_mysql(
+row_check_index_for_mysql(
 /*======================*/
-	row_prebuilt_t*	prebuilt);	/*!< in: prebuilt struct in MySQL
-					handle */
+	row_prebuilt_t*		prebuilt,	/*!< in: prebuilt struct
+						in MySQL handle */
+	const dict_index_t*	index,		/*!< in: index */
+	ulint*			n_rows);	/*!< out: number of entries
+						seen in the consistent read */
 
 /*********************************************************************//**
 Determines if a table is a magic monitor table.
diff --git a/storage/innodb_plugin/row/row0mysql.c b/storage/innodb_plugin/row/row0mysql.c
index cf0a632aee8..ad78c559141 100644
--- a/storage/innodb_plugin/row/row0mysql.c
+++ b/storage/innodb_plugin/row/row0mysql.c
@@ -4007,14 +4007,15 @@ Checks that the index contains entries in an ascending order, unique
 constraint is not broken, and calculates the number of index entries
 in the read view of the current transaction.
 @return	TRUE if ok */
-static
+UNIV_INTERN
 ibool
-row_scan_and_check_index(
-/*=====================*/
-	row_prebuilt_t*	prebuilt,	/*!< in: prebuilt struct in MySQL */
-	dict_index_t*	index,		/*!< in: index */
-	ulint*		n_rows)		/*!< out: number of entries seen in the
-					current consistent read */
+row_check_index_for_mysql(
+/*======================*/
+	row_prebuilt_t*		prebuilt,	/*!< in: prebuilt struct
+						in MySQL handle */
+	const dict_index_t*	index,		/*!< in: index */
+	ulint*			n_rows)		/*!< out: number of entries
+						seen in the consistent read */
 {
 	dtuple_t*	prev_entry	= NULL;
 	ulint		matched_fields;
@@ -4035,31 +4036,9 @@ row_scan_and_check_index(
 
 	*n_rows = 0;
 
-	if (!row_merge_is_index_usable(prebuilt->trx, index)) {
-		/* A newly created index may lack some delete-marked
-		records that may exist in the read view of
-		prebuilt->trx.  Thus, such indexes must not be
-		accessed by consistent read. */
-		return(is_ok);
-	}
-
 	buf = mem_alloc(UNIV_PAGE_SIZE);
 	heap = mem_heap_create(100);
 
-	/* Make a dummy template in prebuilt, which we will use
-	in scanning the index entries */
-
-	prebuilt->index = index;
-	/* row_merge_is_index_usable() was already checked above. */
-	prebuilt->index_usable = TRUE;
-	prebuilt->sql_stat_start = TRUE;
-	prebuilt->template_type = ROW_MYSQL_DUMMY_TEMPLATE;
-	prebuilt->n_template = 0;
-	prebuilt->need_to_access_clustered = FALSE;
-
-	dtuple_set_n_fields(prebuilt->search_tuple, 0);
-
-	prebuilt->select_lock_type = LOCK_NONE;
 	cnt = 1000;
 
 	ret = row_search_for_mysql(buf, PAGE_CUR_G, prebuilt, 0, 0);
@@ -4177,119 +4156,6 @@ not_ok:
 	goto loop;
 }
 
-/*********************************************************************//**
-Checks a table for corruption.
-@return	DB_ERROR or DB_SUCCESS */
-UNIV_INTERN
-ulint
-row_check_table_for_mysql(
-/*======================*/
-	row_prebuilt_t*	prebuilt)	/*!< in: prebuilt struct in MySQL
-					handle */
-{
-	dict_table_t*	table		= prebuilt->table;
-	dict_index_t*	index;
-	ulint		n_rows;
-	ulint		n_rows_in_table	= ULINT_UNDEFINED;
-	ulint		ret		= DB_SUCCESS;
-	ulint		old_isolation_level;
-
-	if (table->ibd_file_missing) {
-		ut_print_timestamp(stderr);
-		fprintf(stderr, "  InnoDB: Error:\n"
-			"InnoDB: MySQL is trying to use a table handle"
-			" but the .ibd file for\n"
-			"InnoDB: table %s does not exist.\n"
-			"InnoDB: Have you deleted the .ibd file"
-			" from the database directory under\n"
-			"InnoDB: the MySQL datadir, or have you"
-			" used DISCARD TABLESPACE?\n"
-			"InnoDB: Look from\n"
-			"InnoDB: " REFMAN "innodb-troubleshooting.html\n"
-			"InnoDB: how you can resolve the problem.\n",
-			table->name);
-		return(DB_ERROR);
-	}
-
-	prebuilt->trx->op_info = "checking table";
-
-	old_isolation_level = prebuilt->trx->isolation_level;
-
-	/* We must run the index record counts at an isolation level
-	>= READ COMMITTED, because a dirty read can see a wrong number
-	of records in some index; to play safe, we use always
-	REPEATABLE READ here */
-
-	prebuilt->trx->isolation_level = TRX_ISO_REPEATABLE_READ;
-
-	/* Enlarge the fatal lock wait timeout during CHECK TABLE. */
-	mutex_enter(&kernel_mutex);
-	srv_fatal_semaphore_wait_threshold += 7200; /* 2 hours */
-	mutex_exit(&kernel_mutex);
-
-	index = dict_table_get_first_index(table);
-
-	while (index != NULL) {
-		/* fputs("Validating index ", stderr);
-		ut_print_name(stderr, trx, FALSE, index->name);
-		putc('\n', stderr); */
-
-		if (!btr_validate_index(index, prebuilt->trx)) {
-			ret = DB_ERROR;
-		} else {
-			if (!row_scan_and_check_index(prebuilt,index, &n_rows)){
-				ret = DB_ERROR;
-			}
-
-			if (trx_is_interrupted(prebuilt->trx)) {
-				ret = DB_INTERRUPTED;
-				break;
-			}
-
-			/* fprintf(stderr, "%lu entries in index %s\n", n_rows,
-			index->name); */
-
-			if (index == dict_table_get_first_index(table)) {
-				n_rows_in_table = n_rows;
-			} else if (n_rows != n_rows_in_table) {
-
-				ret = DB_ERROR;
-
-				fputs("Error: ", stderr);
-				dict_index_name_print(stderr,
-						      prebuilt->trx, index);
-				fprintf(stderr,
-					" contains %lu entries,"
-					" should be %lu\n",
-					(ulong) n_rows,
-					(ulong) n_rows_in_table);
-			}
-		}
-
-		index = dict_table_get_next_index(index);
-	}
-
-	/* Restore the original isolation level */
-	prebuilt->trx->isolation_level = old_isolation_level;
-
-	/* We validate also the whole adaptive hash index for all tables
-	at every CHECK TABLE */
-
-	if (!btr_search_validate()) {
-
-		ret = DB_ERROR;
-	}
-
-	/* Restore the fatal lock wait timeout after CHECK TABLE. */
-	mutex_enter(&kernel_mutex);
-	srv_fatal_semaphore_wait_threshold -= 7200; /* 2 hours */
-	mutex_exit(&kernel_mutex);
-
-	prebuilt->trx->op_info = "";
-
-	return(ret);
-}
-
 /*********************************************************************//**
 Determines if a table is a magic monitor table.
 @return	TRUE if monitor table */

From 409e884391dbfdfa97716cc26e35e693e1b5c219 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:36:54 +0400
Subject: [PATCH 042/116] Applying InnoDB snapshot, fixes BUG#35077.

Detailed revision comments:

r6548 | marko | 2010-02-03 15:01:39 +0200 (Wed, 03 Feb 2010) | 11 lines
branches/zip: buf_LRU_invalidate_tablespace(): Ensure that prev_bpage
is not relocated when freeing a compressed block.  This avoids the
costly rescan of the LRU list.  (Bug #35077, Issue #449)

At most one buffer-fix will be active at a time, affecting two blocks:
the buf_page_t and the compressed page frame. This should not block
the memory defragmentation in buf0buddy.c too much.  In fact, it may
avoid unnecessary copying if also prev_bpage belongs to the tablespace
that is being invalidated.

rb://240
---
 storage/innodb_plugin/ChangeLog     |  6 +++
 storage/innodb_plugin/buf/buf0lru.c | 62 ++++++++++++++++++++++++-----
 2 files changed, 58 insertions(+), 10 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 7338bf5cbab..5566aeb1291 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,9 @@
+2010-02-03	The InnoDB Team
+
+	* buf/buf0lru.c:
+	Fix Bug#35077 Very slow DROP TABLE (ALTER TABLE, OPTIMIZE TABLE)
+	on compressed tables
+
 2010-02-03	The InnoDB Team
 
 	* handler/ha_innodb.cc, include/row0mysql.h, row/row0mysql.c:
diff --git a/storage/innodb_plugin/buf/buf0lru.c b/storage/innodb_plugin/buf/buf0lru.c
index eeef43de83b..62198491939 100644
--- a/storage/innodb_plugin/buf/buf0lru.c
+++ b/storage/innodb_plugin/buf/buf0lru.c
@@ -347,11 +347,11 @@ scan_again:
 
 	all_freed = TRUE;
 
-rescan:
 	bpage = UT_LIST_GET_LAST(buf_pool->LRU);
 
 	while (bpage != NULL) {
 		buf_page_t*	prev_bpage;
+		ibool		prev_bpage_buf_fix = FALSE;
 
 		ut_a(buf_page_in_file(bpage));
 
@@ -394,8 +394,40 @@ rescan:
 					(ulong) buf_page_get_page_no(bpage));
 			}
 #endif
-			if (buf_page_get_state(bpage) == BUF_BLOCK_FILE_PAGE
-			    && ((buf_block_t*) bpage)->is_hashed) {
+			if (buf_page_get_state(bpage) != BUF_BLOCK_FILE_PAGE) {
+				/* This is a compressed-only block
+				descriptor.  Ensure that prev_bpage
+				cannot be relocated when bpage is freed. */
+				if (UNIV_LIKELY(prev_bpage != NULL)) {
+					switch (buf_page_get_state(
+							prev_bpage)) {
+					case BUF_BLOCK_FILE_PAGE:
+						/* Descriptors of uncompressed
+						blocks will not be relocated,
+						because we are holding the
+						buf_pool_mutex. */
+						break;
+					case BUF_BLOCK_ZIP_PAGE:
+					case BUF_BLOCK_ZIP_DIRTY:
+						/* Descriptors of compressed-
+						only blocks can be relocated,
+						unless they are buffer-fixed.
+						Because both bpage and
+						prev_bpage are protected by
+						buf_pool_zip_mutex, it is
+						not necessary to acquire
+						further mutexes. */
+						ut_ad(&buf_pool_zip_mutex
+						      == block_mutex);
+						ut_ad(mutex_own(block_mutex));
+						prev_bpage_buf_fix = TRUE;
+						prev_bpage->buf_fix_count++;
+						break;
+					default:
+						ut_error;
+					}
+				}
+			} else if (((buf_block_t*) bpage)->is_hashed) {
 				ulint	page_no;
 				ulint	zip_size;
 
@@ -419,7 +451,8 @@ rescan:
 				buf_flush_remove(bpage);
 			}
 
-			/* Remove from the LRU list */
+			/* Remove from the LRU list. */
+
 			if (buf_LRU_block_remove_hashed_page(bpage, TRUE)
 			    != BUF_BLOCK_ZIP_FREE) {
 				buf_LRU_block_free_hashed_page((buf_block_t*)
@@ -431,18 +464,27 @@ rescan:
 				ut_ad(block_mutex == &buf_pool_zip_mutex);
 				ut_ad(!mutex_own(block_mutex));
 
-				/* The compressed block descriptor
-				(bpage) has been deallocated and
-				block_mutex released.  Also,
-				buf_buddy_free() may have relocated
-				prev_bpage.  Rescan the LRU list. */
+				if (prev_bpage_buf_fix) {
+					/* We temporarily buffer-fixed
+					prev_bpage, so that
+					buf_buddy_free() could not
+					relocate it, in case it was a
+					compressed-only block
+					descriptor. */
 
-				goto rescan;
+					mutex_enter(block_mutex);
+					ut_ad(prev_bpage->buf_fix_count > 0);
+					prev_bpage->buf_fix_count--;
+					mutex_exit(block_mutex);
+				}
+
+				goto next_page_no_mutex;
 			}
 next_page:
 			mutex_exit(block_mutex);
 		}
 
+next_page_no_mutex:
 		bpage = prev_bpage;
 	}
 

From c3a7bf0d1ab56a033bcc0a58a3a6afd760815e1b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:37:45 +0400
Subject: [PATCH 043/116] Applying InnoDB snapshot

Detailed revision comments:

r6559 | marko | 2010-02-04 13:21:18 +0200 (Thu, 04 Feb 2010) | 14 lines
branches/zip: Pass the file name and line number of the caller of the
b-tree cursor functions to the buffer pool requests, in order to make
the latch diagnostics more accurate.

buf_page_optimistic_get_func(): Renamed to buf_page_optimistic_get().

btr_page_get_father_node_ptr(), btr_insert_on_non_leaf_level(),
btr_pcur_open(), btr_pcur_open_with_no_init(), btr_pcur_open_on_user_rec(),
btr_pcur_open_at_rnd_pos(), btr_pcur_restore_position(),
btr_cur_open_at_index_side(), btr_cur_open_at_rnd_pos():
Rename the function to _func and add the parameters file, line.
Define wrapper macros with __FILE__, __LINE__.

btr_cur_search_to_nth_level(): Add the parameters file, line.
---
 storage/innodb_plugin/ChangeLog           |  9 +++++
 storage/innodb_plugin/btr/btr0btr.c       | 20 +++++++----
 storage/innodb_plugin/btr/btr0cur.c       | 26 +++++++++-----
 storage/innodb_plugin/btr/btr0pcur.c      | 22 +++++++-----
 storage/innodb_plugin/buf/buf0buf.c       |  4 +--
 storage/innodb_plugin/include/btr0btr.h   |  8 +++--
 storage/innodb_plugin/include/btr0cur.h   | 18 +++++++---
 storage/innodb_plugin/include/btr0pcur.h  | 41 +++++++++++++++++------
 storage/innodb_plugin/include/btr0pcur.ic | 28 ++++++++++------
 storage/innodb_plugin/include/buf0buf.h   | 10 ++----
 storage/innodb_plugin/row/row0ins.c       |  8 +++--
 11 files changed, 132 insertions(+), 62 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 5566aeb1291..d727501ea9a 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,12 @@
+2010-02-04	The InnoDB Team
+
+	* btr/btr0btr.c, btr/btr0cur.c, btr/btr0pcur.c, buf/buf0buf.c,
+	include/btr0btr.h, include/btr0cur.h, include/btr0pcur.h,
+	include/btr0pcur.ic, include/buf0buf.h, row/row0ins.c, row/row0sel.c:
+	Pass the file name and line number of the caller of the
+	b-tree cursor functions to the buffer pool requests, in order
+	to make the latch diagnostics more accurate.
+
 2010-02-03	The InnoDB Team
 
 	* buf/buf0lru.c:
diff --git a/storage/innodb_plugin/btr/btr0btr.c b/storage/innodb_plugin/btr/btr0btr.c
index 086b3a0a599..91f5e3a31b9 100644
--- a/storage/innodb_plugin/btr/btr0btr.c
+++ b/storage/innodb_plugin/btr/btr0btr.c
@@ -592,13 +592,15 @@ an x-latch on the tree.
 @return	rec_get_offsets() of the node pointer record */
 static
 ulint*
-btr_page_get_father_node_ptr(
-/*=========================*/
+btr_page_get_father_node_ptr_func(
+/*==============================*/
 	ulint*		offsets,/*!< in: work area for the return value */
 	mem_heap_t*	heap,	/*!< in: memory heap to use */
 	btr_cur_t*	cursor,	/*!< in: cursor pointing to user record,
 				out: cursor on node pointer record,
 				its page x-latched */
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line where called */
 	mtr_t*		mtr)	/*!< in: mtr */
 {
 	dtuple_t*	tuple;
@@ -622,7 +624,8 @@ btr_page_get_father_node_ptr(
 	tuple = dict_index_build_node_ptr(index, user_rec, 0, heap, level);
 
 	btr_cur_search_to_nth_level(index, level + 1, tuple, PAGE_CUR_LE,
-				    BTR_CONT_MODIFY_TREE, cursor, 0, mtr);
+				    BTR_CONT_MODIFY_TREE, cursor, 0,
+				    file, line, mtr);
 
 	node_ptr = btr_cur_get_rec(cursor);
 	ut_ad(!page_rec_is_comp(node_ptr)
@@ -670,6 +673,9 @@ btr_page_get_father_node_ptr(
 	return(offsets);
 }
 
+#define btr_page_get_father_node_ptr(of,heap,cur,mtr)			\
+	btr_page_get_father_node_ptr_func(of,heap,cur,__FILE__,__LINE__,mtr)
+
 /************************************************************//**
 Returns the upper level node pointer to a page. It is assumed that mtr holds
 an x-latch on the tree.
@@ -1662,11 +1668,13 @@ Inserts a data tuple to a tree on a non-leaf level. It is assumed
 that mtr holds an x-latch on the tree. */
 UNIV_INTERN
 void
-btr_insert_on_non_leaf_level(
-/*=========================*/
+btr_insert_on_non_leaf_level_func(
+/*==============================*/
 	dict_index_t*	index,	/*!< in: index */
 	ulint		level,	/*!< in: level, must be > 0 */
 	dtuple_t*	tuple,	/*!< in: the record to be inserted */
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line where called */
 	mtr_t*		mtr)	/*!< in: mtr */
 {
 	big_rec_t*	dummy_big_rec;
@@ -1678,7 +1686,7 @@ btr_insert_on_non_leaf_level(
 
 	btr_cur_search_to_nth_level(index, level, tuple, PAGE_CUR_LE,
 				    BTR_CONT_MODIFY_TREE,
-				    &cursor, 0, mtr);
+				    &cursor, 0, file, line, mtr);
 
 	err = btr_cur_pessimistic_insert(BTR_NO_LOCKING_FLAG
 					 | BTR_KEEP_SYS_FLAG
diff --git a/storage/innodb_plugin/btr/btr0cur.c b/storage/innodb_plugin/btr/btr0cur.c
index 46dfb5d1a46..df21b8d4440 100644
--- a/storage/innodb_plugin/btr/btr0cur.c
+++ b/storage/innodb_plugin/btr/btr0cur.c
@@ -342,6 +342,8 @@ btr_cur_search_to_nth_level(
 	ulint		has_search_latch,/*!< in: info on the latch mode the
 				caller currently has on btr_search_latch:
 				RW_S_LATCH, or 0 */
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line where called */
 	mtr_t*		mtr)	/*!< in: mtr */
 {
 	page_cur_t*	page_cursor;
@@ -520,7 +522,7 @@ btr_cur_search_to_nth_level(
 retry_page_get:
 		block = buf_page_get_gen(space, zip_size, page_no,
 					 rw_latch, guess, buf_mode,
-					 __FILE__, __LINE__, mtr);
+					 file, line, mtr);
 		if (block == NULL) {
 			/* This must be a search to perform an insert;
 			try insert to the insert buffer */
@@ -677,13 +679,15 @@ func_exit:
 Opens a cursor at either end of an index. */
 UNIV_INTERN
 void
-btr_cur_open_at_index_side(
-/*=======================*/
+btr_cur_open_at_index_side_func(
+/*============================*/
 	ibool		from_left,	/*!< in: TRUE if open to the low end,
 					FALSE if to the high end */
 	dict_index_t*	index,		/*!< in: index */
 	ulint		latch_mode,	/*!< in: latch mode */
 	btr_cur_t*	cursor,		/*!< in: cursor */
+	const char*	file,		/*!< in: file name */
+	ulint		line,		/*!< in: line where called */
 	mtr_t*		mtr)		/*!< in: mtr */
 {
 	page_cur_t*	page_cursor;
@@ -728,7 +732,7 @@ btr_cur_open_at_index_side(
 		page_t*		page;
 		block = buf_page_get_gen(space, zip_size, page_no,
 					 RW_NO_LATCH, NULL, BUF_GET,
-					 __FILE__, __LINE__, mtr);
+					 file, line, mtr);
 		page = buf_block_get_frame(block);
 		ut_ad(0 == ut_dulint_cmp(index->id,
 					 btr_page_get_index_id(page)));
@@ -808,11 +812,13 @@ btr_cur_open_at_index_side(
 Positions a cursor at a randomly chosen position within a B-tree. */
 UNIV_INTERN
 void
-btr_cur_open_at_rnd_pos(
-/*====================*/
+btr_cur_open_at_rnd_pos_func(
+/*=========================*/
 	dict_index_t*	index,		/*!< in: index */
 	ulint		latch_mode,	/*!< in: BTR_SEARCH_LEAF, ... */
 	btr_cur_t*	cursor,		/*!< in/out: B-tree cursor */
+	const char*	file,		/*!< in: file name */
+	ulint		line,		/*!< in: line where called */
 	mtr_t*		mtr)		/*!< in: mtr */
 {
 	page_cur_t*	page_cursor;
@@ -847,7 +853,7 @@ btr_cur_open_at_rnd_pos(
 
 		block = buf_page_get_gen(space, zip_size, page_no,
 					 RW_NO_LATCH, NULL, BUF_GET,
-					 __FILE__, __LINE__, mtr);
+					 file, line, mtr);
 		page = buf_block_get_frame(block);
 		ut_ad(0 == ut_dulint_cmp(index->id,
 					 btr_page_get_index_id(page)));
@@ -3100,7 +3106,8 @@ btr_estimate_n_rows_in_range(
 
 		btr_cur_search_to_nth_level(index, 0, tuple1, mode1,
 					    BTR_SEARCH_LEAF | BTR_ESTIMATE,
-					    &cursor, 0, &mtr);
+					    &cursor, 0,
+					    __FILE__, __LINE__, &mtr);
 	} else {
 		btr_cur_open_at_index_side(TRUE, index,
 					   BTR_SEARCH_LEAF | BTR_ESTIMATE,
@@ -3117,7 +3124,8 @@ btr_estimate_n_rows_in_range(
 
 		btr_cur_search_to_nth_level(index, 0, tuple2, mode2,
 					    BTR_SEARCH_LEAF | BTR_ESTIMATE,
-					    &cursor, 0, &mtr);
+					    &cursor, 0,
+					    __FILE__, __LINE__, &mtr);
 	} else {
 		btr_cur_open_at_index_side(FALSE, index,
 					   BTR_SEARCH_LEAF | BTR_ESTIMATE,
diff --git a/storage/innodb_plugin/btr/btr0pcur.c b/storage/innodb_plugin/btr/btr0pcur.c
index ec98692c35b..436d1f252c3 100644
--- a/storage/innodb_plugin/btr/btr0pcur.c
+++ b/storage/innodb_plugin/btr/btr0pcur.c
@@ -205,10 +205,12 @@ record and it can be restored on a user record whose ordering fields
 are identical to the ones of the original user record */
 UNIV_INTERN
 ibool
-btr_pcur_restore_position(
-/*======================*/
+btr_pcur_restore_position_func(
+/*===========================*/
 	ulint		latch_mode,	/*!< in: BTR_SEARCH_LEAF, ... */
 	btr_pcur_t*	cursor,		/*!< in: detached persistent cursor */
+	const char*	file,		/*!< in: file name */
+	ulint		line,		/*!< in: line where called */
 	mtr_t*		mtr)		/*!< in: mtr */
 {
 	dict_index_t*	index;
@@ -257,7 +259,8 @@ btr_pcur_restore_position(
 		if (UNIV_LIKELY(buf_page_optimistic_get(
 					latch_mode,
 					cursor->block_when_stored,
-					cursor->modify_clock, mtr))) {
+					cursor->modify_clock,
+					file, line, mtr))) {
 			cursor->pos_state = BTR_PCUR_IS_POSITIONED;
 
 			buf_block_dbg_add_level(btr_pcur_get_block(cursor),
@@ -312,8 +315,8 @@ btr_pcur_restore_position(
 		mode = PAGE_CUR_L;
 	}
 
-	btr_pcur_open_with_no_init(index, tuple, mode, latch_mode,
-				   cursor, 0, mtr);
+	btr_pcur_open_with_no_init_func(index, tuple, mode, latch_mode,
+					cursor, 0, file, line, mtr);
 
 	/* Restore the old search mode */
 	cursor->search_mode = old_mode;
@@ -553,8 +556,8 @@ before first in tree. The latching mode must be BTR_SEARCH_LEAF or
 BTR_MODIFY_LEAF. */
 UNIV_INTERN
 void
-btr_pcur_open_on_user_rec(
-/*======================*/
+btr_pcur_open_on_user_rec_func(
+/*===========================*/
 	dict_index_t*	index,		/*!< in: index */
 	const dtuple_t*	tuple,		/*!< in: tuple on which search done */
 	ulint		mode,		/*!< in: PAGE_CUR_L, ... */
@@ -562,9 +565,12 @@ btr_pcur_open_on_user_rec(
 					BTR_MODIFY_LEAF */
 	btr_pcur_t*	cursor,		/*!< in: memory buffer for persistent
 					cursor */
+	const char*	file,		/*!< in: file name */
+	ulint		line,		/*!< in: line where called */
 	mtr_t*		mtr)		/*!< in: mtr */
 {
-	btr_pcur_open(index, tuple, mode, latch_mode, cursor, mtr);
+	btr_pcur_open_func(index, tuple, mode, latch_mode, cursor,
+			   file, line, mtr);
 
 	if ((mode == PAGE_CUR_GE) || (mode == PAGE_CUR_G)) {
 
diff --git a/storage/innodb_plugin/buf/buf0buf.c b/storage/innodb_plugin/buf/buf0buf.c
index fd5b07330cb..c4afff1e3e7 100644
--- a/storage/innodb_plugin/buf/buf0buf.c
+++ b/storage/innodb_plugin/buf/buf0buf.c
@@ -2358,8 +2358,8 @@ page.
 @return	TRUE if success */
 UNIV_INTERN
 ibool
-buf_page_optimistic_get_func(
-/*=========================*/
+buf_page_optimistic_get(
+/*====================*/
 	ulint		rw_latch,/*!< in: RW_S_LATCH, RW_X_LATCH */
 	buf_block_t*	block,	/*!< in: guessed buffer block */
 	ib_uint64_t	modify_clock,/*!< in: modify clock value if mode is
diff --git a/storage/innodb_plugin/include/btr0btr.h b/storage/innodb_plugin/include/btr0btr.h
index d5c8258513c..11cacfc2244 100644
--- a/storage/innodb_plugin/include/btr0btr.h
+++ b/storage/innodb_plugin/include/btr0btr.h
@@ -317,12 +317,16 @@ Inserts a data tuple to a tree on a non-leaf level. It is assumed
 that mtr holds an x-latch on the tree. */
 UNIV_INTERN
 void
-btr_insert_on_non_leaf_level(
-/*=========================*/
+btr_insert_on_non_leaf_level_func(
+/*==============================*/
 	dict_index_t*	index,	/*!< in: index */
 	ulint		level,	/*!< in: level, must be > 0 */
 	dtuple_t*	tuple,	/*!< in: the record to be inserted */
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line where called */
 	mtr_t*		mtr);	/*!< in: mtr */
+# define btr_insert_on_non_leaf_level(i,l,t,m)				\
+	btr_insert_on_non_leaf_level_func(i,l,t,__FILE__,__LINE__,m)
 #endif /* !UNIV_HOTBACKUP */
 /****************************************************************//**
 Sets a record as the predefined minimum record. */
diff --git a/storage/innodb_plugin/include/btr0cur.h b/storage/innodb_plugin/include/btr0cur.h
index 480a3877e54..8be3672f22d 100644
--- a/storage/innodb_plugin/include/btr0cur.h
+++ b/storage/innodb_plugin/include/btr0cur.h
@@ -152,29 +152,39 @@ btr_cur_search_to_nth_level(
 	ulint		has_search_latch,/*!< in: latch mode the caller
 				currently has on btr_search_latch:
 				RW_S_LATCH, or 0 */
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line where called */
 	mtr_t*		mtr);	/*!< in: mtr */
 /*****************************************************************//**
 Opens a cursor at either end of an index. */
 UNIV_INTERN
 void
-btr_cur_open_at_index_side(
-/*=======================*/
+btr_cur_open_at_index_side_func(
+/*============================*/
 	ibool		from_left,	/*!< in: TRUE if open to the low end,
 					FALSE if to the high end */
 	dict_index_t*	index,		/*!< in: index */
 	ulint		latch_mode,	/*!< in: latch mode */
 	btr_cur_t*	cursor,		/*!< in: cursor */
+	const char*	file,		/*!< in: file name */
+	ulint		line,		/*!< in: line where called */
 	mtr_t*		mtr);		/*!< in: mtr */
+#define btr_cur_open_at_index_side(f,i,l,c,m)				\
+	btr_cur_open_at_index_side_func(f,i,l,c,__FILE__,__LINE__,m)
 /**********************************************************************//**
 Positions a cursor at a randomly chosen position within a B-tree. */
 UNIV_INTERN
 void
-btr_cur_open_at_rnd_pos(
-/*====================*/
+btr_cur_open_at_rnd_pos_func(
+/*=========================*/
 	dict_index_t*	index,		/*!< in: index */
 	ulint		latch_mode,	/*!< in: BTR_SEARCH_LEAF, ... */
 	btr_cur_t*	cursor,		/*!< in/out: B-tree cursor */
+	const char*	file,		/*!< in: file name */
+	ulint		line,		/*!< in: line where called */
 	mtr_t*		mtr);		/*!< in: mtr */
+#define btr_cur_open_at_rnd_pos(i,l,c,m)				\
+	btr_cur_open_at_rnd_pos_func(i,l,c,__FILE__,__LINE__,m)
 /*************************************************************//**
 Tries to perform an insert to a page in an index tree, next to cursor.
 It is assumed that mtr holds an x-latch on the page. The operation does
diff --git a/storage/innodb_plugin/include/btr0pcur.h b/storage/innodb_plugin/include/btr0pcur.h
index 12b1375d8b7..7d8bac85f3e 100644
--- a/storage/innodb_plugin/include/btr0pcur.h
+++ b/storage/innodb_plugin/include/btr0pcur.h
@@ -82,8 +82,8 @@ Initializes and opens a persistent cursor to an index tree. It should be
 closed with btr_pcur_close. */
 UNIV_INLINE
 void
-btr_pcur_open(
-/*==========*/
+btr_pcur_open_func(
+/*===============*/
 	dict_index_t*	index,	/*!< in: index */
 	const dtuple_t*	tuple,	/*!< in: tuple on which search done */
 	ulint		mode,	/*!< in: PAGE_CUR_L, ...;
@@ -94,14 +94,18 @@ btr_pcur_open(
 				record! */
 	ulint		latch_mode,/*!< in: BTR_SEARCH_LEAF, ... */
 	btr_pcur_t*	cursor, /*!< in: memory buffer for persistent cursor */
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line where called */
 	mtr_t*		mtr);	/*!< in: mtr */
+#define btr_pcur_open(i,t,md,l,c,m)				\
+	btr_pcur_open_func(i,t,md,l,c,__FILE__,__LINE__,m)
 /**************************************************************//**
 Opens an persistent cursor to an index tree without initializing the
 cursor. */
 UNIV_INLINE
 void
-btr_pcur_open_with_no_init(
-/*=======================*/
+btr_pcur_open_with_no_init_func(
+/*============================*/
 	dict_index_t*	index,	/*!< in: index */
 	const dtuple_t*	tuple,	/*!< in: tuple on which search done */
 	ulint		mode,	/*!< in: PAGE_CUR_L, ...;
@@ -119,7 +123,12 @@ btr_pcur_open_with_no_init(
 	ulint		has_search_latch,/*!< in: latch mode the caller
 				currently has on btr_search_latch:
 				RW_S_LATCH, or 0 */
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line where called */
 	mtr_t*		mtr);	/*!< in: mtr */
+#define btr_pcur_open_with_no_init(ix,t,md,l,cur,has,m)			\
+	btr_pcur_open_with_no_init_func(ix,t,md,l,cur,has,__FILE__,__LINE__,m)
+
 /*****************************************************************//**
 Opens a persistent cursor at either end of an index. */
 UNIV_INLINE
@@ -160,8 +169,8 @@ before first in tree. The latching mode must be BTR_SEARCH_LEAF or
 BTR_MODIFY_LEAF. */
 UNIV_INTERN
 void
-btr_pcur_open_on_user_rec(
-/*======================*/
+btr_pcur_open_on_user_rec_func(
+/*===========================*/
 	dict_index_t*	index,		/*!< in: index */
 	const dtuple_t*	tuple,		/*!< in: tuple on which search done */
 	ulint		mode,		/*!< in: PAGE_CUR_L, ... */
@@ -169,17 +178,25 @@ btr_pcur_open_on_user_rec(
 					BTR_MODIFY_LEAF */
 	btr_pcur_t*	cursor,		/*!< in: memory buffer for persistent
 					cursor */
+	const char*	file,		/*!< in: file name */
+	ulint		line,		/*!< in: line where called */
 	mtr_t*		mtr);		/*!< in: mtr */
+#define btr_pcur_open_on_user_rec(i,t,md,l,c,m)				\
+	btr_pcur_open_on_user_rec_func(i,t,md,l,c,__FILE__,__LINE__,m)
 /**********************************************************************//**
 Positions a cursor at a randomly chosen position within a B-tree. */
 UNIV_INLINE
 void
-btr_pcur_open_at_rnd_pos(
-/*=====================*/
+btr_pcur_open_at_rnd_pos_func(
+/*==========================*/
 	dict_index_t*	index,		/*!< in: index */
 	ulint		latch_mode,	/*!< in: BTR_SEARCH_LEAF, ... */
 	btr_pcur_t*	cursor,		/*!< in/out: B-tree pcur */
+	const char*	file,		/*!< in: file name */
+	ulint		line,		/*!< in: line where called */
 	mtr_t*		mtr);		/*!< in: mtr */
+#define btr_pcur_open_at_rnd_pos(i,l,c,m)				\
+	btr_pcur_open_at_rnd_pos_func(i,l,c,__FILE__,__LINE__,m)
 /**************************************************************//**
 Frees the possible old_rec_buf buffer of a persistent cursor and sets the
 latch mode of the persistent cursor to BTR_NO_LATCHES. */
@@ -218,11 +235,15 @@ record and it can be restored on a user record whose ordering fields
 are identical to the ones of the original user record */
 UNIV_INTERN
 ibool
-btr_pcur_restore_position(
-/*======================*/
+btr_pcur_restore_position_func(
+/*===========================*/
 	ulint		latch_mode,	/*!< in: BTR_SEARCH_LEAF, ... */
 	btr_pcur_t*	cursor,		/*!< in: detached persistent cursor */
+	const char*	file,		/*!< in: file name */
+	ulint		line,		/*!< in: line where called */
 	mtr_t*		mtr);		/*!< in: mtr */
+#define btr_pcur_restore_position(l,cur,mtr)				\
+	btr_pcur_restore_position_func(l,cur,__FILE__,__LINE__,mtr)
 /**************************************************************//**
 If the latch mode of the cursor is BTR_LEAF_SEARCH or BTR_LEAF_MODIFY,
 releases the page latch and bufferfix reserved by the cursor.
diff --git a/storage/innodb_plugin/include/btr0pcur.ic b/storage/innodb_plugin/include/btr0pcur.ic
index 0ca7223f861..b11d7dce21c 100644
--- a/storage/innodb_plugin/include/btr0pcur.ic
+++ b/storage/innodb_plugin/include/btr0pcur.ic
@@ -483,8 +483,8 @@ Initializes and opens a persistent cursor to an index tree. It should be
 closed with btr_pcur_close. */
 UNIV_INLINE
 void
-btr_pcur_open(
-/*==========*/
+btr_pcur_open_func(
+/*===============*/
 	dict_index_t*	index,	/*!< in: index */
 	const dtuple_t*	tuple,	/*!< in: tuple on which search done */
 	ulint		mode,	/*!< in: PAGE_CUR_L, ...;
@@ -495,6 +495,8 @@ btr_pcur_open(
 				record! */
 	ulint		latch_mode,/*!< in: BTR_SEARCH_LEAF, ... */
 	btr_pcur_t*	cursor, /*!< in: memory buffer for persistent cursor */
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line where called */
 	mtr_t*		mtr)	/*!< in: mtr */
 {
 	btr_cur_t*	btr_cursor;
@@ -511,7 +513,7 @@ btr_pcur_open(
 	btr_cursor = btr_pcur_get_btr_cur(cursor);
 
 	btr_cur_search_to_nth_level(index, 0, tuple, mode, latch_mode,
-				    btr_cursor, 0, mtr);
+				    btr_cursor, 0, file, line, mtr);
 	cursor->pos_state = BTR_PCUR_IS_POSITIONED;
 
 	cursor->trx_if_known = NULL;
@@ -522,8 +524,8 @@ Opens an persistent cursor to an index tree without initializing the
 cursor. */
 UNIV_INLINE
 void
-btr_pcur_open_with_no_init(
-/*=======================*/
+btr_pcur_open_with_no_init_func(
+/*============================*/
 	dict_index_t*	index,	/*!< in: index */
 	const dtuple_t*	tuple,	/*!< in: tuple on which search done */
 	ulint		mode,	/*!< in: PAGE_CUR_L, ...;
@@ -541,6 +543,8 @@ btr_pcur_open_with_no_init(
 	ulint		has_search_latch,/*!< in: latch mode the caller
 				currently has on btr_search_latch:
 				RW_S_LATCH, or 0 */
+	const char*	file,	/*!< in: file name */
+	ulint		line,	/*!< in: line where called */
 	mtr_t*		mtr)	/*!< in: mtr */
 {
 	btr_cur_t*	btr_cursor;
@@ -553,7 +557,8 @@ btr_pcur_open_with_no_init(
 	btr_cursor = btr_pcur_get_btr_cur(cursor);
 
 	btr_cur_search_to_nth_level(index, 0, tuple, mode, latch_mode,
-				    btr_cursor, has_search_latch, mtr);
+				    btr_cursor, has_search_latch,
+				    file, line, mtr);
 	cursor->pos_state = BTR_PCUR_IS_POSITIONED;
 
 	cursor->old_stored = BTR_PCUR_OLD_NOT_STORED;
@@ -600,11 +605,13 @@ btr_pcur_open_at_index_side(
 Positions a cursor at a randomly chosen position within a B-tree. */
 UNIV_INLINE
 void
-btr_pcur_open_at_rnd_pos(
-/*=====================*/
+btr_pcur_open_at_rnd_pos_func(
+/*==========================*/
 	dict_index_t*	index,		/*!< in: index */
 	ulint		latch_mode,	/*!< in: BTR_SEARCH_LEAF, ... */
 	btr_pcur_t*	cursor,		/*!< in/out: B-tree pcur */
+	const char*	file,		/*!< in: file name */
+	ulint		line,		/*!< in: line where called */
 	mtr_t*		mtr)		/*!< in: mtr */
 {
 	/* Initialize the cursor */
@@ -614,8 +621,9 @@ btr_pcur_open_at_rnd_pos(
 
 	btr_pcur_init(cursor);
 
-	btr_cur_open_at_rnd_pos(index, latch_mode,
-				btr_pcur_get_btr_cur(cursor), mtr);
+	btr_cur_open_at_rnd_pos_func(index, latch_mode,
+				     btr_pcur_get_btr_cur(cursor),
+				     file, line, mtr);
 	cursor->pos_state = BTR_PCUR_IS_POSITIONED;
 	cursor->old_stored = BTR_PCUR_OLD_NOT_STORED;
 
diff --git a/storage/innodb_plugin/include/buf0buf.h b/storage/innodb_plugin/include/buf0buf.h
index 927ff893e39..cf47042423c 100644
--- a/storage/innodb_plugin/include/buf0buf.h
+++ b/storage/innodb_plugin/include/buf0buf.h
@@ -202,20 +202,14 @@ with care. */
 #define buf_page_get_with_no_latch(SP, ZS, OF, MTR)	   buf_page_get_gen(\
 				SP, ZS, OF, RW_NO_LATCH, NULL,\
 				BUF_GET_NO_LATCH, __FILE__, __LINE__, MTR)
-/**************************************************************//**
-NOTE! The following macros should be used instead of
-buf_page_optimistic_get_func, to improve debugging. Only values RW_S_LATCH and
-RW_X_LATCH are allowed as LA! */
-#define buf_page_optimistic_get(LA, BL, MC, MTR)			     \
-	buf_page_optimistic_get_func(LA, BL, MC, __FILE__, __LINE__, MTR)
 /********************************************************************//**
 This is the general function used to get optimistic access to a database
 page.
 @return	TRUE if success */
 UNIV_INTERN
 ibool
-buf_page_optimistic_get_func(
-/*=========================*/
+buf_page_optimistic_get(
+/*====================*/
 	ulint		rw_latch,/*!< in: RW_S_LATCH, RW_X_LATCH */
 	buf_block_t*	block,	/*!< in: guessed block */
 	ib_uint64_t	modify_clock,/*!< in: modify clock value if mode is
diff --git a/storage/innodb_plugin/row/row0ins.c b/storage/innodb_plugin/row/row0ins.c
index fe51fce82c4..ece983eb957 100644
--- a/storage/innodb_plugin/row/row0ins.c
+++ b/storage/innodb_plugin/row/row0ins.c
@@ -1991,7 +1991,7 @@ row_ins_index_entry_low(
 
 	btr_cur_search_to_nth_level(index, 0, entry, PAGE_CUR_LE,
 				    mode | BTR_INSERT | ignore_sec_unique,
-				    &cursor, 0, &mtr);
+				    &cursor, 0, __FILE__, __LINE__, &mtr);
 
 	if (cursor.flag == BTR_CUR_INSERT_TO_IBUF) {
 		/* The insertion was made to the insert buffer already during
@@ -2049,7 +2049,8 @@ row_ins_index_entry_low(
 			btr_cur_search_to_nth_level(index, 0, entry,
 						    PAGE_CUR_LE,
 						    mode | BTR_INSERT,
-						    &cursor, 0, &mtr);
+						    &cursor, 0,
+						    __FILE__, __LINE__, &mtr);
 		}
 	}
 
@@ -2104,7 +2105,8 @@ function_exit:
 		mtr_start(&mtr);
 
 		btr_cur_search_to_nth_level(index, 0, entry, PAGE_CUR_LE,
-					    BTR_MODIFY_TREE, &cursor, 0, &mtr);
+					    BTR_MODIFY_TREE, &cursor, 0,
+					    __FILE__, __LINE__, &mtr);
 		rec = btr_cur_get_rec(&cursor);
 		offsets = rec_get_offsets(rec, index, NULL,
 					  ULINT_UNDEFINED, &heap);

From 662a8f4a1966d13828d2943fcf3a6e2a7238359f Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:38:11 +0400
Subject: [PATCH 044/116] Applying InnoDB snapshot

Detailed revision comments:

r6560 | sunny | 2010-02-04 16:11:23 +0200 (Thu, 04 Feb 2010) | 7 lines
branches/zip: Remove the additional check introduced in r6534 which tries
to check if the joining transaction has any other transactions waiting on
its locks. This optimization results in excessive deadlocks when running
Sysbench with a large number of threads. The function seems to return
FALSE positives.

rb://250
---
 storage/innodb_plugin/lock/lock0lock.c | 63 --------------------------
 1 file changed, 63 deletions(-)

diff --git a/storage/innodb_plugin/lock/lock0lock.c b/storage/innodb_plugin/lock/lock0lock.c
index 4552c1e7cde..11a5232b0dc 100644
--- a/storage/innodb_plugin/lock/lock0lock.c
+++ b/storage/innodb_plugin/lock/lock0lock.c
@@ -3311,64 +3311,6 @@ retry:
 	return(FALSE);
 }
 
-/********************************************************************//**
-Check that no other transaction is waiting on this transaction's locks.
-@return TRUE if some other transaction is waiting for this lock. */
-static
-ulint
-lock_trx_has_no_waiters(
-/*====================*/
-	const trx_t*	trx)		/*!< in: the transaction to check */
-{
-	const lock_t*	lock;
-
-	ut_ad(mutex_own(&kernel_mutex));
-
-	for (lock = UT_LIST_GET_FIRST(trx->trx_locks);
-	     lock != NULL;
-	     lock = UT_LIST_GET_NEXT(trx_locks, lock)) {
-
-		const lock_t*	wait_lock = lock;
-
-		/* Look for all transactions that could be waiting on this
-		transaction's locks. For that we need to search forward. */
-		if (lock_get_type_low(lock) == LOCK_REC) {
-
-			ulint	heap_no;
-
-			/* It's possible for heap_no to be undefined here.
-			This can happen during lock move from one page to
-			another when we split. */
-
-			heap_no = lock_rec_find_set_bit(lock);
-
-			do {
-				wait_lock = lock_rec_get_next(
-					heap_no, (lock_t*) wait_lock);
-
-				if (wait_lock != NULL
-				    && lock_has_to_wait(wait_lock, lock)) {
-
-					return(FALSE);
-				}
-			} while (wait_lock != NULL);
-		} else {
-			do {
-				wait_lock = UT_LIST_GET_NEXT(
-					un_member.tab_lock.locks, wait_lock);
-
-				if (wait_lock != NULL
-				    && lock_has_to_wait(wait_lock, lock) ) {
-
-					return(FALSE);
-				}
-			} while (wait_lock != NULL);
-		}
-	}
-
-	return(TRUE);
-}
-
 /********************************************************************//**
 Looks recursively for a deadlock.
 @return 0 if no deadlock found, LOCK_VICTIM_IS_START if there was a
@@ -3404,16 +3346,11 @@ lock_deadlock_recursive(
 		/* We have already exhaustively searched the subtree starting
 		from this trx */
 
-		return(0);
-	} else if (lock_trx_has_no_waiters(trx)) {
-		/* If no other transaction is waiting for this transaction
-		to release its locks then no deadlock can occur. */
 		return(0);
 	}
 
 	*cost = *cost + 1;
 
-
 	if (lock_get_type_low(wait_lock) == LOCK_REC) {
 		ulint		space;
 		ulint		page_no;

From 7f1dd013ec2e4317b5614b426fa4789c51f8a4d0 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:38:35 +0400
Subject: [PATCH 045/116] Applying InnoDB snapshot

Detailed revision comments:

r6591 | marko | 2010-02-08 10:06:39 +0200 (Mon, 08 Feb 2010) | 3 lines
branches/zip: row_merge_drop_index(): Remove redundant condition
on SYS_INDEXES.TABLE_ID.  INDEX_ID must be instance-widely unique,
because SYS_FIELDS is not indexed by TABLE_ID.
---
 storage/innodb_plugin/row/row0merge.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/storage/innodb_plugin/row/row0merge.c b/storage/innodb_plugin/row/row0merge.c
index a9bcb7df55c..7be9bd412a3 100644
--- a/storage/innodb_plugin/row/row0merge.c
+++ b/storage/innodb_plugin/row/row0merge.c
@@ -1977,14 +1977,12 @@ row_merge_drop_index(
 		/* Drop the field definitions of the index. */
 		"DELETE FROM SYS_FIELDS WHERE INDEX_ID = :indexid;\n"
 		/* Drop the index definition and the B-tree. */
-		"DELETE FROM SYS_INDEXES WHERE ID = :indexid\n"
-		"		AND TABLE_ID = :tableid;\n"
+		"DELETE FROM SYS_INDEXES WHERE ID = :indexid;\n"
 		"END;\n";
 
 	ut_ad(index && table && trx);
 
 	pars_info_add_dulint_literal(info, "indexid", index->id);
-	pars_info_add_dulint_literal(info, "tableid", table->id);
 
 	trx_start_if_not_started(trx);
 	trx->op_info = "dropping index";

From 41531fc8ca1ecafef20523e12501a71518a24bce Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:39:03 +0400
Subject: [PATCH 046/116] Applying InnoDB snapshot

Detailed revision comments:

r6594 | marko | 2010-02-08 12:55:04 +0200 (Mon, 08 Feb 2010) | 2 lines
branches/zip: rec_get_nth_field_offs_old():
Replace if (!cond) ut_error; tests with ut_a(cond).
---
 storage/innodb_plugin/rem/rem0rec.c | 16 +++-------------
 1 file changed, 3 insertions(+), 13 deletions(-)

diff --git a/storage/innodb_plugin/rem/rem0rec.c b/storage/innodb_plugin/rem/rem0rec.c
index 1c8b3fd8c1e..b2a8f45f89b 100644
--- a/storage/innodb_plugin/rem/rem0rec.c
+++ b/storage/innodb_plugin/rem/rem0rec.c
@@ -695,19 +695,9 @@ rec_get_nth_field_offs_old(
 	ulint	os;
 	ulint	next_os;
 
-	ut_ad(rec && len);
-	ut_ad(n < rec_get_n_fields_old(rec));
-
-	if (UNIV_UNLIKELY(n > REC_MAX_N_FIELDS)) {
-		fprintf(stderr, "Error: trying to access field %lu in rec\n",
-			(ulong) n);
-		ut_error;
-	}
-
-	if (UNIV_UNLIKELY(rec == NULL)) {
-		fputs("Error: rec is NULL pointer\n", stderr);
-		ut_error;
-	}
+	ut_ad(len);
+	ut_a(rec);
+	ut_a(n < rec_get_n_fields_old(rec));
 
 	if (rec_get_1byte_offs_flag(rec)) {
 		os = rec_1_get_field_start_offs(rec, n);

From def53688ec168b8eae096e2b2949bb62fd187d3c Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:39:29 +0400
Subject: [PATCH 047/116] Applying InnoDB snapshot

Detailed revision comments:

r6595 | marko | 2010-02-08 13:53:02 +0200 (Mon, 08 Feb 2010) | 1 line
branches/zip: btr_pcur_commit(): Unused function, remove.
---
 storage/innodb_plugin/include/btr0pcur.h  |  9 +--------
 storage/innodb_plugin/include/btr0pcur.ic | 19 +------------------
 2 files changed, 2 insertions(+), 26 deletions(-)

diff --git a/storage/innodb_plugin/include/btr0pcur.h b/storage/innodb_plugin/include/btr0pcur.h
index 7d8bac85f3e..509e61b9c8d 100644
--- a/storage/innodb_plugin/include/btr0pcur.h
+++ b/storage/innodb_plugin/include/btr0pcur.h
@@ -281,20 +281,13 @@ btr_pcur_get_mtr(
 /*=============*/
 	btr_pcur_t*	cursor);	/*!< in: persistent cursor */
 /**************************************************************//**
-Commits the pcur mtr and sets the pcur latch mode to BTR_NO_LATCHES,
+Commits the mtr and sets the pcur latch mode to BTR_NO_LATCHES,
 that is, the cursor becomes detached. If there have been modifications
 to the page where pcur is positioned, this can be used instead of
 btr_pcur_release_leaf. Function btr_pcur_store_position should be used
 before calling this, if restoration of cursor is wanted later. */
 UNIV_INLINE
 void
-btr_pcur_commit(
-/*============*/
-	btr_pcur_t*	pcur);	/*!< in: persistent cursor */
-/**************************************************************//**
-Differs from btr_pcur_commit in that we can specify the mtr to commit. */
-UNIV_INLINE
-void
 btr_pcur_commit_specify_mtr(
 /*========================*/
 	btr_pcur_t*	pcur,	/*!< in: persistent cursor */
diff --git a/storage/innodb_plugin/include/btr0pcur.ic b/storage/innodb_plugin/include/btr0pcur.ic
index b11d7dce21c..68acb68c0d3 100644
--- a/storage/innodb_plugin/include/btr0pcur.ic
+++ b/storage/innodb_plugin/include/btr0pcur.ic
@@ -395,30 +395,13 @@ btr_pcur_move_to_next(
 }
 
 /**************************************************************//**
-Commits the pcur mtr and sets the pcur latch mode to BTR_NO_LATCHES,
+Commits the mtr and sets the pcur latch mode to BTR_NO_LATCHES,
 that is, the cursor becomes detached. If there have been modifications
 to the page where pcur is positioned, this can be used instead of
 btr_pcur_release_leaf. Function btr_pcur_store_position should be used
 before calling this, if restoration of cursor is wanted later. */
 UNIV_INLINE
 void
-btr_pcur_commit(
-/*============*/
-	btr_pcur_t*	pcur)	/*!< in: persistent cursor */
-{
-	ut_a(pcur->pos_state == BTR_PCUR_IS_POSITIONED);
-
-	pcur->latch_mode = BTR_NO_LATCHES;
-
-	mtr_commit(pcur->mtr);
-
-	pcur->pos_state = BTR_PCUR_WAS_POSITIONED;
-}
-
-/**************************************************************//**
-Differs from btr_pcur_commit in that we can specify the mtr to commit. */
-UNIV_INLINE
-void
 btr_pcur_commit_specify_mtr(
 /*========================*/
 	btr_pcur_t*	pcur,	/*!< in: persistent cursor */

From 41bd9dfe618023386afef10e34e97704a10efb51 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:39:56 +0400
Subject: [PATCH 048/116] Applying InnoDB snapshot

Detailed revision comments:

r6608 | marko | 2010-02-09 11:02:37 +0200 (Tue, 09 Feb 2010) | 1 line
branches/zip: ha_innobase::add_index(): Check for !innodb_table.
---
 storage/innodb_plugin/handler/handler0alter.cc | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/storage/innodb_plugin/handler/handler0alter.cc b/storage/innodb_plugin/handler/handler0alter.cc
index 47999ae37f8..fbbcfd6149b 100644
--- a/storage/innodb_plugin/handler/handler0alter.cc
+++ b/storage/innodb_plugin/handler/handler0alter.cc
@@ -656,6 +656,11 @@ ha_innobase::add_index(
 	innodb_table = indexed_table
 		= dict_table_get(prebuilt->table->name, FALSE);
 
+	if (UNIV_UNLIKELY(!innodb_table)) {
+		error = HA_ERR_NO_SUCH_TABLE;
+		goto err_exit;
+	}
+
 	/* Check if the index name is reserved. */
 	if (innobase_index_name_is_reserved(trx, key_info, num_of_keys)) {
 		error = -1;

From a9b1a67b2928931debfe49e15160aa7021dcfdfe Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:40:13 +0400
Subject: [PATCH 049/116] Applying InnoDB snapshot

Detailed revision comments:

r6609 | marko | 2010-02-09 13:45:40 +0200 (Tue, 09 Feb 2010) | 1 line
branches/zip: dict_field_print_low(): Add const qualifier.
---
 storage/innodb_plugin/dict/dict0dict.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/storage/innodb_plugin/dict/dict0dict.c b/storage/innodb_plugin/dict/dict0dict.c
index ca129c29d20..35b856cd79f 100644
--- a/storage/innodb_plugin/dict/dict0dict.c
+++ b/storage/innodb_plugin/dict/dict0dict.c
@@ -140,7 +140,7 @@ static
 void
 dict_field_print_low(
 /*=================*/
-	dict_field_t*	field);	/*!< in: field */
+	const dict_field_t*	field);	/*!< in: field */
 /*********************************************************************//**
 Frees a foreign key struct. */
 static
@@ -4403,7 +4403,7 @@ static
 void
 dict_field_print_low(
 /*=================*/
-	dict_field_t*	field)	/*!< in: field */
+	const dict_field_t*	field)	/*!< in: field */
 {
 	ut_ad(mutex_own(&(dict_sys->mutex)));
 

From cb10d756b4ec46ea3a5ca4dfa671712047cd1d17 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:40:44 +0400
Subject: [PATCH 050/116] Applying InnoDB snapshot

Detailed revision comments:

r6610 | marko | 2010-02-09 13:53:59 +0200 (Tue, 09 Feb 2010) | 17 lines
branches/zip: When dropping temporary indexes and tables at startup,
first load them to the data dictionary cache and use the normal
routines for dropping tables or indexes.  This should reduce the
risk of bugs and also make the code compatible with the upcoming
TablespaceDictionary implementation.

DICT_SYS_INDEXES_NAME_FIELD: The clustered index position of SYS_INDEXES.NAME.

row_merge_drop_temp_indexes(): Scan SYS_INDEXES for tables containing
temporary indexes, and load the tables as needed. Invoke
row_merge_drop_index() to drop the indexes.

row_mysql_drop_temp_tables(): Scan SYS_TABLES for temporary tables,
load them with dict_load_table() and drop them with
row_drop_table_for_mysql().

rb://251, not yet reviewed
---
 storage/innodb_plugin/dict/dict0boot.c    |   5 +-
 storage/innodb_plugin/include/dict0boot.h |   1 +
 storage/innodb_plugin/row/row0merge.c     |  93 ++++++++++-----
 storage/innodb_plugin/row/row0mysql.c     | 133 ++++++++++------------
 4 files changed, 134 insertions(+), 98 deletions(-)

diff --git a/storage/innodb_plugin/dict/dict0boot.c b/storage/innodb_plugin/dict/dict0boot.c
index 8f948c06c51..c1df2bac0bf 100644
--- a/storage/innodb_plugin/dict/dict0boot.c
+++ b/storage/innodb_plugin/dict/dict0boot.c
@@ -358,7 +358,7 @@ dict_boot(void)
 	dict_mem_table_add_col(table, heap, "SPACE", DATA_INT, 0, 4);
 	dict_mem_table_add_col(table, heap, "PAGE_NO", DATA_INT, 0, 4);
 
-	/* The '+ 2' below comes from the 2 system fields */
+	/* The '+ 2' below comes from the fields DB_TRX_ID, DB_ROLL_PTR */
 #if DICT_SYS_INDEXES_PAGE_NO_FIELD != 6 + 2
 #error "DICT_SYS_INDEXES_PAGE_NO_FIELD != 6 + 2"
 #endif
@@ -367,6 +367,9 @@ dict_boot(void)
 #endif
 #if DICT_SYS_INDEXES_TYPE_FIELD != 4 + 2
 #error "DICT_SYS_INDEXES_TYPE_FIELD != 4 + 2"
+#endif
+#if DICT_SYS_INDEXES_NAME_FIELD != 1 + 2
+#error "DICT_SYS_INDEXES_NAME_FIELD != 1 + 2"
 #endif
 
 	table->id = DICT_INDEXES_ID;
diff --git a/storage/innodb_plugin/include/dict0boot.h b/storage/innodb_plugin/include/dict0boot.h
index 51d37ee98d1..404fc23bfa3 100644
--- a/storage/innodb_plugin/include/dict0boot.h
+++ b/storage/innodb_plugin/include/dict0boot.h
@@ -137,6 +137,7 @@ clustered index */
 #define DICT_SYS_INDEXES_PAGE_NO_FIELD	 8
 #define DICT_SYS_INDEXES_SPACE_NO_FIELD	 7
 #define DICT_SYS_INDEXES_TYPE_FIELD	 6
+#define DICT_SYS_INDEXES_NAME_FIELD	 3
 
 /* When a row id which is zero modulo this number (which must be a power of
 two) is assigned, the field DICT_HDR_ROW_ID on the dictionary header page is
diff --git a/storage/innodb_plugin/row/row0merge.c b/storage/innodb_plugin/row/row0merge.c
index 7be9bd412a3..1608751dc89 100644
--- a/storage/innodb_plugin/row/row0merge.c
+++ b/storage/innodb_plugin/row/row0merge.c
@@ -2031,38 +2031,79 @@ row_merge_drop_temp_indexes(void)
 /*=============================*/
 {
 	trx_t*		trx;
-	ulint		err;
-
-	/* We use the private SQL parser of Innobase to generate the
-	query graphs needed in deleting the dictionary data from system
-	tables in Innobase. Deleting a row from SYS_INDEXES table also
-	frees the file segments of the B-tree associated with the index. */
-	static const char drop_temp_indexes[] =
-		"PROCEDURE DROP_TEMP_INDEXES_PROC () IS\n"
-		"indexid CHAR;\n"
-		"DECLARE CURSOR c IS SELECT ID FROM SYS_INDEXES\n"
-		"WHERE SUBSTR(NAME,0,1)='" TEMP_INDEX_PREFIX_STR "';\n"
-		"BEGIN\n"
-		"\tOPEN c;\n"
-		"\tWHILE 1=1 LOOP\n"
-		"\t\tFETCH c INTO indexid;\n"
-		"\t\tIF (SQL % NOTFOUND) THEN\n"
-		"\t\t\tEXIT;\n"
-		"\t\tEND IF;\n"
-		"\t\tDELETE FROM SYS_FIELDS WHERE INDEX_ID = indexid;\n"
-		"\t\tDELETE FROM SYS_INDEXES WHERE ID = indexid;\n"
-		"\tEND LOOP;\n"
-		"\tCLOSE c;\n"
-		"\tCOMMIT WORK;\n"
-		"END;\n";
+	btr_pcur_t	pcur;
+	mtr_t		mtr;
 
+	/* Load the table definitions that contain partially defined
+	indexes, so that the data dictionary information can be checked
+	when accessing the tablename.ibd files. */
 	trx = trx_allocate_for_background();
 	trx->op_info = "dropping partially created indexes";
 	row_mysql_lock_data_dictionary(trx);
 
-	err = que_eval_sql(NULL, drop_temp_indexes, FALSE, trx);
-	ut_a(err == DB_SUCCESS);
+	mtr_start(&mtr);
 
+	btr_pcur_open_at_index_side(
+		TRUE,
+		dict_table_get_first_index(dict_sys->sys_indexes),
+		BTR_SEARCH_LEAF, &pcur, TRUE, &mtr);
+
+	for (;;) {
+		const rec_t*	rec;
+		const byte*	field;
+		ulint		len;
+		dulint		table_id;
+		dict_table_t*	table;
+
+		btr_pcur_move_to_next_user_rec(&pcur, &mtr);
+
+		if (!btr_pcur_is_on_user_rec(&pcur)) {
+			break;
+		}
+
+		rec = btr_pcur_get_rec(&pcur);
+		field = rec_get_nth_field_old(rec, DICT_SYS_INDEXES_NAME_FIELD,
+					      &len);
+		if (len == UNIV_SQL_NULL || len == 0
+		    || mach_read_from_1(field) != (ulint) TEMP_INDEX_PREFIX) {
+			continue;
+		}
+
+		/* This is a temporary index. */
+
+		field = rec_get_nth_field_old(rec, 0/*TABLE_ID*/, &len);
+		if (len != 8) {
+			/* Corrupted TABLE_ID */
+			continue;
+		}
+
+		table_id = mach_read_from_8(field);
+
+		btr_pcur_store_position(&pcur, &mtr);
+		btr_pcur_commit_specify_mtr(&pcur, &mtr);
+
+		table = dict_load_table_on_id(table_id);
+
+		if (table) {
+			dict_index_t*	index;
+
+			for (index = dict_table_get_first_index(table);
+			     index; index = dict_table_get_next_index(index)) {
+
+				if (*index->name == TEMP_INDEX_PREFIX) {
+					row_merge_drop_index(index, table, trx);
+					trx_commit_for_mysql(trx);
+				}
+			}
+		}
+
+		mtr_start(&mtr);
+		btr_pcur_restore_position(BTR_SEARCH_LEAF,
+					  &pcur, &mtr);
+	}
+
+	btr_pcur_close(&pcur);
+	mtr_commit(&mtr);
 	row_mysql_unlock_data_dictionary(trx);
 	trx_free_for_background(trx);
 }
diff --git a/storage/innodb_plugin/row/row0mysql.c b/storage/innodb_plugin/row/row0mysql.c
index ad78c559141..0223a389b5d 100644
--- a/storage/innodb_plugin/row/row0mysql.c
+++ b/storage/innodb_plugin/row/row0mysql.c
@@ -3370,88 +3370,79 @@ void
 row_mysql_drop_temp_tables(void)
 /*============================*/
 {
-	trx_t*	trx;
-	ulint	err;
+	trx_t*		trx;
+	btr_pcur_t	pcur;
+	mtr_t		mtr;
+	mem_heap_t*	heap;
 
 	trx = trx_allocate_for_background();
 	trx->op_info = "dropping temporary tables";
 	row_mysql_lock_data_dictionary(trx);
 
-	err = que_eval_sql(
-		NULL,
-		"PROCEDURE DROP_TEMP_TABLES_PROC () IS\n"
-		"table_name CHAR;\n"
-		"table_id CHAR;\n"
-		"foreign_id CHAR;\n"
-		"index_id CHAR;\n"
-		"DECLARE CURSOR c IS SELECT NAME,ID FROM SYS_TABLES\n"
-		"WHERE N_COLS > 2147483647\n"
-		/* N_COLS>>31 is set unless ROW_FORMAT=REDUNDANT,
-		and MIX_LEN may be garbage for those tables */
-		"AND MIX_LEN=(MIX_LEN/2*2+1);\n"
-		/* MIX_LEN & 1 is set for temporary tables */
-#if DICT_TF2_TEMPORARY != 1
-# error "DICT_TF2_TEMPORARY != 1"
-#endif
-		"BEGIN\n"
-		"OPEN c;\n"
-		"WHILE 1=1 LOOP\n"
-		"	FETCH c INTO table_name, table_id;\n"
-		"	IF (SQL % NOTFOUND) THEN\n"
-		"		EXIT;\n"
-		"	END IF;\n"
-		"	WHILE 1=1 LOOP\n"
-		"		SELECT ID INTO index_id\n"
-		"		FROM SYS_INDEXES\n"
-		"		WHERE TABLE_ID = table_id\n"
-		"		LOCK IN SHARE MODE;\n"
-		"		IF (SQL % NOTFOUND) THEN\n"
-		"			EXIT;\n"
-		"		END IF;\n"
+	heap = mem_heap_create(200);
 
-		/* Do not drop tables for which there exist
-		foreign key constraints. */
-		"		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;"
-		"		IF NOT (SQL % NOTFOUND) THEN\n"
-		"			EXIT;\n"
-		"		END IF;\n"
+	mtr_start(&mtr);
 
-		"		SELECT ID INTO foreign_id\n"
-		"		FROM SYS_FOREIGN\n"
-		"		WHERE REF_NAME = table_name\n"
-		"		AND TO_BINARY(REF_NAME)\n"
-		"		  = TO_BINARY(table_name)\n;"
-		"		IF NOT (SQL % NOTFOUND) THEN\n"
-		"			EXIT;\n"
-		"		END IF;\n"
+	btr_pcur_open_at_index_side(
+		TRUE,
+		dict_table_get_first_index(dict_sys->sys_tables),
+		BTR_SEARCH_LEAF, &pcur, TRUE, &mtr);
 
-		"		DELETE FROM SYS_FIELDS\n"
-		"		WHERE INDEX_ID = index_id;\n"
-		"		DELETE FROM SYS_INDEXES\n"
-		"		WHERE ID = index_id\n"
-		"		AND TABLE_ID = table_id;\n"
-		"	END LOOP;\n"
-		"	DELETE FROM SYS_COLUMNS\n"
-		"	WHERE TABLE_ID = table_id;\n"
-		"	DELETE FROM SYS_TABLES\n"
-		"	WHERE ID = table_id;\n"
-		"END LOOP;\n"
-		"COMMIT WORK;\n"
-		"END;\n"
-		, FALSE, trx);
+	for (;;) {
+		const rec_t*	rec;
+		const byte*	field;
+		ulint		len;
+		const char*	table_name;
+		dict_table_t*	table;
 
-	if (err != DB_SUCCESS) {
-		ut_print_timestamp(stderr);
-		fprintf(stderr,
-			"  InnoDB: Failed to drop temporary tables:"
-			" error %lu occurred\n",
-			(ulong) err);
+		btr_pcur_move_to_next_user_rec(&pcur, &mtr);
+
+		if (!btr_pcur_is_on_user_rec(&pcur)) {
+			break;
+		}
+
+		rec = btr_pcur_get_rec(&pcur);
+		field = rec_get_nth_field_old(rec, 4/*N_COLS*/, &len);
+		if (len != 4 || !(mach_read_from_4(field) & 0x80000000UL)) {
+			continue;
+		}
+
+		/* Because this is not a ROW_FORMAT=REDUNDANT table,
+		the is_temp flag is valid.  Examine it. */
+
+		field = rec_get_nth_field_old(rec, 7/*MIX_LEN*/, &len);
+		if (len != 4
+		    || !(mach_read_from_4(field) & DICT_TF2_TEMPORARY)) {
+			continue;
+		}
+
+		/* This is a temporary table. */
+		field = rec_get_nth_field_old(rec, 0/*NAME*/, &len);
+		if (len == UNIV_SQL_NULL || len == 0) {
+			/* Corrupted SYS_TABLES.NAME */
+			continue;
+		}
+
+		table_name = mem_heap_strdupl(heap, (const char*) field, len);
+
+		btr_pcur_store_position(&pcur, &mtr);
+		btr_pcur_commit_specify_mtr(&pcur, &mtr);
+
+		table = dict_load_table(table_name);
+
+		if (table) {
+			row_drop_table_for_mysql(table_name, trx, FALSE);
+			trx_commit_for_mysql(trx);
+		}
+
+		mtr_start(&mtr);
+		btr_pcur_restore_position(BTR_SEARCH_LEAF,
+					  &pcur, &mtr);
 	}
 
+	btr_pcur_close(&pcur);
+	mtr_commit(&mtr);
+	mem_heap_free(heap);
 	row_mysql_unlock_data_dictionary(trx);
 	trx_free_for_background(trx);
 }

From 752aaebeb92f7a26af79835132611fddae17e046 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:41:04 +0400
Subject: [PATCH 051/116] Applying InnoDB snapshot

Detailed revision comments:

r6611 | marko | 2010-02-09 14:28:25 +0200 (Tue, 09 Feb 2010) | 11 lines
branches/zip: Roll back dictionary transaction(s) before scanning *.ibd files

innobase_start_or_create_for_mysql(): Roll back data dictionary
transactions before scanning the *.ibd files. Then, data dictionary
records can be loaded to the cache before opening the *.ibd files.

recv_recovery_rollback_active(): Refactored from
recv_recovery_from_checkpoint_finish().

rb://235, committing without review, because this is needed for
TablespaceDictionary.
---
 storage/innodb_plugin/include/log0recv.h |  6 ++++++
 storage/innodb_plugin/log/log0recv.c     | 12 ++++++++++--
 storage/innodb_plugin/srv/srv0start.c    | 24 ++++++++++--------------
 3 files changed, 26 insertions(+), 16 deletions(-)

diff --git a/storage/innodb_plugin/include/log0recv.h b/storage/innodb_plugin/include/log0recv.h
index 35576bb579d..3cbb85bbec8 100644
--- a/storage/innodb_plugin/include/log0recv.h
+++ b/storage/innodb_plugin/include/log0recv.h
@@ -176,6 +176,12 @@ UNIV_INTERN
 void
 recv_recovery_from_checkpoint_finish(void);
 /*======================================*/
+/********************************************************//**
+Initiates the rollback of active transactions. */
+UNIV_INTERN
+void
+recv_recovery_rollback_active(void);
+/*===============================*/
 /*******************************************************//**
 Scans log from a buffer and stores new log data to the parsing buffer.
 Parses and hashes the log records if new data found.  Unless
diff --git a/storage/innodb_plugin/log/log0recv.c b/storage/innodb_plugin/log/log0recv.c
index b060c9e6e24..b14404ca1da 100644
--- a/storage/innodb_plugin/log/log0recv.c
+++ b/storage/innodb_plugin/log/log0recv.c
@@ -3212,8 +3212,6 @@ void
 recv_recovery_from_checkpoint_finish(void)
 /*======================================*/
 {
-	int		i;
-
 	/* Apply the hashed log records to the respective file pages */
 
 	if (srv_force_recovery < SRV_FORCE_NO_LOG_REDO) {
@@ -3261,6 +3259,16 @@ recv_recovery_from_checkpoint_finish(void)
 	The data dictionary latch should guarantee that there is at
 	most one data dictionary transaction active at a time. */
 	trx_rollback_or_clean_recovered(FALSE);
+}
+
+/********************************************************//**
+Initiates the rollback of active transactions. */
+UNIV_INTERN
+void
+recv_recovery_rollback_active(void)
+/*===============================*/
+{
+	int		i;
 
 	/* Drop partially created indexes. */
 	row_merge_drop_temp_indexes();
diff --git a/storage/innodb_plugin/srv/srv0start.c b/storage/innodb_plugin/srv/srv0start.c
index 71e84d1c185..3293fbb5bb7 100644
--- a/storage/innodb_plugin/srv/srv0start.c
+++ b/storage/innodb_plugin/srv/srv0start.c
@@ -1570,6 +1570,14 @@ innobase_start_or_create_for_mysql(void)
 		dict_boot();
 		trx_sys_init_at_db_start();
 
+		/* Initialize the fsp free limit global variable in the log
+		system */
+		fsp_header_get_free_limit();
+
+		/* recv_recovery_from_checkpoint_finish needs trx lists which
+		are initialized in trx_sys_init_at_db_start(). */
+
+		recv_recovery_from_checkpoint_finish();
 		if (srv_force_recovery < SRV_FORCE_NO_IBUF_MERGE) {
 			/* The following call is necessary for the insert
 			buffer to work with multiple tablespaces. We must
@@ -1585,26 +1593,14 @@ innobase_start_or_create_for_mysql(void)
 			every table in the InnoDB data dictionary that has
 			an .ibd file.
 
-			We also determine the maximum tablespace id used.
-
-			TODO: We may have incomplete transactions in the
-			data dictionary tables. Does that harm the scanning of
-			the data dictionary below? */
+			We also determine the maximum tablespace id used. */
 
 			dict_check_tablespaces_and_store_max_id(
 				recv_needed_recovery);
 		}
 
 		srv_startup_is_before_trx_rollback_phase = FALSE;
-
-		/* Initialize the fsp free limit global variable in the log
-		system */
-		fsp_header_get_free_limit();
-
-		/* recv_recovery_from_checkpoint_finish needs trx lists which
-		are initialized in trx_sys_init_at_db_start(). */
-
-		recv_recovery_from_checkpoint_finish();
+		recv_recovery_rollback_active();
 
 		/* It is possible that file_format tag has never
 		been set. In this case we initialize it to minimum

From 41c82f8897290ad0c25e8065602e182540752cc2 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:41:23 +0400
Subject: [PATCH 052/116] Applying InnoDB snapshot

Detailed revision comments:

r6612 | marko | 2010-02-09 14:32:39 +0200 (Tue, 09 Feb 2010) | 3 lines
branches/zip: recv_recovery_rollback_active():
Drop the temporary tables and indexes after enabling sync order checks.
This should not make any difference. This could have been done in r6611.
---
 storage/innodb_plugin/log/log0recv.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/storage/innodb_plugin/log/log0recv.c b/storage/innodb_plugin/log/log0recv.c
index b14404ca1da..eb20cba7e32 100644
--- a/storage/innodb_plugin/log/log0recv.c
+++ b/storage/innodb_plugin/log/log0recv.c
@@ -3270,11 +3270,6 @@ recv_recovery_rollback_active(void)
 {
 	int		i;
 
-	/* Drop partially created indexes. */
-	row_merge_drop_temp_indexes();
-	/* Drop temporary tables. */
-	row_mysql_drop_temp_tables();
-
 #ifdef UNIV_SYNC_DEBUG
 	/* Wait for a while so that created threads have time to suspend
 	themselves before we switch the latching order checks on */
@@ -3283,6 +3278,11 @@ recv_recovery_rollback_active(void)
 	/* Switch latching order checks on in sync0sync.c */
 	sync_order_checks_on = TRUE;
 #endif
+	/* Drop partially created indexes. */
+	row_merge_drop_temp_indexes();
+	/* Drop temporary tables. */
+	row_mysql_drop_temp_tables();
+
 	if (srv_force_recovery < SRV_FORCE_NO_TRX_UNDO) {
 		/* Rollback the uncommitted transactions which have no user
 		session */

From 6833b977659537cf72a500c585b65099d00c808a Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:41:47 +0400
Subject: [PATCH 053/116] Applying InnoDB snapshot

Detailed revision comments:

r6614 | inaam | 2010-02-09 20:26:23 +0200 (Tue, 09 Feb 2010) | 7 lines
branches/plugin rb://242

Let the master thread sleep if the amount of work to be done is
calibrated as taking less than a second.

Approved by: Heikki
---
 storage/innodb_plugin/srv/srv0srv.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/srv/srv0srv.c b/storage/innodb_plugin/srv/srv0srv.c
index 23ec3b8db4c..1dba8f3ba2b 100644
--- a/storage/innodb_plugin/srv/srv0srv.c
+++ b/storage/innodb_plugin/srv/srv0srv.c
@@ -2524,7 +2524,10 @@ loop:
 						BUF_FLUSH_LIST,
 						n_flush,
 						IB_ULONGLONG_MAX);
-				skip_sleep = TRUE;
+
+				if (n_flush == PCT_IO(100)) {
+					skip_sleep = TRUE;
+				}
 			}
 		}
 

From e2bf3f2c6976b02e7cfdf32615c82747f99f4329 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:42:04 +0400
Subject: [PATCH 054/116] Applying InnoDB snapshot

Detailed revision comments:

r6631 | marko | 2010-02-10 09:19:52 +0200 (Wed, 10 Feb 2010) | 1 line
branches/zip: Document r6614 in ChangeLog.
---
 storage/innodb_plugin/ChangeLog | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index d727501ea9a..e6a52543798 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,9 @@
+2010-02-09	The InnoDB Team
+
+	* srv/srv0srv.c:
+	Let the master thread sleep if the amount of work to be done is
+	calibrated as taking less than a second.
+
 2010-02-04	The InnoDB Team
 
 	* btr/btr0btr.c, btr/btr0cur.c, btr/btr0pcur.c, buf/buf0buf.c,

From bcdd00576a0b8c684b318147a93aeb5578ebe38c Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:42:24 +0400
Subject: [PATCH 055/116] Applying InnoDB snapshot

Detailed revision comments:

r6633 | marko | 2010-02-10 10:40:55 +0200 (Wed, 10 Feb 2010) | 31 lines
branches/zip: Merge revisions 6538:6613 from branches/5.1:

  ------------------------------------------------------------------------
  r6545 | jyang | 2010-02-03 03:57:32 +0200 (Wed, 03 Feb 2010) | 8 lines
  Changed paths:
     M /branches/5.1/lock/lock0lock.c

  branches/5.1: Fix bug #49001, "SHOW INNODB STATUS deadlock info
  incorrect when deadlock detection aborts". Print the correct
  lock owner when recursive function lock_deadlock_recursive()
  exceeds its maximum depth LOCK_MAX_DEPTH_IN_DEADLOCK_CHECK.

  rb://217, approved by Marko.
  ------------------------------------------------------------------------
  r6613 | inaam | 2010-02-09 20:23:09 +0200 (Tue, 09 Feb 2010) | 11 lines
  Changed paths:
     M /branches/5.1/buf/buf0buf.c
     M /branches/5.1/buf/buf0rea.c
     M /branches/5.1/include/buf0rea.h

  branches/5.1: Fix Bug #38901
  InnoDB logs error repeatedly when trying to load page into buffer pool

  In buf_page_get_gen() if we are unable to read a page (because of
  corruption or some other reason) we keep on retrying. This fills up
  error log with millions of entries in no time and we'd eventually run
  out of disk space. This patch limits the number of attempts that we
  make (currently set to 100) and after that we abort with a message.

  rb://241 Approved by: Heikki
  ------------------------------------------------------------------------
---
 storage/innodb_plugin/ChangeLog        | 12 ++++
 storage/innodb_plugin/buf/buf0buf.c    | 27 +++++++-
 storage/innodb_plugin/lock/lock0lock.c | 96 ++++++++++++++++----------
 3 files changed, 99 insertions(+), 36 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index e6a52543798..c66c4b01cda 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,9 @@
+2010-02-09	The InnoDB Team
+
+	* buf/buf0buf.c:
+	Fix Bug #38901 InnoDB logs error repeatedly when trying to load
+	page into buffer pool
+
 2010-02-09	The InnoDB Team
 
 	* srv/srv0srv.c:
@@ -13,6 +19,12 @@
 	b-tree cursor functions to the buffer pool requests, in order
 	to make the latch diagnostics more accurate.
 
+2010-02-03	The InnoDB Team
+
+	* lock/lock0lock.c:
+	Fix Bug#49001 SHOW INNODB STATUS deadlock info incorrect
+	when deadlock detection aborts
+
 2010-02-03	The InnoDB Team
 
 	* buf/buf0lru.c:
diff --git a/storage/innodb_plugin/buf/buf0buf.c b/storage/innodb_plugin/buf/buf0buf.c
index c4afff1e3e7..5312f772213 100644
--- a/storage/innodb_plugin/buf/buf0buf.c
+++ b/storage/innodb_plugin/buf/buf0buf.c
@@ -242,6 +242,8 @@ the read requests for the whole area.
 #ifndef UNIV_HOTBACKUP
 /** Value in microseconds */
 static const int WAIT_FOR_READ	= 5000;
+/** Number of attemtps made to read in a page in the buffer pool */
+static const ulint BUF_PAGE_READ_MAX_RETRIES = 100;
 
 /** The buffer buf_pool of the database */
 UNIV_INTERN buf_pool_t*	buf_pool = NULL;
@@ -2036,6 +2038,7 @@ buf_page_get_gen(
 	unsigned	access_time;
 	ulint		fix_type;
 	ibool		must_read;
+	ulint		retries = 0;
 
 	ut_ad(mtr);
 	ut_ad((rw_latch == RW_S_LATCH)
@@ -2090,7 +2093,29 @@ loop2:
 			return(NULL);
 		}
 
-		buf_read_page(space, zip_size, offset);
+		if (buf_read_page(space, zip_size, offset)) {
+			retries = 0;
+		} else if (retries < BUF_PAGE_READ_MAX_RETRIES) {
+			++retries;
+		} else {
+			fprintf(stderr, "InnoDB: Error: Unable"
+				" to read tablespace %lu page no"
+				" %lu into the buffer pool after"
+				" %lu attempts\n"
+				"InnoDB: The most probable cause"
+				" of this error may be that the"
+				" table has been corrupted.\n"
+				"InnoDB: You can try to fix this"
+				" problem by using"
+				" innodb_force_recovery.\n"
+				"InnoDB: Please see reference manual"
+				" for more details.\n"
+				"InnoDB: Aborting...\n",
+				space, offset,
+				BUF_PAGE_READ_MAX_RETRIES);
+
+			ut_error;
+		}
 
 #if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
 		ut_a(++buf_dbg_counter % 37 || buf_validate());
diff --git a/storage/innodb_plugin/lock/lock0lock.c b/storage/innodb_plugin/lock/lock0lock.c
index 11a5232b0dc..1de612ef637 100644
--- a/storage/innodb_plugin/lock/lock0lock.c
+++ b/storage/innodb_plugin/lock/lock0lock.c
@@ -376,6 +376,7 @@ UNIV_INTERN FILE*	lock_latest_err_file;
 /* Flags for recursive deadlock search */
 #define LOCK_VICTIM_IS_START	1
 #define LOCK_VICTIM_IS_OTHER	2
+#define LOCK_EXCEED_MAX_DEPTH	3
 
 /********************************************************************//**
 Checks if a lock request results in a deadlock.
@@ -394,7 +395,8 @@ Looks recursively for a deadlock.
 deadlock and we chose 'start' as the victim, LOCK_VICTIM_IS_OTHER if a
 deadlock was found and we chose some other trx as a victim: we must do
 the search again in this last case because there may be another
-deadlock! */
+deadlock!
+LOCK_EXCEED_MAX_DEPTH if the lock search exceeds max steps or max depth. */
 static
 ulint
 lock_deadlock_recursive(
@@ -404,10 +406,10 @@ lock_deadlock_recursive(
 	lock_t*	wait_lock,	/*!< in:  lock that is waiting to be granted */
 	ulint*	cost,		/*!< in/out: number of calculation steps thus
 				far: if this exceeds LOCK_MAX_N_STEPS_...
-				we return LOCK_VICTIM_IS_START */
+				we return LOCK_EXCEED_MAX_DEPTH */
 	ulint	depth);		/*!< in: recursion depth: if this exceeds
 				LOCK_MAX_DEPTH_IN_DEADLOCK_CHECK, we
-				return LOCK_VICTIM_IS_START */
+				return LOCK_EXCEED_MAX_DEPTH */
 
 /*********************************************************************//**
 Gets the nth bit of a record lock.
@@ -3261,8 +3263,6 @@ lock_deadlock_occurs(
 	lock_t*	lock,	/*!< in: lock the transaction is requesting */
 	trx_t*	trx)	/*!< in: transaction */
 {
-	dict_table_t*	table;
-	dict_index_t*	index;
 	trx_t*		mark_trx;
 	ulint		ret;
 	ulint		cost	= 0;
@@ -3284,31 +3284,50 @@ retry:
 
 	ret = lock_deadlock_recursive(trx, trx, lock, &cost, 0);
 
-	if (ret == LOCK_VICTIM_IS_OTHER) {
+	switch (ret) {
+	case LOCK_VICTIM_IS_OTHER:
 		/* We chose some other trx as a victim: retry if there still
 		is a deadlock */
-
 		goto retry;
-	}
 
-	if (UNIV_UNLIKELY(ret == LOCK_VICTIM_IS_START)) {
-		if (lock_get_type_low(lock) & LOCK_TABLE) {
-			table = lock->un_member.tab_lock.table;
-			index = NULL;
-		} else {
-			index = lock->index;
-			table = index->table;
-		}
+	case LOCK_EXCEED_MAX_DEPTH:
+		/* If the lock search exceeds the max step
+		or the max depth, the current trx will be
+		the victim. Print its information. */
+		rewind(lock_latest_err_file);
+		ut_print_timestamp(lock_latest_err_file);
 
-		lock_deadlock_found = TRUE;
-
-		fputs("*** WE ROLL BACK TRANSACTION (2)\n",
+		fputs("TOO DEEP OR LONG SEARCH IN THE LOCK TABLE"
+		      " WAITS-FOR GRAPH, WE WILL ROLL BACK"
+		      " FOLLOWING TRANSACTION \n",
 		      lock_latest_err_file);
 
-		return(TRUE);
+		fputs("\n*** TRANSACTION:\n", lock_latest_err_file);
+		      trx_print(lock_latest_err_file, trx, 3000);
+
+		fputs("*** WAITING FOR THIS LOCK TO BE GRANTED:\n",
+		      lock_latest_err_file);
+
+		if (lock_get_type(lock) == LOCK_REC) {
+			lock_rec_print(lock_latest_err_file, lock);
+		} else {
+			lock_table_print(lock_latest_err_file, lock);
+		}
+		break;
+
+	case LOCK_VICTIM_IS_START:
+		fputs("*** WE ROLL BACK TRANSACTION (2)\n",
+		      lock_latest_err_file);
+		break;
+
+	default:
+		/* No deadlock detected*/
+		return(FALSE);
 	}
 
-	return(FALSE);
+	lock_deadlock_found = TRUE;
+
+	return(TRUE);
 }
 
 /********************************************************************//**
@@ -3317,7 +3336,8 @@ Looks recursively for a deadlock.
 deadlock and we chose 'start' as the victim, LOCK_VICTIM_IS_OTHER if a
 deadlock was found and we chose some other trx as a victim: we must do
 the search again in this last case because there may be another
-deadlock! */
+deadlock!
+LOCK_EXCEED_MAX_DEPTH if the lock search exceeds max steps or max depth. */
 static
 ulint
 lock_deadlock_recursive(
@@ -3327,10 +3347,10 @@ lock_deadlock_recursive(
 	lock_t*	wait_lock,	/*!< in: lock that is waiting to be granted */
 	ulint*	cost,		/*!< in/out: number of calculation steps thus
 				far: if this exceeds LOCK_MAX_N_STEPS_...
-				we return LOCK_VICTIM_IS_START */
+				we return LOCK_EXCEED_MAX_DEPTH */
 	ulint	depth)		/*!< in: recursion depth: if this exceeds
 				LOCK_MAX_DEPTH_IN_DEADLOCK_CHECK, we
-				return LOCK_VICTIM_IS_START */
+				return LOCK_EXCEED_MAX_DEPTH */
 {
 	ulint	ret;
 	lock_t*	lock;
@@ -3406,7 +3426,7 @@ lock_deadlock_recursive(
 
 			lock_trx = lock->trx;
 
-			if (lock_trx == start || too_far) {
+			if (lock_trx == start) {
 
 				/* We came back to the recursion starting
 				point: a deadlock detected; or we have
@@ -3453,19 +3473,10 @@ lock_deadlock_recursive(
 				}
 #ifdef UNIV_DEBUG
 				if (lock_print_waits) {
-					fputs("Deadlock detected"
-					      " or too long search\n",
+					fputs("Deadlock detected\n",
 					      stderr);
 				}
 #endif /* UNIV_DEBUG */
-				if (too_far) {
-
-					fputs("TOO DEEP OR LONG SEARCH"
-					      " IN THE LOCK TABLE"
-					      " WAITS-FOR GRAPH\n", ef);
-
-					return(LOCK_VICTIM_IS_START);
-				}
 
 				if (trx_weight_cmp(wait_lock->trx,
 						   start) >= 0) {
@@ -3501,6 +3512,21 @@ lock_deadlock_recursive(
 				return(LOCK_VICTIM_IS_OTHER);
 			}
 
+			if (too_far) {
+
+#ifdef UNIV_DEBUG
+				if (lock_print_waits) {
+					fputs("Deadlock search exceeds"
+					      " max steps or depth.\n",
+					      stderr);
+				}
+#endif /* UNIV_DEBUG */
+				/* The information about transaction/lock
+				to be rolled back is available in the top
+				level. Do not print anything here. */
+				return(LOCK_EXCEED_MAX_DEPTH);
+			}
+
 			if (lock_trx->que_state == TRX_QUE_LOCK_WAIT) {
 
 				/* Another trx ahead has requested lock	in an

From 709e5a9839756704def6d6bb93fd0bcbfa0905e4 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:42:55 +0400
Subject: [PATCH 056/116] Applying InnoDB snapshot

Detailed revision comments:

r6635 | marko | 2010-02-10 11:07:05 +0200 (Wed, 10 Feb 2010) | 4 lines
branches/zip: Clean up after r6559.  Now that
btr_pcur_open_with_no_init() is a macro, do not mix preprocessor
directives in the macro invocation, because it is implementation-defined
whether that is going to work.
---
 storage/innodb_plugin/row/row0sel.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/storage/innodb_plugin/row/row0sel.c b/storage/innodb_plugin/row/row0sel.c
index e14f29d8d64..1656657f9e9 100644
--- a/storage/innodb_plugin/row/row0sel.c
+++ b/storage/innodb_plugin/row/row0sel.c
@@ -3200,14 +3200,17 @@ row_sel_try_search_shortcut_for_mysql(
 	ut_ad(dict_index_is_clust(index));
 	ut_ad(!prebuilt->templ_contains_blob);
 
+#ifndef UNIV_SEARCH_DEBUG
 	btr_pcur_open_with_no_init(index, search_tuple, PAGE_CUR_GE,
 				   BTR_SEARCH_LEAF, pcur,
-#ifndef UNIV_SEARCH_DEBUG
 				   RW_S_LATCH,
-#else
-				   0,
-#endif
 				   mtr);
+#else /* UNIV_SEARCH_DEBUG */
+	btr_pcur_open_with_no_init(index, search_tuple, PAGE_CUR_GE,
+				   BTR_SEARCH_LEAF, pcur,
+				   0,
+				   mtr);
+#endif /* UNIV_SEARCH_DEBUG */
 	rec = btr_pcur_get_rec(pcur);
 
 	if (!page_rec_is_user_rec(rec)) {

From a2c5231e3d3843485842d2324c91335d0c395016 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:51:08 +0400
Subject: [PATCH 057/116] Applying InnoDB snapshot

Detailed revision comments:

r6639 | marko | 2010-02-10 13:11:04 +0200 (Wed, 10 Feb 2010) | 1 line
branches/zip: trx_rseg_create(): Unused function, remove.
---
 storage/innodb_plugin/include/trx0rseg.h | 11 --------
 storage/innodb_plugin/trx/trx0rseg.c     | 36 ------------------------
 2 files changed, 47 deletions(-)

diff --git a/storage/innodb_plugin/include/trx0rseg.h b/storage/innodb_plugin/include/trx0rseg.h
index ba1fc88b6c4..e8f84c97dd9 100644
--- a/storage/innodb_plugin/include/trx0rseg.h
+++ b/storage/innodb_plugin/include/trx0rseg.h
@@ -114,17 +114,6 @@ trx_rseg_list_and_array_init(
 /*=========================*/
 	trx_sysf_t*	sys_header,	/*!< in: trx system header */
 	mtr_t*		mtr);		/*!< in: mtr */
-/****************************************************************//**
-Creates a new rollback segment to the database.
-@return	the created segment object, NULL if fail */
-UNIV_INTERN
-trx_rseg_t*
-trx_rseg_create(
-/*============*/
-	ulint	space,		/*!< in: space id */
-	ulint	max_size,	/*!< in: max size in pages */
-	ulint*	id,		/*!< out: rseg id */
-	mtr_t*	mtr);		/*!< in: mtr */
 /***************************************************************************
 Free's an instance of the rollback segment in memory. */
 UNIV_INTERN
diff --git a/storage/innodb_plugin/trx/trx0rseg.c b/storage/innodb_plugin/trx/trx0rseg.c
index 8d754788e2a..3bc497bc135 100644
--- a/storage/innodb_plugin/trx/trx0rseg.c
+++ b/storage/innodb_plugin/trx/trx0rseg.c
@@ -286,39 +286,3 @@ trx_rseg_list_and_array_init(
 		}
 	}
 }
-
-/****************************************************************//**
-Creates a new rollback segment to the database.
-@return	the created segment object, NULL if fail */
-UNIV_INTERN
-trx_rseg_t*
-trx_rseg_create(
-/*============*/
-	ulint	space,		/*!< in: space id */
-	ulint	max_size,	/*!< in: max size in pages */
-	ulint*	id,		/*!< out: rseg id */
-	mtr_t*	mtr)		/*!< in: mtr */
-{
-	ulint		flags;
-	ulint		zip_size;
-	ulint		page_no;
-	trx_rseg_t*	rseg;
-
-	mtr_x_lock(fil_space_get_latch(space, &flags), mtr);
-	zip_size = dict_table_flags_to_zip_size(flags);
-	mutex_enter(&kernel_mutex);
-
-	page_no = trx_rseg_header_create(space, zip_size, max_size, id, mtr);
-
-	if (page_no == FIL_NULL) {
-
-		mutex_exit(&kernel_mutex);
-		return(NULL);
-	}
-
-	rseg = trx_rseg_mem_create(*id, space, zip_size, page_no, mtr);
-
-	mutex_exit(&kernel_mutex);
-
-	return(rseg);
-}

From 9da177c6bfcf54088f21b31c1a5f79335ae8ad0a Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:51:26 +0400
Subject: [PATCH 058/116] Applying InnoDB snapshot

Detailed revision comments:

r6660 | marko | 2010-02-11 11:21:11 +0200 (Thu, 11 Feb 2010) | 7 lines
branches/zip: Clarify the rollback of INSERT by UPDATE of delete-marked rec.

row_undo_mod_remove_clust_low(): Augment the function comment.

row_undo_mod_remove_clust_low(), row_undo_mod_del_mark_or_remove_sec_low(),
row_undo_mod_del_mark_or_remove_sec(), row_undo_mod_upd_del_sec():
Add ut_ad(node->rec_type == TRX_UNDO_UPD_DEL_REC);
---
 storage/innodb_plugin/row/row0umod.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/storage/innodb_plugin/row/row0umod.c b/storage/innodb_plugin/row/row0umod.c
index fb7fd460769..ec96c8fc845 100644
--- a/storage/innodb_plugin/row/row0umod.c
+++ b/storage/innodb_plugin/row/row0umod.c
@@ -144,6 +144,10 @@ row_undo_mod_clust_low(
 
 /***********************************************************//**
 Removes a clustered index record after undo if possible.
+This is attempted when the record was inserted by updating a
+delete-marked record and there no longer exist transactions
+that would see the delete-marked record.  In other words, we
+roll back the insert by purging the record.
 @return	DB_SUCCESS, DB_FAIL, or error code: we may run out of file space */
 static
 ulint
@@ -159,6 +163,7 @@ row_undo_mod_remove_clust_low(
 	ulint		err;
 	ibool		success;
 
+	ut_ad(node->rec_type == TRX_UNDO_UPD_DEL_REC);
 	pcur = &(node->pcur);
 	btr_cur = btr_pcur_get_btr_cur(pcur);
 
@@ -316,6 +321,7 @@ row_undo_mod_del_mark_or_remove_sec_low(
 	mtr_t		mtr;
 	mtr_t		mtr_vers;
 
+	ut_ad(node->rec_type == TRX_UNDO_UPD_DEL_REC);
 	log_free_check();
 	mtr_start(&mtr);
 
@@ -412,6 +418,8 @@ row_undo_mod_del_mark_or_remove_sec(
 {
 	ulint	err;
 
+	ut_ad(node->rec_type == TRX_UNDO_UPD_DEL_REC);
+
 	err = row_undo_mod_del_mark_or_remove_sec_low(node, thr, index,
 						      entry, BTR_MODIFY_LEAF);
 	if (err == DB_SUCCESS) {
@@ -533,6 +541,7 @@ row_undo_mod_upd_del_sec(
 	dict_index_t*	index;
 	ulint		err	= DB_SUCCESS;
 
+	ut_ad(node->rec_type == TRX_UNDO_UPD_DEL_REC);
 	heap = mem_heap_create(1024);
 
 	while (node->index != NULL) {

From bc18342a0aaa3277e51521488e158020080c627b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:51:48 +0400
Subject: [PATCH 059/116] Applying InnoDB snapshot

Detailed revision comments:

r6672 | marko | 2010-02-11 13:01:18 +0200 (Thu, 11 Feb 2010) | 1 line
branches/zip: Introduce thr_is_recv().
---
 storage/innodb_plugin/include/que0que.h  | 11 +++++++++++
 storage/innodb_plugin/include/que0que.ic | 14 ++++++++++++++
 storage/innodb_plugin/row/row0umod.c     |  2 +-
 3 files changed, 26 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/include/que0que.h b/storage/innodb_plugin/include/que0que.h
index 420f34550e2..c35b75d7d51 100644
--- a/storage/innodb_plugin/include/que0que.h
+++ b/storage/innodb_plugin/include/que0que.h
@@ -30,6 +30,7 @@ Created 5/27/1996 Heikki Tuuri
 #include "data0data.h"
 #include "dict0types.h"
 #include "trx0trx.h"
+#include "trx0roll.h"
 #include "srv0srv.h"
 #include "usr0types.h"
 #include "que0types.h"
@@ -215,6 +216,16 @@ trx_t*
 thr_get_trx(
 /*========*/
 	que_thr_t*	thr);	/*!< in: query thread */
+/*******************************************************************//**
+Determines if this thread is rolling back an incomplete transaction
+in crash recovery.
+@return TRUE if thr is rolling back an incomplete transaction in crash
+recovery */
+UNIV_INLINE
+ibool
+thr_is_recv(
+/*========*/
+	const que_thr_t*	thr);	/*!< in: query thread */
 /***********************************************************************//**
 Gets the type of a graph node. */
 UNIV_INLINE
diff --git a/storage/innodb_plugin/include/que0que.ic b/storage/innodb_plugin/include/que0que.ic
index a1c0dc1e77a..f027c748230 100644
--- a/storage/innodb_plugin/include/que0que.ic
+++ b/storage/innodb_plugin/include/que0que.ic
@@ -38,6 +38,20 @@ thr_get_trx(
 	return(thr->graph->trx);
 }
 
+/*******************************************************************//**
+Determines if this thread is rolling back an incomplete transaction
+in crash recovery.
+@return TRUE if thr is rolling back an incomplete transaction in crash
+recovery */
+UNIV_INLINE
+ibool
+thr_is_recv(
+/*========*/
+	const que_thr_t*	thr)	/*!< in: query thread */
+{
+	return(trx_is_recv(thr->graph->trx));
+}
+
 /***********************************************************************//**
 Gets the first thr in a fork. */
 UNIV_INLINE
diff --git a/storage/innodb_plugin/row/row0umod.c b/storage/innodb_plugin/row/row0umod.c
index ec96c8fc845..674f0fbe09a 100644
--- a/storage/innodb_plugin/row/row0umod.c
+++ b/storage/innodb_plugin/row/row0umod.c
@@ -559,7 +559,7 @@ row_undo_mod_upd_del_sec(
 			does not exist.  However, this situation may
 			only occur during the rollback of incomplete
 			transactions. */
-			ut_a(trx_is_recv(thr_get_trx(thr)));
+			ut_a(thr_is_recv(thr));
 		} else {
 			err = row_undo_mod_del_mark_or_remove_sec(
 				node, thr, index, entry);

From 503708894f341fa2865a0b944326a92b3a29abd4 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:52:08 +0400
Subject: [PATCH 060/116] Applying InnoDB snapshot

Detailed revision comments:

r6673 | marko | 2010-02-11 13:09:48 +0200 (Thu, 11 Feb 2010) | 9 lines
branches/zip: Relax a debug assertion about a missing BLOB. (Issue #452)
When rolling back an incomplete transaction in purge, tolerate missing
BLOBs also in update undo, when undoing an INSERT by updating a delete-marked
record, and the delete-marked record is no longer needed.
Previously, we only tolerated missing BLOBs in insert undo.
This merely fixes a debug assertion; the code performed correctly
without UNIV_DEBUG.

rb://249 approved by Sunny Bains.
---
 storage/innodb_plugin/btr/btr0cur.c       |  4 ++--
 storage/innodb_plugin/include/trx0types.h |  7 +++++++
 storage/innodb_plugin/row/row0umod.c      | 21 ++++++++++++---------
 3 files changed, 21 insertions(+), 11 deletions(-)

diff --git a/storage/innodb_plugin/btr/btr0cur.c b/storage/innodb_plugin/btr/btr0cur.c
index df21b8d4440..88827d33c61 100644
--- a/storage/innodb_plugin/btr/btr0cur.c
+++ b/storage/innodb_plugin/btr/btr0cur.c
@@ -4260,7 +4260,7 @@ btr_free_externally_stored_field(
 		/* In the rollback of uncommitted transactions, we may
 		encounter a clustered index record whose BLOBs have
 		not been written.  There is nothing to free then. */
-		ut_a(rb_ctx == RB_RECOVERY);
+		ut_a(rb_ctx == RB_RECOVERY || rb_ctx == RB_RECOVERY_PURGE_REC);
 		return;
 	}
 
@@ -4306,7 +4306,7 @@ btr_free_externally_stored_field(
 		    || (mach_read_from_1(field_ref + BTR_EXTERN_LEN)
 			& BTR_EXTERN_OWNER_FLAG)
 		    /* Rollback and inherited field */
-		    || (rb_ctx != RB_NONE
+		    || ((rb_ctx == RB_NORMAL || rb_ctx == RB_RECOVERY)
 			&& (mach_read_from_1(field_ref + BTR_EXTERN_LEN)
 			    & BTR_EXTERN_INHERITED_FLAG))) {
 
diff --git a/storage/innodb_plugin/include/trx0types.h b/storage/innodb_plugin/include/trx0types.h
index 24cf57d53d5..ae4451636cd 100644
--- a/storage/innodb_plugin/include/trx0types.h
+++ b/storage/innodb_plugin/include/trx0types.h
@@ -70,6 +70,13 @@ typedef struct trx_named_savept_struct trx_named_savept_t;
 enum trx_rb_ctx {
 	RB_NONE = 0,	/*!< no rollback */
 	RB_NORMAL,	/*!< normal rollback */
+	RB_RECOVERY_PURGE_REC,
+			/*!< rolling back an incomplete transaction,
+			in crash recovery, rolling back an
+			INSERT that was performed by updating a
+			delete-marked record; if the delete-marked record
+			no longer exists in an active read view, it will
+			be purged */
 	RB_RECOVERY	/*!< rolling back an incomplete transaction,
 			in crash recovery */
 };
diff --git a/storage/innodb_plugin/row/row0umod.c b/storage/innodb_plugin/row/row0umod.c
index 674f0fbe09a..3adcee0a7ff 100644
--- a/storage/innodb_plugin/row/row0umod.c
+++ b/storage/innodb_plugin/row/row0umod.c
@@ -154,7 +154,7 @@ ulint
 row_undo_mod_remove_clust_low(
 /*==========================*/
 	undo_node_t*	node,	/*!< in: row undo node */
-	que_thr_t*	thr __attribute__((unused)), /*!< in: query thread */
+	que_thr_t*	thr,	/*!< in: query thread */
 	mtr_t*		mtr,	/*!< in: mtr */
 	ulint		mode)	/*!< in: BTR_MODIFY_LEAF or BTR_MODIFY_TREE */
 {
@@ -195,11 +195,13 @@ row_undo_mod_remove_clust_low(
 	} else {
 		ut_ad(mode == BTR_MODIFY_TREE);
 
-		/* Note that since this operation is analogous to purge,
-		we can free also inherited externally stored fields:
-		hence the RB_NONE in the call below */
+		/* This operation is analogous to purge, we can free also
+		inherited externally stored fields */
 
-		btr_cur_pessimistic_delete(&err, FALSE, btr_cur, RB_NONE, mtr);
+		btr_cur_pessimistic_delete(&err, FALSE, btr_cur,
+					   thr_is_recv(thr)
+					   ? RB_RECOVERY_PURGE_REC
+					   : RB_NONE, mtr);
 
 		/* The delete operation may fail if we have little
 		file space left: TODO: easiest to crash the database
@@ -376,10 +378,11 @@ row_undo_mod_del_mark_or_remove_sec_low(
 		} else {
 			ut_ad(mode == BTR_MODIFY_TREE);
 
-			/* No need to distinguish RB_RECOVERY here, because we
-			are deleting a secondary index record: the distinction
-			between RB_NORMAL and RB_RECOVERY only matters when
-			deleting a record that contains externally stored
+			/* No need to distinguish RB_RECOVERY_PURGE here,
+			because we are deleting a secondary index record:
+			the distinction between RB_NORMAL and
+			RB_RECOVERY_PURGE only matters when deleting a
+			record that contains externally stored
 			columns. */
 			ut_ad(!dict_index_is_clust(index));
 			btr_cur_pessimistic_delete(&err, FALSE, btr_cur,

From 67925345f8e19ebb73a0b9abdd5824d9f41aa05e Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:52:35 +0400
Subject: [PATCH 061/116] Applying InnoDB snapshot, fixes BUG#49535.

Detailed revision comments:

r6674 | inaam | 2010-02-11 17:54:44 +0200 (Thu, 11 Feb 2010) | 16 lines
branches/zip bug# 49535

This is a backport of r4924.
mem_heap_get_size() scans all allocated blocks to calculate the total
size of the heap. This patch introduces a new, total_size, field in
mem_block_info_struct. This field is valid only for base block
(i.e.: the first block allocated for the heap) and is set to
ULINT_UNDEFINED in other blocks.
This considerably improves the performance of redo scan during recovery.

rb://108 issue#216

Approved by: Heikki
---
 storage/innodb_plugin/include/mem0mem.h  |  3 +++
 storage/innodb_plugin/include/mem0mem.ic |  8 +-------
 storage/innodb_plugin/mem/mem0mem.c      | 18 ++++++++++++++++++
 3 files changed, 22 insertions(+), 7 deletions(-)

diff --git a/storage/innodb_plugin/include/mem0mem.h b/storage/innodb_plugin/include/mem0mem.h
index 98f8748e529..21d300a8ce5 100644
--- a/storage/innodb_plugin/include/mem0mem.h
+++ b/storage/innodb_plugin/include/mem0mem.h
@@ -359,6 +359,9 @@ struct mem_block_info_struct {
 			to the heap is also the first block in this list,
 			though it also contains the base node of the list. */
 	ulint	len;	/*!< physical length of this block in bytes */
+	ulint	total_size; /* physical length in bytes of all blocks
+			in the heap. This is defined only in the base
+			node and is set to ULINT_UNDEFINED in others. */
 	ulint	type;	/*!< type of heap: MEM_HEAP_DYNAMIC, or
 			MEM_HEAP_BUF possibly ORed to MEM_HEAP_BTR_SEARCH */
 	ulint	free;	/*!< offset in bytes of the first free position for
diff --git a/storage/innodb_plugin/include/mem0mem.ic b/storage/innodb_plugin/include/mem0mem.ic
index e7080d8c508..cb681c3f724 100644
--- a/storage/innodb_plugin/include/mem0mem.ic
+++ b/storage/innodb_plugin/include/mem0mem.ic
@@ -579,18 +579,12 @@ mem_heap_get_size(
 /*==============*/
 	mem_heap_t*	heap)	/*!< in: heap */
 {
-	mem_block_t*	block;
 	ulint		size	= 0;
 
 	ut_ad(mem_heap_check(heap));
 
-	block = heap;
+	size = heap->total_size;
 
-	while (block != NULL) {
-
-		size += mem_block_get_len(block);
-		block = UT_LIST_GET_NEXT(list, block);
-	}
 #ifndef UNIV_HOTBACKUP
 	if (heap->free_block) {
 		size += UNIV_PAGE_SIZE;
diff --git a/storage/innodb_plugin/mem/mem0mem.c b/storage/innodb_plugin/mem/mem0mem.c
index ccb2fd8a7b4..39bbfc90313 100644
--- a/storage/innodb_plugin/mem/mem0mem.c
+++ b/storage/innodb_plugin/mem/mem0mem.c
@@ -383,6 +383,20 @@ mem_heap_create_block(
 	mem_block_set_free(block, MEM_BLOCK_HEADER_SIZE);
 	mem_block_set_start(block, MEM_BLOCK_HEADER_SIZE);
 
+	if (UNIV_UNLIKELY(heap == NULL)) {
+		/* This is the first block of the heap. The field
+		total_size should be initialized here */
+		block->total_size = len;
+	} else {
+		/* Not the first allocation for the heap. This block's
+		total_length field should be set to undefined. */
+		ut_d(block->total_size = ULINT_UNDEFINED);
+		UNIV_MEM_INVALID(&block->total_size,
+				 sizeof block->total_size);
+
+		heap->total_size += len;
+	}
+
 	ut_ad((ulint)MEM_BLOCK_HEADER_SIZE < len);
 
 	return(block);
@@ -471,6 +485,10 @@ mem_heap_block_free(
 
 	mem_pool_mutex_exit();
 #endif
+
+	ut_ad(heap->total_size >= block->len);
+	heap->total_size -= block->len;
+
 	type = heap->type;
 	len = block->len;
 	block->magic_n = MEM_FREED_BLOCK_MAGIC_N;

From 965b7219584e52c9b145a5b462463af064b8085e Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:53:41 +0400
Subject: [PATCH 062/116] Applying InnoDB snapshot

Detailed revision comments:

r6675 | marko | 2010-02-11 22:41:11 +0200 (Thu, 11 Feb 2010) | 1 line
branches/zip: Remove bogus debug assertions introduced in r6660.
---
 storage/innodb_plugin/row/row0umod.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/storage/innodb_plugin/row/row0umod.c b/storage/innodb_plugin/row/row0umod.c
index 3adcee0a7ff..12cd05766dd 100644
--- a/storage/innodb_plugin/row/row0umod.c
+++ b/storage/innodb_plugin/row/row0umod.c
@@ -323,7 +323,6 @@ row_undo_mod_del_mark_or_remove_sec_low(
 	mtr_t		mtr;
 	mtr_t		mtr_vers;
 
-	ut_ad(node->rec_type == TRX_UNDO_UPD_DEL_REC);
 	log_free_check();
 	mtr_start(&mtr);
 
@@ -421,8 +420,6 @@ row_undo_mod_del_mark_or_remove_sec(
 {
 	ulint	err;
 
-	ut_ad(node->rec_type == TRX_UNDO_UPD_DEL_REC);
-
 	err = row_undo_mod_del_mark_or_remove_sec_low(node, thr, index,
 						      entry, BTR_MODIFY_LEAF);
 	if (err == DB_SUCCESS) {

From 9c27e3606a5a3d7db54995705000ee1f4bd69c3c Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:54:01 +0400
Subject: [PATCH 063/116] Applying InnoDB snapshot

Detailed revision comments:

r6707 | inaam | 2010-02-12 19:22:35 +0200 (Fri, 12 Feb 2010) | 4 lines
branches/zip

ChangeLog entry for r6674.
---
 storage/innodb_plugin/ChangeLog | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index c66c4b01cda..3d26ebb82bb 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,9 @@
+2010-02-11	The InnoDB Team
+
+	* include/mem0mem.h, include/mem0mem.ic, mem/mem0mem.c:
+	Fix bug #49535 Available memory check slows down crash
+	recovery tens of times
+
 2010-02-09	The InnoDB Team
 
 	* buf/buf0buf.c:

From 0c2e8ba37a97c69bbaef56c86a594d0fc759238c Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:54:23 +0400
Subject: [PATCH 064/116] Applying InnoDB snapshot

Detailed revision comments:

r6712 | marko | 2010-02-16 10:05:36 +0200 (Tue, 16 Feb 2010) | 2 lines
branches/zip: trx_lists_init_at_db_start(): Assert that the kernel_mutex
is held by the caller.
---
 storage/innodb_plugin/trx/trx0trx.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/storage/innodb_plugin/trx/trx0trx.c b/storage/innodb_plugin/trx/trx0trx.c
index 0951b98b79f..7fe0bfb4da8 100644
--- a/storage/innodb_plugin/trx/trx0trx.c
+++ b/storage/innodb_plugin/trx/trx0trx.c
@@ -425,6 +425,7 @@ trx_lists_init_at_db_start(void)
 	trx_undo_t*	undo;
 	trx_t*		trx;
 
+	ut_ad(mutex_own(&kernel_mutex));
 	UT_LIST_INIT(trx_sys->trx_list);
 
 	/* Look from the rollback segments if there exist undo logs for

From 1a4de3768a8cb80cbf42826c849ce38b21365a0c Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:54:41 +0400
Subject: [PATCH 065/116] Applying InnoDB snapshot

Detailed revision comments:

r6713 | sunny | 2010-02-16 10:12:17 +0200 (Tue, 16 Feb 2010) | 7 lines
branches/zip: Change the bit fields back to ulint. Bit fields were causing
problems with concurrency on SMP systems because of word packing issues.
The number of trx_t's in a system is not sufficient enough to require that
we try and save a few bytes in the data structure.

See rb://255 for details.
---
 storage/innodb_plugin/include/trx0trx.h | 77 ++++++++++++++-----------
 1 file changed, 43 insertions(+), 34 deletions(-)

diff --git a/storage/innodb_plugin/include/trx0trx.h b/storage/innodb_plugin/include/trx0trx.h
index e2ad1f2f722..6e611e08a96 100644
--- a/storage/innodb_plugin/include/trx0trx.h
+++ b/storage/innodb_plugin/include/trx0trx.h
@@ -349,7 +349,7 @@ trx_print(
 				   use the default max length */
 
 /** Type of data dictionary operation */
-enum trx_dict_op {
+typedef enum trx_dict_op {
 	/** The transaction is not modifying the data dictionary. */
 	TRX_DICT_OP_NONE = 0,
 	/** The transaction is creating a table or an index, or
@@ -361,7 +361,7 @@ enum trx_dict_op {
 	existing table.  In crash recovery, the data dictionary
 	must be locked, but the table must not be dropped. */
 	TRX_DICT_OP_INDEX = 2
-};
+} trx_dict_op_t;
 
 /**********************************************************************//**
 Determine if a transaction is a dictionary operation.
@@ -463,72 +463,81 @@ rolling back after a database recovery */
 
 struct trx_struct{
 	ulint		magic_n;
-	/* All the next fields are protected by the kernel mutex, except the
-	undo logs which are protected by undo_mutex */
-	const char*	op_info;	/*!< English text describing the
-					current operation, or an empty
-					string */
-	unsigned	is_purge:1;	/*!< 0=user transaction, 1=purge */
-	unsigned	is_recovered:1;	/*!< 0=normal transaction,
-					1=recovered, must be rolled back */
-	unsigned	conc_state:2;	/*!< state of the trx from the point
-					of view of concurrency control:
-					TRX_ACTIVE, TRX_COMMITTED_IN_MEMORY,
-					... */
-	unsigned	que_state:2;	/*!< valid when conc_state == TRX_ACTIVE:
-					TRX_QUE_RUNNING, TRX_QUE_LOCK_WAIT,
-					... */
-	unsigned	isolation_level:2;/* TRX_ISO_REPEATABLE_READ, ... */
-	unsigned	check_foreigns:1;/* normally TRUE, but if the user
+
+	/* These fields are not protected by any mute. */
+	ulint		isolation_level;/* TRX_ISO_REPEATABLE_READ, ... */
+	ulint		check_foreigns;/* normally TRUE, but if the user
 					wants to suppress foreign key checks,
 					(in table imports, for example) we
 					set this FALSE */
-	unsigned	check_unique_secondary:1;
+	ulint		check_unique_secondary;
 					/* normally TRUE, but if the user
 					wants to speed up inserts by
 					suppressing unique key checks
 					for secondary indexes when we decide
 					if we can use the insert buffer for
 					them, we set this FALSE */
-	unsigned	support_xa:1;	/*!< normally we do the XA two-phase
+	ulint		support_xa;	/*!< normally we do the XA two-phase
 					commit steps, but by setting this to
 					FALSE, one can save CPU time and about
 					150 bytes in the undo log size as then
 					we skip XA steps */
-	unsigned	flush_log_later:1;/* In 2PC, we hold the
+	ulint		flush_log_later;/* In 2PC, we hold the
 					prepare_commit mutex across
 					both phases. In that case, we
 					defer flush of the logs to disk
 					until after we release the
 					mutex. */
-	unsigned	must_flush_log_later:1;/* this flag is set to TRUE in
+	ulint		must_flush_log_later;/* this flag is set to TRUE in
 					trx_commit_off_kernel() if
 					flush_log_later was TRUE, and there
 					were modifications by the transaction;
 					in that case we must flush the log
 					in trx_commit_complete_for_mysql() */
-	unsigned	dict_operation:2;/**< @see enum trx_dict_op */
-	unsigned	duplicates:2;	/*!< TRX_DUP_IGNORE | TRX_DUP_REPLACE */
-	unsigned	active_trans:2;	/*!< 1 - if a transaction in MySQL
+	ulint		duplicates;	/*!< TRX_DUP_IGNORE | TRX_DUP_REPLACE */
+	ulint		active_trans;	/*!< 1 - if a transaction in MySQL
 					is active. 2 - if prepare_commit_mutex
 					was taken */
-	unsigned	has_search_latch:1;
+	ulint		has_search_latch;
 					/* TRUE if this trx has latched the
 					search system latch in S-mode */
-	unsigned	declared_to_be_inside_innodb:1;
+	ulint		deadlock_mark;	/*!< a mark field used in deadlock
+					checking algorithm.  */
+
+	/* Fields protected by the srv_conc_mutex. */
+	ulint		declared_to_be_inside_innodb;
 					/* this is TRUE if we have declared
 					this transaction in
 					srv_conc_enter_innodb to be inside the
 					InnoDB engine */
-	unsigned	handling_signals:1;/* this is TRUE as long as the trx
-					is handling signals */
-	unsigned	dict_operation_lock_mode:2;
+
+	/* Fields set when we are holding the kernel mutex, undo log mutex
+	and when not holding the mutex. */
+	trx_dict_op_t	dict_operation;	/**< @see enum trx_dict_op */
+
+	/* Fields covered by the dictionary mutex. */
+	ulint		dict_operation_lock_mode;
 					/* 0, RW_S_LATCH, or RW_X_LATCH:
 					the latch mode trx currently holds
 					on dict_operation_lock */
-	unsigned	deadlock_mark:1;/*!< a mark field used in deadlock
-					checking algorithm. Always protected
-					by the kernel_mutex. */
+
+	/* All the next fields are protected by the kernel mutex, except the
+	undo logs which are protected by undo_mutex */
+	const char*	op_info;	/*!< English text describing the
+					current operation, or an empty
+					string */
+	ulint		is_purge;	/*!< 0=user transaction, 1=purge */
+	ulint		is_recovered;	/*!< 0=normal transaction,
+					1=recovered, must be rolled back */
+	ulint		conc_state;	/*!< state of the trx from the point
+					of view of concurrency control:
+					TRX_ACTIVE, TRX_COMMITTED_IN_MEMORY,
+					... */
+	ulint		que_state;	/*!< valid when conc_state
+					== TRX_ACTIVE: TRX_QUE_RUNNING,
+				       	TRX_QUE_LOCK_WAIT, ... */
+	ulint		handling_signals;/* this is TRUE as long as the trx
+					is handling signals */
 	time_t		start_time;	/*!< time the trx object was created
 					or the state last time became
 					TRX_ACTIVE */

From d0422e5a600dc29d4d8aaaf7f2c00039189f8d7c Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:55:01 +0400
Subject: [PATCH 066/116] Applying InnoDB snapshot

Detailed revision comments:

r6714 | sunny | 2010-02-16 10:12:25 +0200 (Tue, 16 Feb 2010) | 2 lines
branches/zip: Update the comments and fix the whitespace issues.
See rb://255 Approved by: Marko
---
 storage/innodb_plugin/include/trx0trx.h | 28 ++++++++++++-------------
 1 file changed, 13 insertions(+), 15 deletions(-)

diff --git a/storage/innodb_plugin/include/trx0trx.h b/storage/innodb_plugin/include/trx0trx.h
index 6e611e08a96..9511f6feb8c 100644
--- a/storage/innodb_plugin/include/trx0trx.h
+++ b/storage/innodb_plugin/include/trx0trx.h
@@ -464,9 +464,16 @@ rolling back after a database recovery */
 struct trx_struct{
 	ulint		magic_n;
 
-	/* These fields are not protected by any mute. */
+	/* These fields are not protected by any mutex. */
+	const char*	op_info;	/*!< English text describing the
+					current operation, or an empty
+					string */
+	ulint		conc_state;	/*!< state of the trx from the point
+					of view of concurrency control:
+					TRX_ACTIVE, TRX_COMMITTED_IN_MEMORY,
+					... */
 	ulint		isolation_level;/* TRX_ISO_REPEATABLE_READ, ... */
-	ulint		check_foreigns;/* normally TRUE, but if the user
+	ulint		check_foreigns;	/* normally TRUE, but if the user
 					wants to suppress foreign key checks,
 					(in table imports, for example) we
 					set this FALSE */
@@ -503,6 +510,7 @@ struct trx_struct{
 					search system latch in S-mode */
 	ulint		deadlock_mark;	/*!< a mark field used in deadlock
 					checking algorithm.  */
+	trx_dict_op_t	dict_operation;	/**< @see enum trx_dict_op */
 
 	/* Fields protected by the srv_conc_mutex. */
 	ulint		declared_to_be_inside_innodb;
@@ -511,11 +519,8 @@ struct trx_struct{
 					srv_conc_enter_innodb to be inside the
 					InnoDB engine */
 
-	/* Fields set when we are holding the kernel mutex, undo log mutex
-	and when not holding the mutex. */
-	trx_dict_op_t	dict_operation;	/**< @see enum trx_dict_op */
-
-	/* Fields covered by the dictionary mutex. */
+	/* Fields protected by dict_operation_loco(). The very latch
+	it is used to track. */
 	ulint		dict_operation_lock_mode;
 					/* 0, RW_S_LATCH, or RW_X_LATCH:
 					the latch mode trx currently holds
@@ -523,19 +528,12 @@ struct trx_struct{
 
 	/* All the next fields are protected by the kernel mutex, except the
 	undo logs which are protected by undo_mutex */
-	const char*	op_info;	/*!< English text describing the
-					current operation, or an empty
-					string */
 	ulint		is_purge;	/*!< 0=user transaction, 1=purge */
 	ulint		is_recovered;	/*!< 0=normal transaction,
 					1=recovered, must be rolled back */
-	ulint		conc_state;	/*!< state of the trx from the point
-					of view of concurrency control:
-					TRX_ACTIVE, TRX_COMMITTED_IN_MEMORY,
-					... */
 	ulint		que_state;	/*!< valid when conc_state
 					== TRX_ACTIVE: TRX_QUE_RUNNING,
-				       	TRX_QUE_LOCK_WAIT, ... */
+					TRX_QUE_LOCK_WAIT, ... */
 	ulint		handling_signals;/* this is TRUE as long as the trx
 					is handling signals */
 	time_t		start_time;	/*!< time the trx object was created

From 378f7121f95f598c73bda2609eec0b3c910995e1 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:55:20 +0400
Subject: [PATCH 067/116] Applying InnoDB snapshot

Detailed revision comments:

r6715 | sunny | 2010-02-16 10:14:21 +0200 (Tue, 16 Feb 2010) | 1 line
branches/zip: Fix comment. Non functional change.
---
 storage/innodb_plugin/include/trx0trx.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/include/trx0trx.h b/storage/innodb_plugin/include/trx0trx.h
index 9511f6feb8c..664b868e762 100644
--- a/storage/innodb_plugin/include/trx0trx.h
+++ b/storage/innodb_plugin/include/trx0trx.h
@@ -519,7 +519,7 @@ struct trx_struct{
 					srv_conc_enter_innodb to be inside the
 					InnoDB engine */
 
-	/* Fields protected by dict_operation_loco(). The very latch
+	/* Fields protected by dict_operation_lock(). The very latch
 	it is used to track. */
 	ulint		dict_operation_lock_mode;
 					/* 0, RW_S_LATCH, or RW_X_LATCH:

From 34799eb5086b728395220dd03b869101a826d52f Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:55:42 +0400
Subject: [PATCH 068/116] Applying InnoDB snapshot

Detailed revision comments:

r6717 | marko | 2010-02-16 14:53:20 +0200 (Tue, 16 Feb 2010) | 2 lines
branches/zip: log_reserve_and_write_fast(): Correct a race condition
in UNIV_LOG_LSN_DEBUG.  This could have caused Issue #440.
---
 storage/innodb_plugin/include/log0log.ic | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/storage/innodb_plugin/include/log0log.ic b/storage/innodb_plugin/include/log0log.ic
index 36d151a3064..098ef47d761 100644
--- a/storage/innodb_plugin/include/log0log.ic
+++ b/storage/innodb_plugin/include/log0log.ic
@@ -314,12 +314,15 @@ log_reserve_and_write_fast(
 	ulint		data_len;
 #ifdef UNIV_LOG_LSN_DEBUG
 	/* length of the LSN pseudo-record */
-	ulint		lsn_len = 1
-		+ mach_get_compressed_size(log_sys->lsn >> 32)
-		+ mach_get_compressed_size(log_sys->lsn & 0xFFFFFFFFUL);
+	ulint		lsn_len;
 #endif /* UNIV_LOG_LSN_DEBUG */
 
 	mutex_enter(&log_sys->mutex);
+#ifdef UNIV_LOG_LSN_DEBUG
+	lsn_len = 1
+		+ mach_get_compressed_size(log_sys->lsn >> 32)
+		+ mach_get_compressed_size(log_sys->lsn & 0xFFFFFFFFUL);
+#endif /* UNIV_LOG_LSN_DEBUG */
 
 	data_len = len
 #ifdef UNIV_LOG_LSN_DEBUG

From e3239bcf518adc8295684aec160d470fae0f0a50 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:56:02 +0400
Subject: [PATCH 069/116] Applying InnoDB snapshot

Detailed revision comments:

r6718 | marko | 2010-02-16 15:06:16 +0200 (Tue, 16 Feb 2010) | 1 line
branches/zip: Fix a comment.
---
 storage/innodb_plugin/include/trx0trx.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/storage/innodb_plugin/include/trx0trx.h b/storage/innodb_plugin/include/trx0trx.h
index 664b868e762..e570f336462 100644
--- a/storage/innodb_plugin/include/trx0trx.h
+++ b/storage/innodb_plugin/include/trx0trx.h
@@ -519,10 +519,10 @@ struct trx_struct{
 					srv_conc_enter_innodb to be inside the
 					InnoDB engine */
 
-	/* Fields protected by dict_operation_lock(). The very latch
+	/* Fields protected by dict_operation_lock. The very latch
 	it is used to track. */
 	ulint		dict_operation_lock_mode;
-					/* 0, RW_S_LATCH, or RW_X_LATCH:
+					/*!< 0, RW_S_LATCH, or RW_X_LATCH:
 					the latch mode trx currently holds
 					on dict_operation_lock */
 

From 2d889edaac758f7dcb22154e97e4496ba22149df Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:56:38 +0400
Subject: [PATCH 070/116] Applying InnoDB snapshot

Detailed revision comments:

r6723 | marko | 2010-02-17 11:48:34 +0200 (Wed, 17 Feb 2010) | 3 lines
branches/zip: lock_table_other_has_incompatible():
Return an incompatible lock or NULL instead of TRUE or FALSE.
Approved by Sunny over IM.
---
 storage/innodb_plugin/lock/lock0lock.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/storage/innodb_plugin/lock/lock0lock.c b/storage/innodb_plugin/lock/lock0lock.c
index 1de612ef637..325ebc846bb 100644
--- a/storage/innodb_plugin/lock/lock0lock.c
+++ b/storage/innodb_plugin/lock/lock0lock.c
@@ -3753,9 +3753,10 @@ lock_table_enqueue_waiting(
 
 /*********************************************************************//**
 Checks if other transactions have an incompatible mode lock request in
-the lock queue. */
+the lock queue.
+@return	lock or NULL */
 UNIV_INLINE
-ibool
+lock_t*
 lock_table_other_has_incompatible(
 /*==============================*/
 	trx_t*		trx,	/*!< in: transaction, or NULL if all
@@ -3777,13 +3778,13 @@ lock_table_other_has_incompatible(
 		    && (!lock_mode_compatible(lock_get_mode(lock), mode))
 		    && (wait || !(lock_get_wait(lock)))) {
 
-			return(TRUE);
+			return(lock);
 		}
 
 		lock = UT_LIST_GET_PREV(un_member.tab_lock.locks, lock);
 	}
 
-	return(FALSE);
+	return(NULL);
 }
 
 /*********************************************************************//**

From 0ee8a7d6e58b5ef84666c53d157cc4e8ae96983d Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:57:04 +0400
Subject: [PATCH 071/116] Applying InnoDB snapshot, fixes BUG#50691.

Detailed revision comments:

r6724 | marko | 2010-02-17 15:52:05 +0200 (Wed, 17 Feb 2010) | 11 lines
branches/zip: Merge revisions 6613:6669 from branches/5.1:
  ------------------------------------------------------------------------
  r6669 | jyang | 2010-02-11 12:24:19 +0200 (Thu, 11 Feb 2010) | 7 lines

  branches/5.1: Fix bug #50691, AIX implementation of readdir_r
  causes InnoDB errors. readdir_r() returns an non-NULL value
  in the case of reaching the end of a directory. It should
  not be treated as an error return.

  rb://238 approved by Marko
  ------------------------------------------------------------------------
---
 storage/innodb_plugin/os/os0file.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/os/os0file.c b/storage/innodb_plugin/os/os0file.c
index 2bb652980b0..f19e587053b 100644
--- a/storage/innodb_plugin/os/os0file.c
+++ b/storage/innodb_plugin/os/os0file.c
@@ -789,7 +789,15 @@ next_file:
 #ifdef HAVE_READDIR_R
 	ret = readdir_r(dir, (struct dirent*)dirent_buf, &ent);
 
-	if (ret != 0) {
+	if (ret != 0
+#ifdef UNIV_AIX
+	    /* On AIX, only if we got non-NULL 'ent' (result) value and
+	    a non-zero 'ret' (return) value, it indicates a failed
+	    readdir_r() call. An NULL 'ent' with an non-zero 'ret'
+	    would indicate the "end of the directory" is reached. */
+	    && ent != NULL
+#endif
+	   ) {
 		fprintf(stderr,
 			"InnoDB: cannot read directory %s, error %lu\n",
 			dirname, (ulong)ret);

From 622c8c8bf46f2c3f09fea5a1d3af902bacb4cdd9 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:57:40 +0400
Subject: [PATCH 072/116] Applying InnoDB snapshot

Detailed revision comments:

r6726 | marko | 2010-02-17 18:49:21 +0200 (Wed, 17 Feb 2010) | 3 lines
branches/zip: FIL_PAGE_FILE_FLUSH_LSN: Note that the field is only valid
for the first page of each ibdata* file, not *.ibd files.
Suggested by Heikki, in connection with the LSN warning noted in Issue #341.
---
 storage/innodb_plugin/include/fil0fil.h | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/storage/innodb_plugin/include/fil0fil.h b/storage/innodb_plugin/include/fil0fil.h
index 74d0fbcdacd..053eb613ad1 100644
--- a/storage/innodb_plugin/include/fil0fil.h
+++ b/storage/innodb_plugin/include/fil0fil.h
@@ -110,9 +110,10 @@ extern fil_addr_t	fil_addr_null;
 					contents of this field is valid
 					for all uncompressed pages. */
 #define FIL_PAGE_FILE_FLUSH_LSN	26	/*!< this is only defined for the
-					first page in a data file: the file
-					has been flushed to disk at least up
-					to this lsn */
+					first page in a system tablespace
+					data file (ibdata*, not *.ibd):
+					the file has been flushed to disk
+					at least up to this lsn */
 #define FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID  34 /*!< starting from 4.1.x this
 					contains the space id of the page */
 #define FIL_PAGE_DATA		38	/*!< start of the data on the page */

From 49a27847af85fbd7ec52fe953e269ae234a3700f Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:58:03 +0400
Subject: [PATCH 073/116] Applying InnoDB snapshot

Detailed revision comments:

r6727 | marko | 2010-02-17 18:50:20 +0200 (Wed, 17 Feb 2010) | 2 lines
branches/zip: fsp_init_file_page_low(): Declare the page uninitialized
for Valgrind.
---
 storage/innodb_plugin/fsp/fsp0fsp.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/storage/innodb_plugin/fsp/fsp0fsp.c b/storage/innodb_plugin/fsp/fsp0fsp.c
index 9b53e5585be..88f8da6981e 100644
--- a/storage/innodb_plugin/fsp/fsp0fsp.c
+++ b/storage/innodb_plugin/fsp/fsp0fsp.c
@@ -872,6 +872,7 @@ fsp_init_file_page_low(
 #ifdef UNIV_BASIC_LOG_DEBUG
 	memset(page, 0xff, UNIV_PAGE_SIZE);
 #endif
+	UNIV_MEM_INVALID(page, UNIV_PAGE_SIZE);
 	mach_write_to_4(page + FIL_PAGE_OFFSET, buf_block_get_page_no(block));
 	memset(page + FIL_PAGE_LSN, 0, 8);
 	mach_write_to_4(page + FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID,

From 3b6cb06fbe719dba56fcebaf4870c197cafe768b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:58:26 +0400
Subject: [PATCH 074/116] Applying InnoDB snapshot

Detailed revision comments:

r6728 | marko | 2010-02-17 18:54:04 +0200 (Wed, 17 Feb 2010) | 3 lines
branches/zip: Remove UNIV_BASIC_LOG_DEBUG.
This fixes the FILE_FLUSH_LSN printouts mentioned in Issue #341.
Suggested by Heikki.
---
 storage/innodb_plugin/fsp/fsp0fsp.c  | 3 ---
 storage/innodb_plugin/include/univ.i | 5 -----
 2 files changed, 8 deletions(-)

diff --git a/storage/innodb_plugin/fsp/fsp0fsp.c b/storage/innodb_plugin/fsp/fsp0fsp.c
index 88f8da6981e..b5329bb9ebe 100644
--- a/storage/innodb_plugin/fsp/fsp0fsp.c
+++ b/storage/innodb_plugin/fsp/fsp0fsp.c
@@ -869,9 +869,6 @@ fsp_init_file_page_low(
 		return;
 	}
 
-#ifdef UNIV_BASIC_LOG_DEBUG
-	memset(page, 0xff, UNIV_PAGE_SIZE);
-#endif
 	UNIV_MEM_INVALID(page, UNIV_PAGE_SIZE);
 	mach_write_to_4(page + FIL_PAGE_OFFSET, buf_block_get_page_no(block));
 	memset(page + FIL_PAGE_LSN, 0, 8);
diff --git a/storage/innodb_plugin/include/univ.i b/storage/innodb_plugin/include/univ.i
index a083620ab6d..4078167a66a 100644
--- a/storage/innodb_plugin/include/univ.i
+++ b/storage/innodb_plugin/include/univ.i
@@ -229,11 +229,6 @@ by one. */
 			/* the above option prevents forcing of log to disk
 			at a buffer page write: it should be tested with this
 			option off; also some ibuf tests are suppressed */
-/*
-#define UNIV_BASIC_LOG_DEBUG
-*/
-			/* the above option enables basic recovery debugging:
-			new allocated file pages are reset */
 
 /* Linkage specifier for non-static InnoDB symbols (variables and functions)
 that are only referenced from within InnoDB, not from MySQL */

From b9a88d944f49cd79354b49f8d2e2b23025cbe774 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:58:48 +0400
Subject: [PATCH 075/116] Applying InnoDB snapshot

Detailed revision comments:

r6740 | sunny | 2010-02-18 13:44:31 +0200 (Thu, 18 Feb 2010) | 6 lines
branches/zip: Don't print the entire lock bit set if the block was not
found in the buffer pool. Only print the bits that are set and that
information is in the lock and not in the block.

See rb://256 approved by Marko.
---
 storage/innodb_plugin/lock/lock0lock.c | 39 +++++++++++++-------------
 1 file changed, 20 insertions(+), 19 deletions(-)

diff --git a/storage/innodb_plugin/lock/lock0lock.c b/storage/innodb_plugin/lock/lock0lock.c
index 325ebc846bb..7650d1a0d22 100644
--- a/storage/innodb_plugin/lock/lock0lock.c
+++ b/storage/innodb_plugin/lock/lock0lock.c
@@ -4309,28 +4309,29 @@ lock_rec_print(
 
 	block = buf_page_try_get(space, page_no, &mtr);
 
-	if (block) {
-		for (i = 0; i < lock_rec_get_n_bits(lock); i++) {
+	for (i = 0; i < lock_rec_get_n_bits(lock); ++i) {
 
-			if (lock_rec_get_nth_bit(lock, i)) {
-
-				const rec_t*	rec
-					= page_find_rec_with_heap_no(
-						buf_block_get_frame(block), i);
-				offsets = rec_get_offsets(
-					rec, lock->index, offsets,
-					ULINT_UNDEFINED, &heap);
-
-				fprintf(file, "Record lock, heap no %lu ",
-					(ulong) i);
-				rec_print_new(file, rec, offsets);
-				putc('\n', file);
-			}
+		if (!lock_rec_get_nth_bit(lock, i)) {
+			continue;
 		}
-	} else {
-		for (i = 0; i < lock_rec_get_n_bits(lock); i++) {
-			fprintf(file, "Record lock, heap no %lu\n", (ulong) i);
+
+		fprintf(file, "Record lock, heap no %lu", (ulong) i);
+
+		if (block) {
+			const rec_t*	rec;
+
+			rec = page_find_rec_with_heap_no(
+				buf_block_get_frame(block), i);
+
+			offsets = rec_get_offsets(
+				rec, lock->index, offsets,
+				ULINT_UNDEFINED, &heap);
+
+			putc(' ', file);
+			rec_print_new(file, rec, offsets);
 		}
+
+		putc('\n', file);
 	}
 
 	mtr_commit(&mtr);

From 22f0da5d94ab22a72aacbc68d46dc02c15b59a84 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 15:59:25 +0400
Subject: [PATCH 076/116] Applying InnoDB snapshot

Detailed revision comments:

r6749 | vasil | 2010-02-20 18:45:41 +0200 (Sat, 20 Feb 2010) | 5 lines
Non-functional change: update copyright year to 2010 of the files
that have been modified after 2010-01-01 according to svn.

for f in $(svn log -v -r{2010-01-01}:HEAD |grep "^   M " |cut -b 16- |sort -u) ; do sed -i "" -E 's/(Copyright \(c\) [0-9]{4},) [0-9]{4}, (.*Innobase Oy.+All Rights Reserved)/\1 2010, \2/' $f ; done
---
 storage/innodb_plugin/btr/btr0btr.c            | 2 +-
 storage/innodb_plugin/btr/btr0cur.c            | 2 +-
 storage/innodb_plugin/btr/btr0pcur.c           | 2 +-
 storage/innodb_plugin/buf/buf0buf.c            | 2 +-
 storage/innodb_plugin/buf/buf0lru.c            | 2 +-
 storage/innodb_plugin/dict/dict0boot.c         | 2 +-
 storage/innodb_plugin/dict/dict0crea.c         | 2 +-
 storage/innodb_plugin/dict/dict0dict.c         | 2 +-
 storage/innodb_plugin/dict/dict0load.c         | 2 +-
 storage/innodb_plugin/fsp/fsp0fsp.c            | 2 +-
 storage/innodb_plugin/handler/ha_innodb.cc     | 2 +-
 storage/innodb_plugin/handler/ha_innodb.h      | 2 +-
 storage/innodb_plugin/handler/handler0alter.cc | 2 +-
 storage/innodb_plugin/include/btr0btr.h        | 2 +-
 storage/innodb_plugin/include/btr0cur.h        | 2 +-
 storage/innodb_plugin/include/btr0pcur.h       | 2 +-
 storage/innodb_plugin/include/btr0pcur.ic      | 2 +-
 storage/innodb_plugin/include/buf0buf.h        | 2 +-
 storage/innodb_plugin/include/dict0boot.h      | 2 +-
 storage/innodb_plugin/include/fil0fil.h        | 2 +-
 storage/innodb_plugin/include/log0log.h        | 2 +-
 storage/innodb_plugin/include/log0log.ic       | 2 +-
 storage/innodb_plugin/include/log0recv.h       | 2 +-
 storage/innodb_plugin/include/mem0dbg.h        | 2 +-
 storage/innodb_plugin/include/mem0dbg.ic       | 2 +-
 storage/innodb_plugin/include/mem0mem.h        | 2 +-
 storage/innodb_plugin/include/mem0mem.ic       | 2 +-
 storage/innodb_plugin/include/os0file.h        | 2 +-
 storage/innodb_plugin/include/que0que.h        | 2 +-
 storage/innodb_plugin/include/que0que.ic       | 2 +-
 storage/innodb_plugin/include/row0mysql.h      | 2 +-
 storage/innodb_plugin/include/srv0srv.h        | 2 +-
 storage/innodb_plugin/include/sync0sync.h      | 2 +-
 storage/innodb_plugin/include/trx0rseg.h       | 2 +-
 storage/innodb_plugin/include/trx0trx.h        | 2 +-
 storage/innodb_plugin/include/trx0types.h      | 2 +-
 storage/innodb_plugin/include/univ.i           | 2 +-
 storage/innodb_plugin/lock/lock0lock.c         | 2 +-
 storage/innodb_plugin/log/log0log.c            | 2 +-
 storage/innodb_plugin/log/log0recv.c           | 2 +-
 storage/innodb_plugin/mem/mem0dbg.c            | 2 +-
 storage/innodb_plugin/mem/mem0mem.c            | 2 +-
 storage/innodb_plugin/os/os0file.c             | 2 +-
 storage/innodb_plugin/page/page0page.c         | 2 +-
 storage/innodb_plugin/rem/rem0rec.c            | 2 +-
 storage/innodb_plugin/row/row0ins.c            | 2 +-
 storage/innodb_plugin/row/row0merge.c          | 2 +-
 storage/innodb_plugin/row/row0mysql.c          | 2 +-
 storage/innodb_plugin/row/row0sel.c            | 2 +-
 storage/innodb_plugin/row/row0umod.c           | 2 +-
 storage/innodb_plugin/srv/srv0srv.c            | 2 +-
 storage/innodb_plugin/srv/srv0start.c          | 2 +-
 storage/innodb_plugin/sync/sync0sync.c         | 2 +-
 storage/innodb_plugin/trx/trx0rseg.c           | 2 +-
 storage/innodb_plugin/trx/trx0trx.c            | 2 +-
 55 files changed, 55 insertions(+), 55 deletions(-)

diff --git a/storage/innodb_plugin/btr/btr0btr.c b/storage/innodb_plugin/btr/btr0btr.c
index 91f5e3a31b9..fa4fc05d96b 100644
--- a/storage/innodb_plugin/btr/btr0btr.c
+++ b/storage/innodb_plugin/btr/btr0btr.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/btr/btr0cur.c b/storage/innodb_plugin/btr/btr0cur.c
index 88827d33c61..504bd162f32 100644
--- a/storage/innodb_plugin/btr/btr0cur.c
+++ b/storage/innodb_plugin/btr/btr0cur.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
diff --git a/storage/innodb_plugin/btr/btr0pcur.c b/storage/innodb_plugin/btr/btr0pcur.c
index 436d1f252c3..2edfd673530 100644
--- a/storage/innodb_plugin/btr/btr0pcur.c
+++ b/storage/innodb_plugin/btr/btr0pcur.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/buf/buf0buf.c b/storage/innodb_plugin/buf/buf0buf.c
index 5312f772213..d5c14155f72 100644
--- a/storage/innodb_plugin/buf/buf0buf.c
+++ b/storage/innodb_plugin/buf/buf0buf.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
diff --git a/storage/innodb_plugin/buf/buf0lru.c b/storage/innodb_plugin/buf/buf0lru.c
index 62198491939..09878f789d4 100644
--- a/storage/innodb_plugin/buf/buf0lru.c
+++ b/storage/innodb_plugin/buf/buf0lru.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/dict/dict0boot.c b/storage/innodb_plugin/dict/dict0boot.c
index c1df2bac0bf..70b5bfa99f7 100644
--- a/storage/innodb_plugin/dict/dict0boot.c
+++ b/storage/innodb_plugin/dict/dict0boot.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/dict/dict0crea.c b/storage/innodb_plugin/dict/dict0crea.c
index 5bbc2d17ddd..4ba7cd8a48c 100644
--- a/storage/innodb_plugin/dict/dict0crea.c
+++ b/storage/innodb_plugin/dict/dict0crea.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/dict/dict0dict.c b/storage/innodb_plugin/dict/dict0dict.c
index 35b856cd79f..8a03151d062 100644
--- a/storage/innodb_plugin/dict/dict0dict.c
+++ b/storage/innodb_plugin/dict/dict0dict.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/dict/dict0load.c b/storage/innodb_plugin/dict/dict0load.c
index 0c72a2e8f81..377818308c5 100644
--- a/storage/innodb_plugin/dict/dict0load.c
+++ b/storage/innodb_plugin/dict/dict0load.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/fsp/fsp0fsp.c b/storage/innodb_plugin/fsp/fsp0fsp.c
index b5329bb9ebe..c7f1a299d8a 100644
--- a/storage/innodb_plugin/fsp/fsp0fsp.c
+++ b/storage/innodb_plugin/fsp/fsp0fsp.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 44196e131a5..01871dc6c9b 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 2000, 2009, MySQL AB & Innobase Oy. All Rights Reserved.
+Copyright (c) 2000, 2010, MySQL AB & Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, 2009 Google Inc.
 Copyright (c) 2009, Percona Inc.
 
diff --git a/storage/innodb_plugin/handler/ha_innodb.h b/storage/innodb_plugin/handler/ha_innodb.h
index 5dd9726ae4c..8a3e1ccff82 100644
--- a/storage/innodb_plugin/handler/ha_innodb.h
+++ b/storage/innodb_plugin/handler/ha_innodb.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 2000, 2009, MySQL AB & Innobase Oy. All Rights Reserved.
+Copyright (c) 2000, 2010, MySQL AB & Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/handler/handler0alter.cc b/storage/innodb_plugin/handler/handler0alter.cc
index fbbcfd6149b..94d1a4b2f4b 100644
--- a/storage/innodb_plugin/handler/handler0alter.cc
+++ b/storage/innodb_plugin/handler/handler0alter.cc
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 2005, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 2005, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/btr0btr.h b/storage/innodb_plugin/include/btr0btr.h
index 11cacfc2244..b6996d47d45 100644
--- a/storage/innodb_plugin/include/btr0btr.h
+++ b/storage/innodb_plugin/include/btr0btr.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/btr0cur.h b/storage/innodb_plugin/include/btr0cur.h
index 8be3672f22d..716f15c4267 100644
--- a/storage/innodb_plugin/include/btr0cur.h
+++ b/storage/innodb_plugin/include/btr0cur.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/btr0pcur.h b/storage/innodb_plugin/include/btr0pcur.h
index 509e61b9c8d..2334a266280 100644
--- a/storage/innodb_plugin/include/btr0pcur.h
+++ b/storage/innodb_plugin/include/btr0pcur.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/btr0pcur.ic b/storage/innodb_plugin/include/btr0pcur.ic
index 68acb68c0d3..0c38797e6c5 100644
--- a/storage/innodb_plugin/include/btr0pcur.ic
+++ b/storage/innodb_plugin/include/btr0pcur.ic
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/buf0buf.h b/storage/innodb_plugin/include/buf0buf.h
index cf47042423c..449f494f3c7 100644
--- a/storage/innodb_plugin/include/buf0buf.h
+++ b/storage/innodb_plugin/include/buf0buf.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/dict0boot.h b/storage/innodb_plugin/include/dict0boot.h
index 404fc23bfa3..e01fafe652d 100644
--- a/storage/innodb_plugin/include/dict0boot.h
+++ b/storage/innodb_plugin/include/dict0boot.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/fil0fil.h b/storage/innodb_plugin/include/fil0fil.h
index 053eb613ad1..de8ef9e9687 100644
--- a/storage/innodb_plugin/include/fil0fil.h
+++ b/storage/innodb_plugin/include/fil0fil.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/log0log.h b/storage/innodb_plugin/include/log0log.h
index 7fc24b58e4c..8fce4ef96bc 100644
--- a/storage/innodb_plugin/include/log0log.h
+++ b/storage/innodb_plugin/include/log0log.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2009, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
diff --git a/storage/innodb_plugin/include/log0log.ic b/storage/innodb_plugin/include/log0log.ic
index 098ef47d761..139f4041a36 100644
--- a/storage/innodb_plugin/include/log0log.ic
+++ b/storage/innodb_plugin/include/log0log.ic
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/log0recv.h b/storage/innodb_plugin/include/log0recv.h
index 3cbb85bbec8..3209799e140 100644
--- a/storage/innodb_plugin/include/log0recv.h
+++ b/storage/innodb_plugin/include/log0recv.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1997, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1997, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/mem0dbg.h b/storage/innodb_plugin/include/mem0dbg.h
index 8ddf4a13cba..d81e1418b2b 100644
--- a/storage/innodb_plugin/include/mem0dbg.h
+++ b/storage/innodb_plugin/include/mem0dbg.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/mem0dbg.ic b/storage/innodb_plugin/include/mem0dbg.ic
index 9c6e5a78263..b0c8178a623 100644
--- a/storage/innodb_plugin/include/mem0dbg.ic
+++ b/storage/innodb_plugin/include/mem0dbg.ic
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/mem0mem.h b/storage/innodb_plugin/include/mem0mem.h
index 21d300a8ce5..ee28cf7b225 100644
--- a/storage/innodb_plugin/include/mem0mem.h
+++ b/storage/innodb_plugin/include/mem0mem.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/mem0mem.ic b/storage/innodb_plugin/include/mem0mem.ic
index cb681c3f724..cbce2edc661 100644
--- a/storage/innodb_plugin/include/mem0mem.ic
+++ b/storage/innodb_plugin/include/mem0mem.ic
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/os0file.h b/storage/innodb_plugin/include/os0file.h
index ad31b37a2a1..d645cae38bb 100644
--- a/storage/innodb_plugin/include/os0file.h
+++ b/storage/innodb_plugin/include/os0file.h
@@ -1,6 +1,6 @@
 /***********************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2009, Percona Inc.
 
 Portions of this file contain modifications contributed and copyrighted
diff --git a/storage/innodb_plugin/include/que0que.h b/storage/innodb_plugin/include/que0que.h
index c35b75d7d51..39f8d07af89 100644
--- a/storage/innodb_plugin/include/que0que.h
+++ b/storage/innodb_plugin/include/que0que.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/que0que.ic b/storage/innodb_plugin/include/que0que.ic
index f027c748230..bd936670e1e 100644
--- a/storage/innodb_plugin/include/que0que.ic
+++ b/storage/innodb_plugin/include/que0que.ic
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/row0mysql.h b/storage/innodb_plugin/include/row0mysql.h
index 03b29fd6538..d2a8734c61f 100644
--- a/storage/innodb_plugin/include/row0mysql.h
+++ b/storage/innodb_plugin/include/row0mysql.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 2000, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 2000, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/srv0srv.h b/storage/innodb_plugin/include/srv0srv.h
index 8700cc6d66e..7aa2ce74720 100644
--- a/storage/innodb_plugin/include/srv0srv.h
+++ b/storage/innodb_plugin/include/srv0srv.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, 2009, Google Inc.
 Copyright (c) 2009, Percona Inc.
 
diff --git a/storage/innodb_plugin/include/sync0sync.h b/storage/innodb_plugin/include/sync0sync.h
index f8d86eb1c84..b93a8d1c080 100644
--- a/storage/innodb_plugin/include/sync0sync.h
+++ b/storage/innodb_plugin/include/sync0sync.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
diff --git a/storage/innodb_plugin/include/trx0rseg.h b/storage/innodb_plugin/include/trx0rseg.h
index e8f84c97dd9..a25d84f1e84 100644
--- a/storage/innodb_plugin/include/trx0rseg.h
+++ b/storage/innodb_plugin/include/trx0rseg.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/trx0trx.h b/storage/innodb_plugin/include/trx0trx.h
index e570f336462..480f265a138 100644
--- a/storage/innodb_plugin/include/trx0trx.h
+++ b/storage/innodb_plugin/include/trx0trx.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/trx0types.h b/storage/innodb_plugin/include/trx0types.h
index ae4451636cd..40a7256cbfd 100644
--- a/storage/innodb_plugin/include/trx0types.h
+++ b/storage/innodb_plugin/include/trx0types.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/univ.i b/storage/innodb_plugin/include/univ.i
index 4078167a66a..0df05999845 100644
--- a/storage/innodb_plugin/include/univ.i
+++ b/storage/innodb_plugin/include/univ.i
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 Copyright (c) 2009, Sun Microsystems, Inc.
 
diff --git a/storage/innodb_plugin/lock/lock0lock.c b/storage/innodb_plugin/lock/lock0lock.c
index 7650d1a0d22..d5fff572aee 100644
--- a/storage/innodb_plugin/lock/lock0lock.c
+++ b/storage/innodb_plugin/lock/lock0lock.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/log/log0log.c b/storage/innodb_plugin/log/log0log.c
index 6ddbcd22f77..183c24d2147 100644
--- a/storage/innodb_plugin/log/log0log.c
+++ b/storage/innodb_plugin/log/log0log.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2009, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
diff --git a/storage/innodb_plugin/log/log0recv.c b/storage/innodb_plugin/log/log0recv.c
index eb20cba7e32..de026cad7b7 100644
--- a/storage/innodb_plugin/log/log0recv.c
+++ b/storage/innodb_plugin/log/log0recv.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1997, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1997, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/mem/mem0dbg.c b/storage/innodb_plugin/mem/mem0dbg.c
index 4973ead4213..1cd2ff15bab 100644
--- a/storage/innodb_plugin/mem/mem0dbg.c
+++ b/storage/innodb_plugin/mem/mem0dbg.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/mem/mem0mem.c b/storage/innodb_plugin/mem/mem0mem.c
index 39bbfc90313..c0ce8a3e1ac 100644
--- a/storage/innodb_plugin/mem/mem0mem.c
+++ b/storage/innodb_plugin/mem/mem0mem.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/os/os0file.c b/storage/innodb_plugin/os/os0file.c
index f19e587053b..e1effd9b177 100644
--- a/storage/innodb_plugin/os/os0file.c
+++ b/storage/innodb_plugin/os/os0file.c
@@ -1,6 +1,6 @@
 /***********************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2009, Percona Inc.
 
 Portions of this file contain modifications contributed and copyrighted
diff --git a/storage/innodb_plugin/page/page0page.c b/storage/innodb_plugin/page/page0page.c
index 17c40170e14..1068a413e0c 100644
--- a/storage/innodb_plugin/page/page0page.c
+++ b/storage/innodb_plugin/page/page0page.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/rem/rem0rec.c b/storage/innodb_plugin/rem/rem0rec.c
index b2a8f45f89b..27c11dacc8c 100644
--- a/storage/innodb_plugin/rem/rem0rec.c
+++ b/storage/innodb_plugin/rem/rem0rec.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/row/row0ins.c b/storage/innodb_plugin/row/row0ins.c
index ece983eb957..230dc45dadc 100644
--- a/storage/innodb_plugin/row/row0ins.c
+++ b/storage/innodb_plugin/row/row0ins.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/row/row0merge.c b/storage/innodb_plugin/row/row0merge.c
index 1608751dc89..3f61410c04c 100644
--- a/storage/innodb_plugin/row/row0merge.c
+++ b/storage/innodb_plugin/row/row0merge.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 2005, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 2005, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/row/row0mysql.c b/storage/innodb_plugin/row/row0mysql.c
index 0223a389b5d..0d8d298453c 100644
--- a/storage/innodb_plugin/row/row0mysql.c
+++ b/storage/innodb_plugin/row/row0mysql.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 2000, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 2000, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/row/row0sel.c b/storage/innodb_plugin/row/row0sel.c
index 1656657f9e9..1769437301e 100644
--- a/storage/innodb_plugin/row/row0sel.c
+++ b/storage/innodb_plugin/row/row0sel.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1997, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1997, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
diff --git a/storage/innodb_plugin/row/row0umod.c b/storage/innodb_plugin/row/row0umod.c
index 12cd05766dd..e7245dbee41 100644
--- a/storage/innodb_plugin/row/row0umod.c
+++ b/storage/innodb_plugin/row/row0umod.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1997, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1997, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/srv/srv0srv.c b/storage/innodb_plugin/srv/srv0srv.c
index 1dba8f3ba2b..61413b8f014 100644
--- a/storage/innodb_plugin/srv/srv0srv.c
+++ b/storage/innodb_plugin/srv/srv0srv.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, 2009 Google Inc.
 Copyright (c) 2009, Percona Inc.
 
diff --git a/storage/innodb_plugin/srv/srv0start.c b/storage/innodb_plugin/srv/srv0start.c
index 3293fbb5bb7..e517b9a86b0 100644
--- a/storage/innodb_plugin/srv/srv0start.c
+++ b/storage/innodb_plugin/srv/srv0start.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 Copyright (c) 2009, Percona Inc.
 
diff --git a/storage/innodb_plugin/sync/sync0sync.c b/storage/innodb_plugin/sync/sync0sync.c
index 669e0667485..2be9d667705 100644
--- a/storage/innodb_plugin/sync/sync0sync.c
+++ b/storage/innodb_plugin/sync/sync0sync.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
diff --git a/storage/innodb_plugin/trx/trx0rseg.c b/storage/innodb_plugin/trx/trx0rseg.c
index 3bc497bc135..36dea9b2a95 100644
--- a/storage/innodb_plugin/trx/trx0rseg.c
+++ b/storage/innodb_plugin/trx/trx0rseg.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/trx/trx0trx.c b/storage/innodb_plugin/trx/trx0trx.c
index 7fe0bfb4da8..e8c98e22918 100644
--- a/storage/innodb_plugin/trx/trx0trx.c
+++ b/storage/innodb_plugin/trx/trx0trx.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software

From 9a2b8e9ebb120725eecc9afe668179c9a285727c Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:00:18 +0400
Subject: [PATCH 077/116] Applying InnoDB snapshot

Detailed revision comments:

r6750 | marko | 2010-02-22 08:57:23 +0200 (Mon, 22 Feb 2010) | 2 lines
branches/zip: row_fetch_store_uint4(): Remove unused function.
This was added to trunk in r435.
---
 storage/innodb_plugin/include/row0sel.h | 11 ---------
 storage/innodb_plugin/row/row0sel.c     | 30 -------------------------
 2 files changed, 41 deletions(-)

diff --git a/storage/innodb_plugin/include/row0sel.h b/storage/innodb_plugin/include/row0sel.h
index 01a5afaa23e..430493e4cde 100644
--- a/storage/innodb_plugin/include/row0sel.h
+++ b/storage/innodb_plugin/include/row0sel.h
@@ -105,17 +105,6 @@ row_fetch_print(
 /*============*/
 	void*	row,		/*!< in:  sel_node_t* */
 	void*	user_arg);	/*!< in:  not used */
-/****************************************************************//**
-Callback function for fetch that stores an unsigned 4 byte integer to the
-location pointed. The column's type must be DATA_INT, DATA_UNSIGNED, length
-= 4.
-@return	always returns NULL */
-UNIV_INTERN
-void*
-row_fetch_store_uint4(
-/*==================*/
-	void*	row,		/*!< in:  sel_node_t* */
-	void*	user_arg);	/*!< in:  data pointer */
 /***********************************************************//**
 Prints a row in a select result.
 @return	query thread to run next or NULL */
diff --git a/storage/innodb_plugin/row/row0sel.c b/storage/innodb_plugin/row/row0sel.c
index 1769437301e..78318bf6461 100644
--- a/storage/innodb_plugin/row/row0sel.c
+++ b/storage/innodb_plugin/row/row0sel.c
@@ -2167,36 +2167,6 @@ row_fetch_print(
 	return((void*)42);
 }
 
-/****************************************************************//**
-Callback function for fetch that stores an unsigned 4 byte integer to the
-location pointed. The column's type must be DATA_INT, DATA_UNSIGNED, length
-= 4.
-@return	always returns NULL */
-UNIV_INTERN
-void*
-row_fetch_store_uint4(
-/*==================*/
-	void*	row,		/*!< in:  sel_node_t* */
-	void*	user_arg)	/*!< in:  data pointer */
-{
-	sel_node_t*	node = row;
-	ib_uint32_t*	val = user_arg;
-	ulint		tmp;
-
-	dfield_t*	dfield = que_node_get_val(node->select_list);
-	const dtype_t*	type = dfield_get_type(dfield);
-	ulint		len = dfield_get_len(dfield);
-
-	ut_a(dtype_get_mtype(type) == DATA_INT);
-	ut_a(dtype_get_prtype(type) & DATA_UNSIGNED);
-	ut_a(len == 4);
-
-	tmp = mach_read_from_4(dfield_get_data(dfield));
-	*val = (ib_uint32_t) tmp;
-
-	return(NULL);
-}
-
 /***********************************************************//**
 Prints a row in a select result.
 @return	query thread to run next or NULL */

From 3fb46ab4467b1d06ccec61749f02127a8f9a19dc Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:01:10 +0400
Subject: [PATCH 078/116] Applying InnoDB snapshot

Detailed revision comments:

r6754 | marko | 2010-02-24 10:56:43 +0200 (Wed, 24 Feb 2010) | 17 lines
branches/zip: Allocate the merge sort buffers from a heap, not stack.

The merge sort can use up to 48KiB of buffers when merging blocks.
That can cause a stack overflow, especially on 64-bit systems when not
building with inlined functions.  This was reported as Issue #462.

row_merge_dup_report(): Allocate buf and offsets from a heap.

row_merge_heap_create(): Allocate space for buf[3] too. Fix bogus
sizeof arithmetics that happened to work, because
sizeof(ulint)==sizeof(void*).

row_merge_blocks(), row_merge_blocks_copy(): Allocate buf[3] from heap.

row_merge_insert_index_tuples(): Allocate buf from graph_heap.

rb://258 approved and tested by Sunny Bains
---
 storage/innodb_plugin/row/row0merge.c | 49 +++++++++++++++------------
 1 file changed, 27 insertions(+), 22 deletions(-)

diff --git a/storage/innodb_plugin/row/row0merge.c b/storage/innodb_plugin/row/row0merge.c
index 3f61410c04c..88a2770032d 100644
--- a/storage/innodb_plugin/row/row0merge.c
+++ b/storage/innodb_plugin/row/row0merge.c
@@ -424,14 +424,13 @@ row_merge_dup_report(
 	row_merge_dup_t*	dup,	/*!< in/out: for reporting duplicates */
 	const dfield_t*		entry)	/*!< in: duplicate index entry */
 {
-	mrec_buf_t 		buf;
+	mrec_buf_t* 		buf;
 	const dtuple_t*		tuple;
 	dtuple_t		tuple_store;
 	const rec_t*		rec;
 	const dict_index_t*	index	= dup->index;
 	ulint			n_fields= dict_index_get_n_fields(index);
-	mem_heap_t*		heap	= NULL;
-	ulint			offsets_[REC_OFFS_NORMAL_SIZE];
+	mem_heap_t*		heap;
 	ulint*			offsets;
 	ulint			n_ext;
 
@@ -441,22 +440,22 @@ row_merge_dup_report(
 		return;
 	}
 
-	rec_offs_init(offsets_);
-
 	/* Convert the tuple to a record and then to MySQL format. */
+	heap = mem_heap_create((1 + REC_OFFS_HEADER_SIZE + n_fields)
+			       * sizeof *offsets
+			       + sizeof *buf);
+
+	buf = mem_heap_alloc(heap, sizeof *buf);
 
 	tuple = dtuple_from_fields(&tuple_store, entry, n_fields);
 	n_ext = dict_index_is_clust(index) ? dtuple_get_n_ext(tuple) : 0;
 
-	rec = rec_convert_dtuple_to_rec(buf, index, tuple, n_ext);
-	offsets = rec_get_offsets(rec, index, offsets_, ULINT_UNDEFINED,
-				  &heap);
+	rec = rec_convert_dtuple_to_rec(*buf, index, tuple, n_ext);
+	offsets = rec_get_offsets(rec, index, NULL, ULINT_UNDEFINED, &heap);
 
 	innobase_rec_to_mysql(dup->table, rec, index, offsets);
 
-	if (UNIV_LIKELY_NULL(heap)) {
-		mem_heap_free(heap);
-	}
+	mem_heap_free(heap);
 }
 
 /*************************************************************//**
@@ -627,22 +626,26 @@ row_merge_buf_write(
 }
 
 /******************************************************//**
-Create a memory heap and allocate space for row_merge_rec_offsets().
+Create a memory heap and allocate space for row_merge_rec_offsets()
+and mrec_buf_t[3].
 @return	memory heap */
 static
 mem_heap_t*
 row_merge_heap_create(
 /*==================*/
 	const dict_index_t*	index,		/*!< in: record descriptor */
+	mrec_buf_t**		buf,		/*!< out: 3 buffers */
 	ulint**			offsets1,	/*!< out: offsets */
 	ulint**			offsets2)	/*!< out: offsets */
 {
 	ulint		i	= 1 + REC_OFFS_HEADER_SIZE
 		+ dict_index_get_n_fields(index);
-	mem_heap_t*	heap	= mem_heap_create(2 * i * sizeof *offsets1);
+	mem_heap_t*	heap	= mem_heap_create(2 * i * sizeof **offsets1
+						  + 3 * sizeof **buf);
 
-	*offsets1 = mem_heap_alloc(heap, i * sizeof *offsets1);
-	*offsets2 = mem_heap_alloc(heap, i * sizeof *offsets2);
+	*buf = mem_heap_alloc(heap, 3 * sizeof **buf);
+	*offsets1 = mem_heap_alloc(heap, i * sizeof **offsets1);
+	*offsets2 = mem_heap_alloc(heap, i * sizeof **offsets2);
 
 	(*offsets1)[0] = (*offsets2)[0] = i;
 	(*offsets1)[1] = (*offsets2)[1] = dict_index_get_n_fields(index);
@@ -1394,7 +1397,8 @@ row_merge_blocks(
 {
 	mem_heap_t*	heap;	/*!< memory heap for offsets0, offsets1 */
 
-	mrec_buf_t	buf[3];	/*!< buffer for handling split mrec in block[] */
+	mrec_buf_t*	buf;	/*!< buffer for handling
+				split mrec in block[] */
 	const byte*	b0;	/*!< pointer to block[0] */
 	const byte*	b1;	/*!< pointer to block[1] */
 	byte*		b2;	/*!< pointer to block[2] */
@@ -1414,7 +1418,7 @@ row_merge_blocks(
 	}
 #endif /* UNIV_DEBUG */
 
-	heap = row_merge_heap_create(index, &offsets0, &offsets1);
+	heap = row_merge_heap_create(index, &buf, &offsets0, &offsets1);
 
 	/* Write a record and read the next record.  Split the output
 	file in two halves, which can be merged on the following pass. */
@@ -1500,7 +1504,7 @@ row_merge_blocks_copy(
 {
 	mem_heap_t*	heap;	/*!< memory heap for offsets0, offsets1 */
 
-	mrec_buf_t	buf[3];	/*!< buffer for handling
+	mrec_buf_t*	buf;	/*!< buffer for handling
 				split mrec in block[] */
 	const byte*	b0;	/*!< pointer to block[0] */
 	byte*		b2;	/*!< pointer to block[2] */
@@ -1518,7 +1522,7 @@ row_merge_blocks_copy(
 	}
 #endif /* UNIV_DEBUG */
 
-	heap = row_merge_heap_create(index, &offsets0, &offsets1);
+	heap = row_merge_heap_create(index, &buf, &offsets0, &offsets1);
 
 	/* Write a record and read the next record.  Split the output
 	file in two halves, which can be merged on the following pass. */
@@ -1760,7 +1764,6 @@ row_merge_insert_index_tuples(
 	int			fd,	/*!< in: file descriptor */
 	row_merge_block_t*	block)	/*!< in/out: file buffer */
 {
-	mrec_buf_t		buf;
 	const byte*		b;
 	que_thr_t*		thr;
 	ins_node_t*		node;
@@ -1779,7 +1782,7 @@ row_merge_insert_index_tuples(
 
 	trx->op_info = "inserting index entries";
 
-	graph_heap = mem_heap_create(500);
+	graph_heap = mem_heap_create(500 + sizeof(mrec_buf_t));
 	node = ins_node_create(INS_DIRECT, table, graph_heap);
 
 	thr = pars_complete_graph_for_exec(node, trx, graph_heap);
@@ -1801,12 +1804,14 @@ row_merge_insert_index_tuples(
 	if (!row_merge_read(fd, foffs, block)) {
 		error = DB_CORRUPTION;
 	} else {
+		mrec_buf_t*	buf = mem_heap_alloc(graph_heap, sizeof *buf);
+
 		for (;;) {
 			const mrec_t*	mrec;
 			dtuple_t*	dtuple;
 			ulint		n_ext;
 
-			b = row_merge_read_rec(block, &buf, b, index,
+			b = row_merge_read_rec(block, buf, b, index,
 					       fd, &foffs, &mrec, offsets);
 			if (UNIV_UNLIKELY(!b)) {
 				/* End of list, or I/O error */

From b866cdae16e3f41302f76eacd65f262b6014b44b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:01:33 +0400
Subject: [PATCH 079/116] Applying InnoDB snashot, fixes BUG#51587.

Detailed revision comments:

r6767 | calvin | 2010-03-01 18:16:10 +0200 (Mon, 01 Mar 2010) | 3 lines
branches/zip: fix bug#51587
Non-functional change.
---
 storage/innodb_plugin/srv/srv0srv.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/storage/innodb_plugin/srv/srv0srv.c b/storage/innodb_plugin/srv/srv0srv.c
index 61413b8f014..78f11cfd2f3 100644
--- a/storage/innodb_plugin/srv/srv0srv.c
+++ b/storage/innodb_plugin/srv/srv0srv.c
@@ -1717,9 +1717,9 @@ srv_printf_innodb_monitor(
 		"Per second averages calculated from the last %lu seconds\n",
 		(ulong)time_elapsed);
 
-	fputs("----------\n"
-		"BACKGROUND THREAD\n"
-		"----------\n", file);
+	fputs("-----------------\n"
+	      "BACKGROUND THREAD\n"
+	      "-----------------\n", file);
 	srv_print_master_thread_info(file);
 
 	fputs("----------\n"

From b8871cd5522333850360eccdfc7020b6ed0e754d Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:02:34 +0400
Subject: [PATCH 080/116] Applying InnoDB snapshot

Detailed revision comments:

r6768 | vasil | 2010-03-02 18:20:48 +0200 (Tue, 02 Mar 2010) | 5 lines
branches/zip:

Add a NOTE to the comment of btr_node_ptr_get_child_page_no()
to prevent mysterious bugs.
---
 storage/innodb_plugin/include/btr0btr.h  | 4 ++++
 storage/innodb_plugin/include/btr0btr.ic | 4 ++++
 2 files changed, 8 insertions(+)

diff --git a/storage/innodb_plugin/include/btr0btr.h b/storage/innodb_plugin/include/btr0btr.h
index b6996d47d45..5e6a76c7d21 100644
--- a/storage/innodb_plugin/include/btr0btr.h
+++ b/storage/innodb_plugin/include/btr0btr.h
@@ -193,6 +193,10 @@ btr_leaf_page_release(
 	mtr_t*		mtr);		/*!< in: mtr */
 /**************************************************************//**
 Gets the child node file address in a node pointer.
+NOTE: the offsets array must contain all offsets for the record since
+we read the last field according to offsets and assume that it contains
+the child page number. In other words offsets must have been retrieved
+with rec_get_offsets(n_fields=ULINT_UNDEFINED).
 @return	child node address */
 UNIV_INLINE
 ulint
diff --git a/storage/innodb_plugin/include/btr0btr.ic b/storage/innodb_plugin/include/btr0btr.ic
index 2259d22c9a6..4ec27117d85 100644
--- a/storage/innodb_plugin/include/btr0btr.ic
+++ b/storage/innodb_plugin/include/btr0btr.ic
@@ -255,6 +255,10 @@ btr_page_set_prev(
 
 /**************************************************************//**
 Gets the child node file address in a node pointer.
+NOTE: the offsets array must contain all offsets for the record since
+we read the last field according to offsets and assume that it contains
+the child page number. In other words offsets must have been retrieved
+with rec_get_offsets(n_fields=ULINT_UNDEFINED).
 @return	child node address */
 UNIV_INLINE
 ulint

From bb8467ecc5be55472702495092373e063021de0b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:04:43 +0400
Subject: [PATCH 081/116] Applying InnoDB snapshot

Detailed revision comments:

r6770 | marko | 2010-03-03 12:52:55 +0200 (Wed, 03 Mar 2010) | 12 lines
branches/zip: Disallow duplicate index name when creating an index.
This should fix Mantis Issue #461.

innodb.test, innodb.result, innodb-index.test, innodb-index.result:
Adjust the test result and mention that the introduced restriction
has been reported as MySQL Bug #51451.

innobase_check_index_keys(): Add a parameter for the InnoDB table and
check that no duplicate index name is added.  Report errors by
my_error() instead of sql_print_error().

rb://260 approved by Sunny Bains
---
 mysql-test/suite/innodb/r/innodb-index.result |  3 +-
 mysql-test/suite/innodb/t/innodb-index.test   |  2 +
 .../innodb_plugin/handler/handler0alter.cc    | 67 +++++++++----------
 3 files changed, 35 insertions(+), 37 deletions(-)

diff --git a/mysql-test/suite/innodb/r/innodb-index.result b/mysql-test/suite/innodb/r/innodb-index.result
index 67fbe0dce02..f86fcd4a8ef 100644
--- a/mysql-test/suite/innodb/r/innodb-index.result
+++ b/mysql-test/suite/innodb/r/innodb-index.result
@@ -441,6 +441,7 @@ t3	CREATE TABLE `t3` (
   KEY `c` (`c`)
 ) ENGINE=InnoDB DEFAULT CHARSET=latin1
 alter table t2 drop index b, add index (b);
+ERROR 42000: Incorrect index name 'b'
 show create table t2;
 Table	Create Table
 t2	CREATE TABLE `t2` (
@@ -451,8 +452,8 @@ t2	CREATE TABLE `t2` (
   `e` int(11) DEFAULT NULL,
   PRIMARY KEY (`a`),
   UNIQUE KEY `dc` (`d`,`c`),
-  KEY `c` (`c`),
   KEY `b` (`b`),
+  KEY `c` (`c`),
   CONSTRAINT `t2_ibfk_1` FOREIGN KEY (`b`) REFERENCES `t1` (`b`) ON DELETE CASCADE,
   CONSTRAINT `t2_ibfk_2` FOREIGN KEY (`c`) REFERENCES `t3` (`c`),
   CONSTRAINT `t2_ibfk_3` FOREIGN KEY (`d`) REFERENCES `t4` (`d`)
diff --git a/mysql-test/suite/innodb/t/innodb-index.test b/mysql-test/suite/innodb/t/innodb-index.test
index 09d6e2f8e3f..eeadc0f4783 100644
--- a/mysql-test/suite/innodb/t/innodb-index.test
+++ b/mysql-test/suite/innodb/t/innodb-index.test
@@ -141,6 +141,8 @@ show create table t4;
 --error ER_CANT_CREATE_TABLE
 alter table t3 add constraint dc foreign key (a) references t1(a);
 show create table t3;
+# this should be fixed by MySQL (see Bug #51451)
+--error ER_WRONG_NAME_FOR_INDEX
 alter table t2 drop index b, add index (b);
 show create table t2;
 --error ER_ROW_IS_REFERENCED_2
diff --git a/storage/innodb_plugin/handler/handler0alter.cc b/storage/innodb_plugin/handler/handler0alter.cc
index 94d1a4b2f4b..1c99cd8e984 100644
--- a/storage/innodb_plugin/handler/handler0alter.cc
+++ b/storage/innodb_plugin/handler/handler0alter.cc
@@ -229,9 +229,11 @@ static
 int
 innobase_check_index_keys(
 /*======================*/
-	const KEY*	key_info,	/*!< in: Indexes to be created */
-	ulint		num_of_keys)	/*!< in: Number of indexes to
-					be created */
+	const KEY*		key_info,	/*!< in: Indexes to be
+						created */
+	ulint			num_of_keys,	/*!< in: Number of
+						indexes to be created */
+	const dict_table_t*	table)		/*!< in: Existing indexes */
 {
 	ulint		key_num;
 
@@ -248,9 +250,22 @@ innobase_check_index_keys(
 			const KEY&	key2 = key_info[i];
 
 			if (0 == strcmp(key.name, key2.name)) {
-				sql_print_error("InnoDB: key name `%s` appears"
-						" twice in CREATE INDEX\n",
-						key.name);
+				my_error(ER_WRONG_NAME_FOR_INDEX, MYF(0),
+					 key.name);
+
+				return(ER_WRONG_NAME_FOR_INDEX);
+			}
+		}
+
+		/* Check that the same index name does not already exist. */
+
+		for (const dict_index_t* index
+			     = dict_table_get_first_index(table);
+		     index; index = dict_table_get_next_index(index)) {
+
+			if (0 == strcmp(key.name, index->name)) {
+				my_error(ER_WRONG_NAME_FOR_INDEX, MYF(0),
+					 key.name);
 
 				return(ER_WRONG_NAME_FOR_INDEX);
 			}
@@ -258,7 +273,7 @@ innobase_check_index_keys(
 
 		/* Check that MySQL does not try to create a column
 		prefix index field on an inappropriate data type and
-		that the same colum does not appear twice in the index. */
+		that the same column does not appear twice in the index. */
 
 		for (ulint i = 0; i < key.key_parts; i++) {
 			const KEY_PART_INFO&	key_part1
@@ -289,14 +304,8 @@ innobase_check_index_keys(
 					}
 				}
 
-				sql_print_error("InnoDB: MySQL is trying to"
-						" create a column prefix"
-						" index field on an"
-						" inappropriate data type."
-						" column `%s`,"
-						" index `%s`.\n",
-						field->field_name,
-						key.name);
+				my_error(ER_WRONG_KEY_COLUMN, MYF(0),
+					 field->field_name);
 				return(ER_WRONG_KEY_COLUMN);
 			}
 
@@ -309,11 +318,8 @@ innobase_check_index_keys(
 					continue;
 				}
 
-				sql_print_error("InnoDB: column `%s`"
-						" is not allowed to occur"
-						" twice in index `%s`.\n",
-						key_part1.field->field_name,
-						key.name);
+				my_error(ER_WRONG_KEY_COLUMN, MYF(0),
+					 key_part1.field->field_name);
 				return(ER_WRONG_KEY_COLUMN);
 			}
 		}
@@ -666,7 +672,8 @@ ha_innobase::add_index(
 		error = -1;
 	} else {
 		/* Check that index keys are sensible */
-		error = innobase_check_index_keys(key_info, num_of_keys);
+		error = innobase_check_index_keys(key_info, num_of_keys,
+						  innodb_table);
 	}
 
 	if (UNIV_UNLIKELY(error)) {
@@ -808,18 +815,6 @@ err_exit:
 					index, num_of_idx, table);
 
 error_handling:
-#ifdef UNIV_DEBUG
-	/* TODO: At the moment we can't handle the following statement
-	in our debugging code below:
-
-	alter table t drop index b, add index (b);
-
-	The fix will have to parse the SQL and note that the index
-	being added has the same name as the one being dropped and
-	ignore that in the dup index check.*/
-	//dict_table_check_for_dup_indexes(prebuilt->table);
-#endif
-
 	/* After an error, remove all those index definitions from the
 	dictionary which were defined. */
 
@@ -831,6 +826,8 @@ error_handling:
 		row_mysql_lock_data_dictionary(trx);
 		dict_locked = TRUE;
 
+		ut_d(dict_table_check_for_dup_indexes(prebuilt->table));
+
 		if (!new_primary) {
 			error = row_merge_rename_indexes(trx, indexed_table);
 
@@ -1211,9 +1208,7 @@ ha_innobase::final_drop_index(
 	valid index entry count in the translation table to zero */
 	share->idx_trans_tbl.index_count = 0;
 
-#ifdef UNIV_DEBUG
-	dict_table_check_for_dup_indexes(prebuilt->table);
-#endif
+	ut_d(dict_table_check_for_dup_indexes(prebuilt->table));
 
 func_exit:
 	trx_commit_for_mysql(trx);

From d1abc1aae6ec30740fe8a221a2a911787d8f0ae8 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:05:25 +0400
Subject: [PATCH 082/116] Applying InnoDB snapshot

Detailed revision comments:

r6771 | marko | 2010-03-03 14:52:43 +0200 (Wed, 03 Mar 2010) | 1 line
Document r6770.
---
 storage/innodb_plugin/ChangeLog | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 3d26ebb82bb..81d60808905 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,7 +1,13 @@
+2010-03-03	The InnoDB Team
+
+	* handler/handler0alter.cc, innodb-index.result, innodb-index.test,
+	innodb.result, innodb.test:
+	Disallow a duplicate index name when creating an index.
+
 2010-02-11	The InnoDB Team
 
 	* include/mem0mem.h, include/mem0mem.ic, mem/mem0mem.c:
-	Fix bug #49535 Available memory check slows down crash
+	Fix Bug #49535 Available memory check slows down crash
 	recovery tens of times
 
 2010-02-09	The InnoDB Team

From bbc9a7fedaaafd5677e98947f59b43f9d0e9a7a2 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:05:52 +0400
Subject: [PATCH 083/116] Applying InnoDB snapshot

Detailed revision comments:

r6773 | marko | 2010-03-03 15:31:54 +0200 (Wed, 03 Mar 2010) | 2 lines
branches/zip: row_raw_format(): Silence a GCC 4.4.2 warning
of possibly uninitialized variable format_in_hex.
---
 storage/innodb_plugin/row/row0row.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/storage/innodb_plugin/row/row0row.c b/storage/innodb_plugin/row/row0row.c
index 128ac3ba3e8..3103d95bfb2 100644
--- a/storage/innodb_plugin/row/row0row.c
+++ b/storage/innodb_plugin/row/row0row.c
@@ -915,6 +915,10 @@ row_raw_format(
 
 		ret = row_raw_format_int(data, data_len, prtype,
 					 buf, buf_size, &format_in_hex);
+		if (format_in_hex) {
+
+			goto format_in_hex;
+		}
 		break;
 	case DATA_CHAR:
 	case DATA_VARCHAR:
@@ -923,14 +927,15 @@ row_raw_format(
 
 		ret = row_raw_format_str(data, data_len, prtype,
 					 buf, buf_size, &format_in_hex);
+		if (format_in_hex) {
+
+			goto format_in_hex;
+		}
+
 		break;
 	/* XXX support more data types */
 	default:
-
-		format_in_hex = TRUE;
-	}
-
-	if (format_in_hex) {
+	format_in_hex:
 
 		if (UNIV_LIKELY(buf_size > 2)) {
 

From 9137d3b355368b190da81c402f0b813c0d5acfce Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:06:16 +0400
Subject: [PATCH 084/116] Applying InnoDB snapshot

Detailed revision comments:

r6777 | marko | 2010-03-04 13:01:25 +0200 (Thu, 04 Mar 2010) | 2 lines
branches/zip: trx_undo_update_rec_get_update(): Silence a bogus GCC warning
about a possibly uninitialized variable.
---
 storage/innodb_plugin/trx/trx0rec.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/storage/innodb_plugin/trx/trx0rec.c b/storage/innodb_plugin/trx/trx0rec.c
index 5097cf18dcd..38a0e4f0f44 100644
--- a/storage/innodb_plugin/trx/trx0rec.c
+++ b/storage/innodb_plugin/trx/trx0rec.c
@@ -977,6 +977,7 @@ trx_undo_update_rec_get_update(
 			fprintf(stderr, "\n"
 				"InnoDB: n_fields = %lu, i = %lu, ptr %p\n",
 				(ulong) n_fields, (ulong) i, ptr);
+			*upd = NULL;
 			return(NULL);
 		}
 

From 654b4752a12da79511095566842a5cb1b7c1c510 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:06:36 +0400
Subject: [PATCH 085/116] Applying InnoDB snapshot

Detailed revision comments:

r6779 | marko | 2010-03-08 14:35:42 +0200 (Mon, 08 Mar 2010) | 6 lines
branches/zip: Fix IMPORT TABLESPACE of compressed tables.  Previously,
a wrong parameter was passed to buf_flush_init_for_writing().

fil_reset_too_high_lsns(): Set up page_zip and use it if needed.

rb://264, Issue #352
---
 storage/innodb_plugin/ChangeLog     |  5 +++++
 storage/innodb_plugin/fil/fil0fil.c | 29 ++++++++++++++++++++---------
 2 files changed, 25 insertions(+), 9 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 81d60808905..340f0e6d7fa 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,8 @@
+2010-03-08	The InnoDB Team
+
+	* fil/fil0fil.c:
+	Fix ALTER TABLE ... IMPORT TABLESPACE of compressed tables.
+
 2010-03-03	The InnoDB Team
 
 	* handler/handler0alter.cc, innodb-index.result, innodb-index.test,
diff --git a/storage/innodb_plugin/fil/fil0fil.c b/storage/innodb_plugin/fil/fil0fil.c
index 8819113f833..16353959cb5 100644
--- a/storage/innodb_plugin/fil/fil0fil.c
+++ b/storage/innodb_plugin/fil/fil0fil.c
@@ -38,6 +38,7 @@ Created 10/25/1995 Heikki Tuuri
 #include "mtr0mtr.h"
 #include "mtr0log.h"
 #include "dict0dict.h"
+#include "page0page.h"
 #include "page0zip.h"
 #ifndef UNIV_HOTBACKUP
 # include "buf0lru.h"
@@ -2788,6 +2789,7 @@ fil_reset_too_high_lsns(
 	ib_int64_t	offset;
 	ulint		zip_size;
 	ibool		success;
+	page_zip_des_t	page_zip;
 
 	filepath = fil_make_ibd_name(name, FALSE);
 
@@ -2835,6 +2837,12 @@ fil_reset_too_high_lsns(
 	space_id = fsp_header_get_space_id(page);
 	zip_size = fsp_header_get_zip_size(page);
 
+	page_zip_des_init(&page_zip);
+	page_zip_set_size(&page_zip, zip_size);
+	if (zip_size) {
+		page_zip.data = page + UNIV_PAGE_SIZE;
+	}
+
 	ut_print_timestamp(stderr);
 	fprintf(stderr,
 		"  InnoDB: Flush lsn in the tablespace file %lu"
@@ -2869,20 +2877,23 @@ fil_reset_too_high_lsns(
 			/* We have to reset the lsn */
 
 			if (zip_size) {
-				memcpy(page + UNIV_PAGE_SIZE, page, zip_size);
+				memcpy(page_zip.data, page, zip_size);
 				buf_flush_init_for_writing(
-					page, page + UNIV_PAGE_SIZE,
-					current_lsn);
+					page, &page_zip, current_lsn);
+				success = os_file_write(
+					filepath, file, page_zip.data,
+					(ulint) offset & 0xFFFFFFFFUL,
+					(ulint) (offset >> 32), zip_size);
 			} else {
 				buf_flush_init_for_writing(
 					page, NULL, current_lsn);
+				success = os_file_write(
+					filepath, file, page,
+					(ulint)(offset & 0xFFFFFFFFUL),
+					(ulint)(offset >> 32),
+					UNIV_PAGE_SIZE);
 			}
-			success = os_file_write(filepath, file, page,
-						(ulint)(offset & 0xFFFFFFFFUL),
-						(ulint)(offset >> 32),
-						zip_size
-						? zip_size
-						: UNIV_PAGE_SIZE);
+
 			if (!success) {
 
 				goto func_exit;

From c7280ed8d183ae0f2419e3ab6710ab95ce38a9fa Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:06:59 +0400
Subject: [PATCH 086/116] Applying InnoDB snapshot

Detailed revision comments:

r6781 | marko | 2010-03-09 09:41:08 +0200 (Tue, 09 Mar 2010) | 4 lines
branches/zip: Make SHOW ENGINE INNODB MUTEX display SUM(os_waits)
for block mutexes and blocks.

Designed by Michael and Marko. rb://188, Issue #358
---
 storage/innodb_plugin/ChangeLog            |   6 +
 storage/innodb_plugin/handler/ha_innodb.cc | 121 ++++++++++++++-------
 2 files changed, 90 insertions(+), 37 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 340f0e6d7fa..d7d1e89c147 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,9 @@
+2010-03-09	The InnoDB Team
+
+	* handler/ha_innodb.cc:
+	Make SHOW ENGINE INNODB MUTEX STATUS display SUM(os_waits)
+	for the buffer pool block mutexes and locks.
+
 2010-03-08	The InnoDB Team
 
 	* fil/fil0fil.c:
diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 01871dc6c9b..86587fee503 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -8659,19 +8659,25 @@ innodb_show_status(
 }
 
 /************************************************************************//**
-Implements the SHOW MUTEX STATUS command. . */
+Implements the SHOW MUTEX STATUS command.
+@return TRUE on failure, FALSE on success. */
 static
 bool
 innodb_mutex_show_status(
 /*=====================*/
-	handlerton*	hton,	/*!< in: the innodb handlerton */
+	handlerton*	hton,		/*!< in: the innodb handlerton */
 	THD*		thd,		/*!< in: the MySQL query thread of the
 					caller */
-	stat_print_fn*	stat_print)
+	stat_print_fn*	stat_print)	/*!< in: function for printing
+					statistics */
 {
 	char buf1[IO_SIZE], buf2[IO_SIZE];
 	mutex_t*	mutex;
 	rw_lock_t*	lock;
+	ulint		block_mutex_oswait_count = 0;
+	ulint		block_lock_oswait_count = 0;
+	mutex_t*	block_mutex = NULL;
+	rw_lock_t*	block_lock = NULL;
 #ifdef UNIV_DEBUG
 	ulint	  rw_lock_count= 0;
 	ulint	  rw_lock_count_spin_loop= 0;
@@ -8686,12 +8692,16 @@ innodb_mutex_show_status(
 
 	mutex_enter(&mutex_list_mutex);
 
-	mutex = UT_LIST_GET_FIRST(mutex_list);
+	for (mutex = UT_LIST_GET_FIRST(mutex_list); mutex != NULL;
+	     mutex = UT_LIST_GET_NEXT(list, mutex)) {
+		if (mutex->count_os_wait == 0) {
+			continue;
+		}
 
-	while (mutex != NULL) {
-		if (mutex->count_os_wait == 0
-		    || buf_pool_is_block_mutex(mutex)) {
-			goto next_mutex;
+		if (buf_pool_is_block_mutex(mutex)) {
+			block_mutex = mutex;
+			block_mutex_oswait_count += mutex->count_os_wait;
+			continue;
 		}
 #ifdef UNIV_DEBUG
 		if (mutex->mutex_type != 1) {
@@ -8718,8 +8728,7 @@ innodb_mutex_show_status(
 					DBUG_RETURN(1);
 				}
 			}
-		}
-		else {
+		} else {
 			rw_lock_count += mutex->count_using;
 			rw_lock_count_spin_loop += mutex->count_spin_loop;
 			rw_lock_count_spin_rounds += mutex->count_spin_rounds;
@@ -8731,7 +8740,7 @@ innodb_mutex_show_status(
 		buf1len= (uint) my_snprintf(buf1, sizeof(buf1), "%s:%lu",
 				     mutex->cfile_name, (ulong) mutex->cline);
 		buf2len= (uint) my_snprintf(buf2, sizeof(buf2), "os_waits=%lu",
-				     mutex->count_os_wait);
+				     (ulong) mutex->count_os_wait);
 
 		if (stat_print(thd, innobase_hton_name,
 			       hton_name_len, buf1, buf1len,
@@ -8740,45 +8749,83 @@ innodb_mutex_show_status(
 			DBUG_RETURN(1);
 		}
 #endif /* UNIV_DEBUG */
+	}
 
-next_mutex:
-		mutex = UT_LIST_GET_NEXT(list, mutex);
+	if (block_mutex) {
+		buf1len = (uint) my_snprintf(buf1, sizeof buf1,
+					     "combined %s:%lu",
+					     block_mutex->cfile_name,
+					     (ulong) block_mutex->cline);
+		buf2len = (uint) my_snprintf(buf2, sizeof buf2,
+					     "os_waits=%lu",
+					     (ulong) block_mutex_oswait_count);
+
+		if (stat_print(thd, innobase_hton_name,
+			       hton_name_len, buf1, buf1len,
+			       buf2, buf2len)) {
+			mutex_exit(&mutex_list_mutex);
+			DBUG_RETURN(1);
+		}
 	}
 
 	mutex_exit(&mutex_list_mutex);
 
 	mutex_enter(&rw_lock_list_mutex);
 
-	lock = UT_LIST_GET_FIRST(rw_lock_list);
-
-	while (lock != NULL) {
-		if (lock->count_os_wait
-		    && !buf_pool_is_block_lock(lock)) {
-			buf1len= my_snprintf(buf1, sizeof(buf1), "%s:%lu",
-                                    lock->cfile_name, (ulong) lock->cline);
-			buf2len= my_snprintf(buf2, sizeof(buf2),
-                                    "os_waits=%lu", lock->count_os_wait);
-
-			if (stat_print(thd, innobase_hton_name,
-				       hton_name_len, buf1, buf1len,
-				       buf2, buf2len)) {
-				mutex_exit(&rw_lock_list_mutex);
-				DBUG_RETURN(1);
-			}
+	for (lock = UT_LIST_GET_FIRST(rw_lock_list); lock != NULL;
+	     lock = UT_LIST_GET_NEXT(list, lock)) {
+		if (lock->count_os_wait) {
+			continue;
+		}
+
+		if (buf_pool_is_block_lock(lock)) {
+			block_lock = lock;
+			block_lock_oswait_count += lock->count_os_wait;
+			continue;
+		}
+
+		buf1len = my_snprintf(buf1, sizeof buf1, "%s:%lu",
+				     lock->cfile_name, (ulong) lock->cline);
+		buf2len = my_snprintf(buf2, sizeof buf2, "os_waits=%lu",
+				      (ulong) lock->count_os_wait);
+
+		if (stat_print(thd, innobase_hton_name,
+			       hton_name_len, buf1, buf1len,
+			       buf2, buf2len)) {
+			mutex_exit(&rw_lock_list_mutex);
+			DBUG_RETURN(1);
+		}
+	}
+
+	if (block_lock) {
+		buf1len = (uint) my_snprintf(buf1, sizeof buf1,
+					     "combined %s:%lu",
+					     block_lock->cfile_name,
+					     (ulong) block_lock->cline);
+		buf2len = (uint) my_snprintf(buf2, sizeof buf2,
+					     "os_waits=%lu",
+					     (ulong) block_lock_oswait_count);
+
+		if (stat_print(thd, innobase_hton_name,
+			       hton_name_len, buf1, buf1len,
+			       buf2, buf2len)) {
+			mutex_exit(&rw_lock_list_mutex);
+			DBUG_RETURN(1);
 		}
-		lock = UT_LIST_GET_NEXT(list, lock);
 	}
 
 	mutex_exit(&rw_lock_list_mutex);
 
 #ifdef UNIV_DEBUG
-	buf2len= my_snprintf(buf2, sizeof(buf2),
-		"count=%lu, spin_waits=%lu, spin_rounds=%lu, "
-		"os_waits=%lu, os_yields=%lu, os_wait_times=%lu",
-		rw_lock_count, rw_lock_count_spin_loop,
-		rw_lock_count_spin_rounds,
-		rw_lock_count_os_wait, rw_lock_count_os_yield,
-		(ulong) (rw_lock_wait_time/1000));
+	buf2len = my_snprintf(buf2, sizeof buf2,
+			     "count=%lu, spin_waits=%lu, spin_rounds=%lu, "
+			     "os_waits=%lu, os_yields=%lu, os_wait_times=%lu",
+			      (ulong) rw_lock_count,
+			      (ulong) rw_lock_count_spin_loop,
+			      (ulong) rw_lock_count_spin_rounds,
+			      (ulong) rw_lock_count_os_wait,
+			      (ulong) rw_lock_count_os_yield,
+			      (ulong) (rw_lock_wait_time / 1000));
 
 	if (stat_print(thd, innobase_hton_name, hton_name_len,
 			STRING_WITH_LEN("rw_lock_mutexes"), buf2, buf2len)) {

From b3370e53407463e1b69fdc99f520c0c10684254d Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:07:16 +0400
Subject: [PATCH 087/116] Applying InnoDB snapshot

Detailed revision comments:

r6782 | marko | 2010-03-09 14:09:26 +0200 (Tue, 09 Mar 2010) | 1 line
branches/zip: fil0fil.c: Update comments on table->flags as of r6252.
---
 storage/innodb_plugin/fil/fil0fil.c | 19 ++++++++++++-------
 1 file changed, 12 insertions(+), 7 deletions(-)

diff --git a/storage/innodb_plugin/fil/fil0fil.c b/storage/innodb_plugin/fil/fil0fil.c
index 16353959cb5..29c9c2e5b6d 100644
--- a/storage/innodb_plugin/fil/fil0fil.c
+++ b/storage/innodb_plugin/fil/fil0fil.c
@@ -1098,9 +1098,11 @@ fil_space_create(
 	fil_space_t*	space;
 
 	/* The tablespace flags (FSP_SPACE_FLAGS) should be 0 for
-	ROW_FORMAT=COMPACT (table->flags == DICT_TF_COMPACT) and
+	ROW_FORMAT=COMPACT
+	((table->flags & ~(~0 << DICT_TF_BITS)) == DICT_TF_COMPACT) and
 	ROW_FORMAT=REDUNDANT (table->flags == 0).  For any other
-	format, the tablespace flags should equal table->flags. */
+	format, the tablespace flags should equal
+	(table->flags & ~(~0 << DICT_TF_BITS)). */
 	ut_a(flags != DICT_TF_COMPACT);
 	ut_a(!(flags & (~0UL << DICT_TF_BITS)));
 
@@ -2584,9 +2586,11 @@ fil_create_new_single_table_tablespace(
 
 	ut_a(size >= FIL_IBD_FILE_INITIAL_SIZE);
 	/* The tablespace flags (FSP_SPACE_FLAGS) should be 0 for
-	ROW_FORMAT=COMPACT (table->flags == DICT_TF_COMPACT) and
+	ROW_FORMAT=COMPACT
+	((table->flags & ~(~0 << DICT_TF_BITS)) == DICT_TF_COMPACT) and
 	ROW_FORMAT=REDUNDANT (table->flags == 0).  For any other
-	format, the tablespace flags should equal table->flags. */
+	format, the tablespace flags should equal
+	(table->flags & ~(~0 << DICT_TF_BITS)). */
 	ut_a(flags != DICT_TF_COMPACT);
 	ut_a(!(flags & (~0UL << DICT_TF_BITS)));
 
@@ -2969,10 +2973,11 @@ fil_open_single_table_tablespace(
 	filepath = fil_make_ibd_name(name, FALSE);
 
 	/* The tablespace flags (FSP_SPACE_FLAGS) should be 0 for
-	ROW_FORMAT=COMPACT (table->flags == DICT_TF_COMPACT) and
+	ROW_FORMAT=COMPACT
+	((table->flags & ~(~0 << DICT_TF_BITS)) == DICT_TF_COMPACT) and
 	ROW_FORMAT=REDUNDANT (table->flags == 0).  For any other
-	format, the tablespace flags should be equal to
-	table->flags & ~(~0 << DICT_TF_BITS). */
+	format, the tablespace flags should equal
+	(table->flags & ~(~0 << DICT_TF_BITS)). */
 	ut_a(flags != DICT_TF_COMPACT);
 	ut_a(!(flags & (~0UL << DICT_TF_BITS)));
 

From d10b72dbd6af48dc15d5a18b8b55eec39f8d89b3 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:07:40 +0400
Subject: [PATCH 088/116] Applying InnoDB snapshot

Detailed revision comments:

r6786 | vasil | 2010-03-10 09:16:50 +0200 (Wed, 10 Mar 2010) | 4 lines
branches/zip:

Fix typo in comment
---
 mysql-test/suite/innodb/t/innodb-consistent.test | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/mysql-test/suite/innodb/t/innodb-consistent.test b/mysql-test/suite/innodb/t/innodb-consistent.test
index b58d0cb0e62..25bdf78e96b 100644
--- a/mysql-test/suite/innodb/t/innodb-consistent.test
+++ b/mysql-test/suite/innodb/t/innodb-consistent.test
@@ -26,7 +26,7 @@ replace into t1 select * from t2;
 connection b;
 set session transaction isolation level read committed;
 set autocommit=0;
-# should not cuase a lock wait.
+# should not cause a lock wait.
 delete from t2 where a=5;
 commit;
 delete from t2;
@@ -42,7 +42,7 @@ insert into t1 select * from t2;
 connection b;
 set session transaction isolation level read committed;
 set autocommit=0;
-# should not cuase a lock wait.
+# should not cause a lock wait.
 delete from t2 where a=5;
 commit;
 delete from t2;

From 434cf1894cab0ae2a2b8a2cf6d452b60b9033338 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:07:58 +0400
Subject: [PATCH 089/116] Applying InnoDB snapshot

Detailed revision comments:

r6787 | marko | 2010-03-10 10:35:06 +0200 (Wed, 10 Mar 2010) | 10 lines
branches/zip: recv_parse_log_rec(): Remove a bogus assertion about page_no.
TODO: We might also consider removing recv_max_parsed_page_no, because
it does not make much sense with *.ibd files.

recv_report_corrupt_log(), recv_scan_log_recs(): Abort when a
corrupted log record has been found, unless innodb_force_recovery has
been set.

This fixes Issue #464.
rb://265 approved by Heikki Tuuri
---
 storage/innodb_plugin/ChangeLog      |  7 +++++++
 storage/innodb_plugin/log/log0recv.c | 27 ++++++++++++++++++---------
 2 files changed, 25 insertions(+), 9 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index d7d1e89c147..57ad1208de0 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,10 @@
+2010-03-10	The InnoDB Team
+
+	* log/log0recv.c:
+	Remove a bogus assertion about page numbers exceeding 0x90000000
+	in the redo log.  Abort when encountering a corrupted redo log
+	record, unless innodb_force_recovery is set.
+
 2010-03-09	The InnoDB Team
 
 	* handler/ha_innodb.cc:
diff --git a/storage/innodb_plugin/log/log0recv.c b/storage/innodb_plugin/log/log0recv.c
index de026cad7b7..e5a67f28de6 100644
--- a/storage/innodb_plugin/log/log0recv.c
+++ b/storage/innodb_plugin/log/log0recv.c
@@ -2052,15 +2052,6 @@ recv_parse_log_rec(
 	}
 #endif /* UNIV_LOG_LSN_DEBUG */
 
-	/* Check that page_no is sensible */
-
-	if (UNIV_UNLIKELY(*page_no > 0x8FFFFFFFUL)) {
-
-		recv_sys->found_corrupt_log = TRUE;
-
-		return(0);
-	}
-
 	new_ptr = recv_parse_or_apply_log_rec_body(*type, new_ptr, end_ptr,
 						   NULL, NULL);
 	if (UNIV_UNLIKELY(new_ptr == NULL)) {
@@ -2169,6 +2160,14 @@ recv_report_corrupt_log(
 		putc('\n', stderr);
 	}
 
+#ifndef UNIV_HOTBACKUP
+	if (!srv_force_recovery) {
+		fputs("InnoDB: Set innodb_force_recovery"
+		      " to ignore this error.\n", stderr);
+		ut_error;
+	}
+#endif /* !UNIV_HOTBACKUP */
+
 	fputs("InnoDB: WARNING: the log file may have been corrupt and it\n"
 	      "InnoDB: is possible that the log scan did not proceed\n"
 	      "InnoDB: far enough in recovery! Please run CHECK TABLE\n"
@@ -2683,6 +2682,16 @@ recv_scan_log_recs(
 
 				recv_sys->found_corrupt_log = TRUE;
 
+#ifndef UNIV_HOTBACKUP
+				if (!srv_force_recovery) {
+					fputs("InnoDB: Set"
+					      " innodb_force_recovery"
+					      " to ignore this error.\n",
+					      stderr);
+					ut_error;
+				}
+#endif /* !UNIV_HOTBACKUP */
+
 			} else if (!recv_sys->found_corrupt_log) {
 				more_data = recv_sys_add_to_parsing_buf(
 					log_block, scanned_lsn);

From 178dccaae5fe4185e045e2e287cb88ddec38454b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:12:44 +0400
Subject: [PATCH 090/116] Applying InnoDB snapshot, fixes BUG#51378.

Detailed revision comments:

r6789 | jyang | 2010-03-10 11:18:18 +0200 (Wed, 10 Mar 2010) | 10 lines
branches/zip: If a unique index is on a column prefix, such
unique index cannot be upgrade to primary index even if there
is no primary index already defined. Also fix possible corruption
when initialize "ref_length" value in case there is a mismatch
between MySQL and InnoDB primary key. Fix bug #51378: "Init
'ref_length'  to correct value, in case an out of bound MySQL
primary_key".
rb://262 approved by Marko.
---
 storage/innodb_plugin/handler/ha_innodb.cc    | 99 ++++++++++++++-----
 .../innodb_plugin/handler/handler0alter.cc    |  8 +-
 2 files changed, 82 insertions(+), 25 deletions(-)

diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 86587fee503..f5627f35a45 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -3304,6 +3304,9 @@ ha_innobase::innobase_initialize_autoinc()
 		ulint		err;
 
 		update_thd(ha_thd());
+
+		ut_a(prebuilt->trx == thd_to_trx(user_thd));
+
 		col_name = field->field_name;
 		index = innobase_get_index(table->s->next_number_index);
 
@@ -3494,31 +3497,86 @@ retry:
 	of length ref_length! */
 
 	if (!row_table_got_default_clust_index(ib_table)) {
-		if (primary_key >= MAX_KEY) {
-		  sql_print_error("Table %s has a primary key in InnoDB data "
-				  "dictionary, but not in MySQL!", name);
-		}
 
 		prebuilt->clust_index_was_generated = FALSE;
 
-		/* MySQL allocates the buffer for ref. key_info->key_length
-		includes space for all key columns + one byte for each column
-		that may be NULL. ref_length must be as exact as possible to
-		save space, because all row reference buffers are allocated
-		based on ref_length. */
+		if (UNIV_UNLIKELY(primary_key >= MAX_KEY)) {
+			sql_print_error("Table %s has a primary key in "
+					"InnoDB data dictionary, but not "
+					"in MySQL!", name);
 
-		ref_length = table->key_info[primary_key].key_length;
+			/* This mismatch could cause further problems
+			if not attended, bring this to the user's attention
+			by printing a warning in addition to log a message
+			in the errorlog */
+			push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
+					    ER_NO_SUCH_INDEX,
+					    "InnoDB: Table %s has a "
+					    "primary key in InnoDB data "
+					    "dictionary, but not in "
+					    "MySQL!", name);
+
+			/* If primary_key >= MAX_KEY, its (primary_key)
+			value could be out of bound if continue to index
+			into key_info[] array. Find InnoDB primary index,
+			and assign its key_length to ref_length.
+			In addition, since MySQL indexes are sorted starting
+			with primary index, unique index etc., initialize
+			ref_length to the first index key length in
+			case we fail to find InnoDB cluster index.
+
+			Please note, this will not resolve the primary
+			index mismatch problem, other side effects are
+			possible if users continue to use the table.
+			However, we allow this table to be opened so
+			that user can adopt necessary measures for the
+			mismatch while still being accessible to the table
+			date. */
+			ref_length = table->key_info[0].key_length;
+
+			/* Find correspoinding cluster index
+			key length in MySQL's key_info[] array */
+			for (ulint i = 0; i < table->s->keys; i++) {
+				dict_index_t*	index;
+				index = innobase_get_index(i);
+				if (dict_index_is_clust(index)) {
+					ref_length =
+						 table->key_info[i].key_length;
+				}
+			}
+		} else {
+			/* MySQL allocates the buffer for ref.
+			key_info->key_length includes space for all key
+			columns + one byte for each column that may be
+			NULL. ref_length must be as exact as possible to
+			save space, because all row reference buffers are
+			allocated based on ref_length. */
+
+			ref_length = table->key_info[primary_key].key_length;
+		}
 	} else {
 		if (primary_key != MAX_KEY) {
-		  sql_print_error("Table %s has no primary key in InnoDB data "
-				  "dictionary, but has one in MySQL! If you "
-				  "created the table with a MySQL version < "
-				  "3.23.54 and did not define a primary key, "
-				  "but defined a unique key with all non-NULL "
-				  "columns, then MySQL internally treats that "
-				  "key as the primary key. You can fix this "
-				  "error by dump + DROP + CREATE + reimport "
-				  "of the table.", name);
+			sql_print_error(
+				"Table %s has no primary key in InnoDB data "
+				"dictionary, but has one in MySQL! If you "
+				"created the table with a MySQL version < "
+				"3.23.54 and did not define a primary key, "
+				"but defined a unique key with all non-NULL "
+				"columns, then MySQL internally treats that "
+				"key as the primary key. You can fix this "
+				"error by dump + DROP + CREATE + reimport "
+				"of the table.", name);
+
+			/* This mismatch could cause further problems
+			if not attended, bring this to the user attention
+			by printing a warning in addition to log a message
+			in the errorlog */
+			push_warning_printf(thd, MYSQL_ERROR::WARN_LEVEL_WARN,
+					    ER_NO_SUCH_INDEX,
+					    "InnoDB: Table %s has no "
+					    "primary key in InnoDB data "
+					    "dictionary, but has one in "
+					    "MySQL!", name);
 		}
 
 		prebuilt->clust_index_was_generated = TRUE;
@@ -5402,9 +5460,6 @@ ha_innobase::innobase_get_index(
 	DBUG_ENTER("innobase_get_index");
 	ha_statistic_increment(&SSV::ha_read_key_count);
 
-	ut_ad(user_thd == ha_thd());
-	ut_a(prebuilt->trx == thd_to_trx(user_thd));
-
 	if (keynr != MAX_KEY && table->s->keys > 0) {
 		key = table->key_info + keynr;
 
diff --git a/storage/innodb_plugin/handler/handler0alter.cc b/storage/innodb_plugin/handler/handler0alter.cc
index 1c99cd8e984..071253d2dae 100644
--- a/storage/innodb_plugin/handler/handler0alter.cc
+++ b/storage/innodb_plugin/handler/handler0alter.cc
@@ -528,12 +528,14 @@ innobase_create_key_def(
 				     key_info->name, "PRIMARY");
 
 	/* If there is a UNIQUE INDEX consisting entirely of NOT NULL
-	columns, MySQL will treat it as a PRIMARY KEY unless the
-	table already has one. */
+	columns and if the index does not contain column prefix(es)
+	(only prefix/part of the column is indexed), MySQL will treat the
+	index as a PRIMARY KEY unless the table already has one. */
 
 	if (!new_primary && (key_info->flags & HA_NOSAME)
+	    && (!(key_info->flags & HA_KEY_HAS_PART_KEY_SEG))
 	    && row_table_got_default_clust_index(table)) {
-		uint	key_part = key_info->key_parts;
+		uint    key_part = key_info->key_parts;
 
 		new_primary = TRUE;
 

From aacb485899f22e5b8a7a61d7fd396e7f98dc9d34 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:14:51 +0400
Subject: [PATCH 091/116] Applying InnoDB snapshot, fixes BUG#51356.

Detailed revision comments:

r6790 | jyang | 2010-03-10 13:09:41 +0200 (Wed, 10 Mar 2010) | 7 lines
branches/zip: Fix bug #51356: "many valgrind errors in error messages
with concurrent ddl". Null terminate the name string returned
from innobase_convert_identifier() call when reporting DB_DUPLICATE_KEY
error in create_table_def().
rb://266 approved by Marko
---
 .../suite/innodb/r/innodb_bug51378.result     | 66 ++++++++++++++++
 .../suite/innodb/t/innodb_bug51378.test       | 77 +++++++++++++++++++
 storage/innodb_plugin/handler/ha_innodb.cc    |  8 +-
 3 files changed, 148 insertions(+), 3 deletions(-)
 create mode 100644 mysql-test/suite/innodb/r/innodb_bug51378.result
 create mode 100644 mysql-test/suite/innodb/t/innodb_bug51378.test

diff --git a/mysql-test/suite/innodb/r/innodb_bug51378.result b/mysql-test/suite/innodb/r/innodb_bug51378.result
new file mode 100644
index 00000000000..a3ca73c16a9
--- /dev/null
+++ b/mysql-test/suite/innodb/r/innodb_bug51378.result
@@ -0,0 +1,66 @@
+create table bug51378 (
+col1 int not null,
+col2 blob not null,
+col3 time not null) engine = innodb;
+create unique index idx on bug51378(col1, col2(31));
+alter table bug51378 add unique index idx2(col1, col2(31));
+create unique index idx3 on bug51378(col1, col3);
+SHOW CREATE TABLE bug51378;
+Table	Create Table
+bug51378	CREATE TABLE `bug51378` (
+  `col1` int(11) NOT NULL,
+  `col2` blob NOT NULL,
+  `col3` time NOT NULL,
+  UNIQUE KEY `idx3` (`col1`,`col3`),
+  UNIQUE KEY `idx` (`col1`,`col2`(31)),
+  UNIQUE KEY `idx2` (`col1`,`col2`(31))
+) ENGINE=InnoDB DEFAULT CHARSET=latin1
+drop index idx3 on bug51378;
+SHOW CREATE TABLE bug51378;
+Table	Create Table
+bug51378	CREATE TABLE `bug51378` (
+  `col1` int(11) NOT NULL,
+  `col2` blob NOT NULL,
+  `col3` time NOT NULL,
+  UNIQUE KEY `idx` (`col1`,`col2`(31)),
+  UNIQUE KEY `idx2` (`col1`,`col2`(31))
+) ENGINE=InnoDB DEFAULT CHARSET=latin1
+alter table bug51378 add primary key idx3(col1, col2(31));
+SHOW CREATE TABLE bug51378;
+Table	Create Table
+bug51378	CREATE TABLE `bug51378` (
+  `col1` int(11) NOT NULL,
+  `col2` blob NOT NULL,
+  `col3` time NOT NULL,
+  PRIMARY KEY (`col1`,`col2`(31)),
+  UNIQUE KEY `idx` (`col1`,`col2`(31)),
+  UNIQUE KEY `idx2` (`col1`,`col2`(31))
+) ENGINE=InnoDB DEFAULT CHARSET=latin1
+drop table bug51378;
+create table bug51378 (
+col1 int not null,
+col2 blob not null,
+col3 time not null, primary key(col1, col2(31))) engine = innodb;
+create unique index idx on bug51378(col1, col2(31));
+SHOW CREATE TABLE bug51378;
+Table	Create Table
+bug51378	CREATE TABLE `bug51378` (
+  `col1` int(11) NOT NULL,
+  `col2` blob NOT NULL,
+  `col3` time NOT NULL,
+  PRIMARY KEY (`col1`,`col2`(31)),
+  UNIQUE KEY `idx` (`col1`,`col2`(31))
+) ENGINE=InnoDB DEFAULT CHARSET=latin1
+drop table bug51378;
+create table bug51378 (
+col1 int not null,
+col2 int ) engine = innodb;
+create unique index idx on bug51378(col1, col2);
+SHOW CREATE TABLE bug51378;
+Table	Create Table
+bug51378	CREATE TABLE `bug51378` (
+  `col1` int(11) NOT NULL,
+  `col2` int(11) DEFAULT NULL,
+  UNIQUE KEY `idx` (`col1`,`col2`)
+) ENGINE=InnoDB DEFAULT CHARSET=latin1
+drop table bug51378;
diff --git a/mysql-test/suite/innodb/t/innodb_bug51378.test b/mysql-test/suite/innodb/t/innodb_bug51378.test
new file mode 100644
index 00000000000..8f7b0b9605a
--- /dev/null
+++ b/mysql-test/suite/innodb/t/innodb_bug51378.test
@@ -0,0 +1,77 @@
+# This is the test for bug 51378. Unique index created
+# through "create index" and "alter table add unique index"
+# interfaces should not be treated as primary index if indexed
+# columns contain one or more column prefix(es) (only prefix/part of
+# the column is indexed)
+# On the other hand, if there is a unique index covers all
+# columns of a table, and they are non-null columns, and
+# full length of the column are indexed, then this index
+# will be created as primary index
+# Following queries test various scenario, no mismatch
+# error message should be printed.
+--source include/have_innodb.inc
+
+# Create a table contains a BLOB column
+create table bug51378 (
+	col1 int not null,
+	col2 blob not null,
+	col3 time not null) engine = innodb;
+
+# Create following unique indexes on 'col1' and 'col2(31)'
+# of the table, the index should not be treated as primary
+# key because it indexes only first 31 bytes of col2.
+# Thus it contains "column prefix", and will not be
+# upgraded to primary index.
+# There should not be mismatch message printed in the
+# errorlog
+create unique index idx on bug51378(col1, col2(31));
+
+alter table bug51378 add unique index idx2(col1, col2(31));
+
+# Unique index on 'col1' and 'col3' will be created as primary index,
+# since the index does not contain column prefix
+create unique index idx3 on bug51378(col1, col3);
+
+# Show create table would show idx3 created as unique index, internally,
+# idx3 is treated as primary index both by MySQL and Innodb
+SHOW CREATE TABLE bug51378;
+
+# "GEN_CLUST_INDEX" will be re-created as default primary index
+# after idx3 is dropped
+drop index idx3 on bug51378;
+
+SHOW CREATE TABLE bug51378;
+
+# Or we can add the primary key through alter table interfaces
+alter table bug51378 add primary key idx3(col1, col2(31));
+
+SHOW CREATE TABLE bug51378;
+
+drop table bug51378;
+
+# Or we can create such primary key through create table interfaces
+create table bug51378 (
+        col1 int not null,
+        col2 blob not null,
+        col3 time not null, primary key(col1, col2(31))) engine = innodb;
+
+# Unique index on one or more column prefix(es) will be created
+# as non-cluster index
+create unique index idx on bug51378(col1, col2(31));
+
+SHOW CREATE TABLE bug51378;
+
+drop table bug51378;
+
+# If a table has a NULLABLE column, unique index on it will not
+# be treated as primary index.
+create table bug51378 (
+	col1 int not null,
+        col2 int ) engine = innodb;
+
+# This will be created as non-cluster index since col2 is nullable
+create unique index idx on bug51378(col1, col2);
+
+SHOW CREATE TABLE bug51378;
+
+drop table bug51378;
diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index f5627f35a45..74a67c2893b 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -6045,9 +6045,11 @@ create_table_def(
 
 	if (error == DB_DUPLICATE_KEY) {
 		char buf[100];
-		innobase_convert_identifier(buf, sizeof buf,
-					    table_name, strlen(table_name),
-					    trx->mysql_thd, TRUE);
+		char* buf_end = innobase_convert_identifier(
+			buf, sizeof buf - 1, table_name, strlen(table_name),
+			trx->mysql_thd, TRUE);
+
+		*buf_end = '\0';
 		my_error(ER_TABLE_EXISTS_ERROR, MYF(0), buf);
 	}
 

From 36f1a009f80ee3343489c5c80a7b186ce531125d Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:15:15 +0400
Subject: [PATCH 092/116] Applying InnoDB snapshot

Detailed revision comments:

r6791 | marko | 2010-03-10 13:39:06 +0200 (Wed, 10 Mar 2010) | 1 line
branches/zip: Add ChangeLog entries for r6789, r6790.
---
 storage/innodb_plugin/ChangeLog | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 57ad1208de0..f0230a64f3c 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,16 @@
+2010-03-10	The InnoDB Team
+
+	* handler/ha_innodb.cc:
+	Fix Bug #51356 Many Valgrind errors in error messages
+	with concurrent DDL
+
+2010-03-10	The InnoDB Team
+
+	* handler/ha_innodb.cc, handler/handler0alter.cc,
+	mysql-test/innodb_bug51378.result, mysql-test/innodb_bug51378.test:
+	Fix Bug #51378 Init 'ref_length' to correct value, in case an out
+	of bound MySQL primary_key
+
 2010-03-10	The InnoDB Team
 
 	* log/log0recv.c:

From 1ce8091b22b0261f45a98fa42830266265b80333 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:20:37 +0400
Subject: [PATCH 093/116] Applying InnoDB snapshot

Detailed revision comments:

r6792 | marko | 2010-03-10 13:56:41 +0200 (Wed, 10 Mar 2010) | 1 line
branches/zip: Copy tests from branches/5.1 that were lost in some merge.
r6793 | marko | 2010-03-10 14:02:19 +0200 (Wed, 10 Mar 2010) | 60 lines
branches/zip: Merge revisions 6669:6788 from branches/5.1:

  ------------------------------------------------------------------------
  r6774 | calvin | 2010-03-03 23:56:10 +0200 (Wed, 03 Mar 2010) | 2 lines
  Changed paths:
     M /branches/5.1/trx/trx0sys.c

  branches/5.1: fix bug#51653: outdated reference to set-variable
  Non functional change.
  ------------------------------------------------------------------------
  r6780 | vasil | 2010-03-08 19:13:20 +0200 (Mon, 08 Mar 2010) | 4 lines
  Changed paths:
     M /branches/5.1/plug.in

  branches/5.1:

  Whitespace fixup.
  ------------------------------------------------------------------------
  r6783 | jyang | 2010-03-09 17:54:14 +0200 (Tue, 09 Mar 2010) | 9 lines
  Changed paths:
     M /branches/5.1/handler/ha_innodb.cc
     M /branches/5.1/mysql-test/innodb_bug21704.result
     A /branches/5.1/mysql-test/innodb_bug47621.result
     A /branches/5.1/mysql-test/innodb_bug47621.test

  branches/5.1: Fix bug #47621 "MySQL and InnoDB data dictionaries
  will become out of sync when renaming columns". MySQL does not
  provide new column name information to storage engine to
  update the system table. To avoid column name mismatch, we shall
  just request a table copy for now.

  rb://246 approved by Marko.
  ------------------------------------------------------------------------
  r6785 | vasil | 2010-03-10 09:04:38 +0200 (Wed, 10 Mar 2010) | 11 lines
  Changed paths:
     M /branches/5.1/mysql-test/innodb_bug38231.test

  branches/5.1:

  Add the missing --reap statements in innodb_bug38231.test. Probably MySQL
  enforced the presence of those recently and the test started failing like:

    main.innodb_bug38231                     [ fail ]
            Test ended at 2010-03-10 08:48:32

    CURRENT_TEST: main.innodb_bug38231
    mysqltest: At line 49: Cannot run query on connection between send and reap
  ------------------------------------------------------------------------
  r6788 | vasil | 2010-03-10 10:53:21 +0200 (Wed, 10 Mar 2010) | 8 lines
  Changed paths:
     M /branches/5.1/mysql-test/innodb_bug38231.test

  branches/5.1:

  In innodb_bug38231.test: replace the fragile sleep 0.2 that depends on timing
  with a more robust condition which waits for the TRUNCATE and LOCK commands
  to appear in information_schema.processlist. This could also break if there
  are other sessions executing the same SQL commands, but there are none during
  the execution of the mysql test.
  ------------------------------------------------------------------------
---
 storage/innodb_plugin/ChangeLog            | 12 ++++
 storage/innodb_plugin/handler/ha_innodb.cc | 72 ++++++++++++++++------
 storage/innodb_plugin/trx/trx0sys.c        |  4 +-
 3 files changed, 68 insertions(+), 20 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index f0230a64f3c..324d0704a30 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,15 @@
+2010-03-10	The InnoDB Team
+
+	* trx/trx0sys.c:
+	Fix Bug #51653 outdated reference to set-variable
+
+2010-03-10	The InnoDB Team
+
+	* handler/ha_innodb.cc, mysql-test/innodb_bug21704.result,
+	mysql-test/innodb_bug47621.result, mysql-test/innodb_bug47621.test:
+	Fix Bug #47621 MySQL and InnoDB data dictionaries will become
+	out of sync when renaming columns
+
 2010-03-10	The InnoDB Team
 
 	* handler/ha_innodb.cc:
diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 74a67c2893b..2f61e164e42 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -9833,33 +9833,60 @@ innobase_set_cursor_view(
 				  (cursor_view_t*) curview);
 }
 
+/*******************************************************************//**
+If col_name is not NULL, check whether the named column is being
+renamed in the table. If col_name is not provided, check
+whether any one of columns in the table is being renamed.
+@return true if the column is being renamed */
+static
+bool
+check_column_being_renamed(
+/*=======================*/
+	const TABLE*	table,		/*!< in: MySQL table */
+	const char*	col_name)	/*!< in: name of the column */
+{
+	uint		k;
+	Field*		field;
 
-/***********************************************************************
-Check whether any of the given columns is being renamed in the table. */
+	for (k = 0; k < table->s->fields; k++) {
+		field = table->field[k];
+
+		if (field->flags & FIELD_IS_RENAMED) {
+
+			/* If col_name is not provided, return
+			if the field is marked as being renamed. */
+			if (!col_name) {
+				return(true);
+			}
+
+			/* If col_name is provided, return only
+			if names match */
+			if (innobase_strcasecmp(field->field_name,
+						col_name) == 0) {
+				return(true);
+			}
+		}
+	}
+
+	return(false);
+}
+
+/*******************************************************************//**
+Check whether any of the given columns is being renamed in the table.
+@return true if any of col_names is being renamed in table */
 static
 bool
 column_is_being_renamed(
 /*====================*/
-					/* out: true if any of col_names is
-					being renamed in table */
-	TABLE*		table,		/* in: MySQL table */
-	uint		n_cols,		/* in: number of columns */
-	const char**	col_names)	/* in: names of the columns */
+	TABLE*		table,		/*!< in: MySQL table */
+	uint		n_cols,		/*!< in: number of columns */
+	const char**	col_names)	/*!< in: names of the columns */
 {
 	uint		j;
-	uint		k;
-	Field*		field;
-	const char*	col_name;
 
 	for (j = 0; j < n_cols; j++) {
-		col_name = col_names[j];
-		for (k = 0; k < table->s->fields; k++) {
-			field = table->field[k];
-			if ((field->flags & FIELD_IS_RENAMED)
-			    && innobase_strcasecmp(field->field_name,
-						   col_name) == 0) {
-				return(true);
-			}
+		if (check_column_being_renamed(table, col_names[j])) {
+			return(true);
 		}
 	}
 
@@ -9943,6 +9970,15 @@ ha_innobase::check_if_incompatible_data(
 		return(COMPATIBLE_DATA_NO);
 	}
 
+	/* For column rename operation, MySQL does not supply enough
+	information (new column name etc.) for InnoDB to make appropriate
+	system metadata change. To avoid system metadata inconsistency,
+	currently we can just request a table rebuild/copy by returning
+	COMPATIBLE_DATA_NO */
+	if (check_column_being_renamed(table, NULL)) {
+		return COMPATIBLE_DATA_NO;
+	}
+
 	/* Check if a column participating in a foreign key is being renamed.
 	There is no mechanism for updating InnoDB foreign key definitions. */
 	if (foreign_key_column_is_being_renamed(prebuilt, table)) {
diff --git a/storage/innodb_plugin/trx/trx0sys.c b/storage/innodb_plugin/trx/trx0sys.c
index 253619545af..ba25662c8fb 100644
--- a/storage/innodb_plugin/trx/trx0sys.c
+++ b/storage/innodb_plugin/trx/trx0sys.c
@@ -584,8 +584,8 @@ trx_sys_doublewrite_init_or_restore_pages(
 						" recover the database"
 						" with the my.cnf\n"
 						"InnoDB: option:\n"
-						"InnoDB: set-variable="
-						"innodb_force_recovery=6\n");
+						"InnoDB:"
+						" innodb_force_recovery=6\n");
 					exit(1);
 				}
 

From 2d33096d870b57fab9f6fcb2bfef3d5a10e3624d Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:21:00 +0400
Subject: [PATCH 094/116] Applying InnoDB snapshot

Detailed revision comments:

r6798 | marko | 2010-03-11 09:53:01 +0200 (Thu, 11 Mar 2010) | 14 lines
branches/zip: Fix and clarify the latching of some buf_block_t members.

buf_block_t::check_index_page_at_flush: Note that this field is not
protected by any mutex. Make it a separate field, not a bitfield that
could share the machine word with other fields.

buf_block_t::lock_hash_val: Note that this field is protected by
buf_block_t::lock (or during block creation, by buf_pool_mutex and
buf_block_t::mutex).

buf_block_get_lock_hash_val(): Assert that block->lock is held by the
current thread.

Issue #465, rb://267 approved by Inaam Rana
---
 storage/innodb_plugin/ChangeLog          |  7 +++++++
 storage/innodb_plugin/include/buf0buf.h  | 12 +++++++++---
 storage/innodb_plugin/include/buf0buf.ic |  6 ++++++
 3 files changed, 22 insertions(+), 3 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 324d0704a30..a8b8c52908d 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,10 @@
+2010-03-11	The InnoDB Team
+
+	* buf0buf.h, buf0buf.ic:
+	Fix and clarify the latching of some buf_block_t members.
+	Note that check_index_page_at_flush is not protected by any mutex.
+	Note and assert that lock_hash_val is protected by the rw-latch.
+
 2010-03-10	The InnoDB Team
 
 	* trx/trx0sys.c:
diff --git a/storage/innodb_plugin/include/buf0buf.h b/storage/innodb_plugin/include/buf0buf.h
index 449f494f3c7..b5ee2bfdaf6 100644
--- a/storage/innodb_plugin/include/buf0buf.h
+++ b/storage/innodb_plugin/include/buf0buf.h
@@ -1179,15 +1179,21 @@ struct buf_block_struct{
 	rw_lock_t	lock;		/*!< read-write lock of the buffer
 					frame */
 	unsigned	lock_hash_val:32;/*!< hashed value of the page address
-					in the record lock hash table */
-	unsigned	check_index_page_at_flush:1;
+					in the record lock hash table;
+					protected by buf_block_t::lock
+					(or buf_block_t::mutex, buf_pool_mutex
+				        in buf_page_get_gen(),
+					buf_page_init_for_read()
+					and buf_page_create()) */
+	ibool		check_index_page_at_flush;
 					/*!< TRUE if we know that this is
 					an index page, and want the database
 					to check its consistency before flush;
 					note that there may be pages in the
 					buffer pool which are index pages,
 					but this flag is not set because
-					we do not keep track of all pages */
+					we do not keep track of all pages;
+					NOT protected by any mutex */
 	/* @} */
 	/** @name Optimistic search field */
 	/* @{ */
diff --git a/storage/innodb_plugin/include/buf0buf.ic b/storage/innodb_plugin/include/buf0buf.ic
index 0f92a59a1c7..e5cbde73ff6 100644
--- a/storage/innodb_plugin/include/buf0buf.ic
+++ b/storage/innodb_plugin/include/buf0buf.ic
@@ -705,6 +705,12 @@ buf_block_get_lock_hash_val(
 /*========================*/
 	const buf_block_t*	block)	/*!< in: block */
 {
+	ut_ad(block);
+	ut_ad(buf_page_in_file(&block->page));
+#ifdef UNIV_SYNC_DEBUG
+	ut_ad(rw_lock_own(&(((buf_block_t*) block)->lock), RW_LOCK_EXCLUSIVE)
+	      || rw_lock_own(&(((buf_block_t*) block)->lock), RW_LOCK_SHARED));
+#endif /* UNIV_SYNC_DEBUG */
 	return(block->lock_hash_val);
 }
 

From 07e86b50086c6334afb9bd7a9a29f60cbd2afec4 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:22:36 +0400
Subject: [PATCH 095/116] Applying InnoDB snapshot

Detailed revision comments:

r6799 | jyang | 2010-03-11 09:59:42 +0200 (Thu, 11 Mar 2010) | 5 lines
branches/zip: Once change in bug #47621 merges into zip branch,
zip only test innodb_bug44571 needs to be updated to reflect the
column name change would be successful be done in InnoDB as well.
---
 mysql-test/suite/innodb/r/innodb_bug44571.result | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/mysql-test/suite/innodb/r/innodb_bug44571.result b/mysql-test/suite/innodb/r/innodb_bug44571.result
index 36374edcb3e..7ee7820a02d 100644
--- a/mysql-test/suite/innodb/r/innodb_bug44571.result
+++ b/mysql-test/suite/innodb/r/innodb_bug44571.result
@@ -2,8 +2,7 @@ CREATE TABLE bug44571 (foo INT) ENGINE=InnoDB;
 ALTER TABLE bug44571 CHANGE foo bar INT;
 ALTER TABLE bug44571 ADD INDEX bug44571b (foo);
 ERROR 42000: Key column 'foo' doesn't exist in table
-ALTER TABLE bug44571 ADD INDEX bug44571b (bar);
-ERROR HY000: Incorrect key file for table 'bug44571'; try to repair it
-CREATE INDEX bug44571b ON bug44571 (bar);
-ERROR HY000: Incorrect key file for table 'bug44571'; try to repair it
+ALTER TABLE bug44571 ADD INDEX bug44571c (bar);
+DROP INDEX bug44571c ON bug44571;
+CREATE INDEX bug44571c ON bug44571 (bar);
 DROP TABLE bug44571;

From edba051bee0ec0f25aaab45d8e4f8e6999d24f33 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:27:13 +0400
Subject: [PATCH 096/116] Applying InnoDB snapshot

Detailed revision comments:

r6800 | marko | 2010-03-11 12:02:57 +0200 (Thu, 11 Mar 2010) | 1 line
branches/zip: Add ut_ad(mtr->state == MTR_ACTIVE) to various places.
---
 mysql-test/suite/innodb/t/innodb_bug44571.test | 17 +++++++++++------
 storage/innodb_plugin/btr/btr0pcur.c           |  3 +++
 storage/innodb_plugin/buf/buf0buf.c            | 10 +++++++++-
 storage/innodb_plugin/include/mtr0mtr.ic       |  3 +++
 4 files changed, 26 insertions(+), 7 deletions(-)

diff --git a/mysql-test/suite/innodb/t/innodb_bug44571.test b/mysql-test/suite/innodb/t/innodb_bug44571.test
index 43f290cde84..58f02dfb5d2 100644
--- a/mysql-test/suite/innodb/t/innodb_bug44571.test
+++ b/mysql-test/suite/innodb/t/innodb_bug44571.test
@@ -1,18 +1,23 @@
 #
 # Bug#44571 InnoDB Plugin crashes on ADD INDEX
 # http://bugs.mysql.com/44571
+# Please also refer to related fix in
+# http://bugs.mysql.com/47621
 #
 -- source include/have_innodb.inc
 -- source suite/innodb/include/have_innodb_plugin.inc
 
 CREATE TABLE bug44571 (foo INT) ENGINE=InnoDB;
 ALTER TABLE bug44571 CHANGE foo bar INT;
+# Create index with the old column name will fail,
+# because the CHANGE foo bar is successful. And
+# the column name change would communicate to
+# InnoDB with the fix from bug #47621
 -- error ER_KEY_COLUMN_DOES_NOT_EXITS
 ALTER TABLE bug44571 ADD INDEX bug44571b (foo);
-# The following will fail, because the CHANGE foo bar was
-# not communicated to InnoDB.
---error ER_NOT_KEYFILE
-ALTER TABLE bug44571 ADD INDEX bug44571b (bar);
---error ER_NOT_KEYFILE
-CREATE INDEX bug44571b ON bug44571 (bar);
+# The following create indexes should succeed,
+# indirectly confirm the CHANGE foo bar is successful.
+ALTER TABLE bug44571 ADD INDEX bug44571c (bar);
+DROP INDEX bug44571c ON bug44571;
+CREATE INDEX bug44571c ON bug44571 (bar);
 DROP TABLE bug44571;
diff --git a/storage/innodb_plugin/btr/btr0pcur.c b/storage/innodb_plugin/btr/btr0pcur.c
index 2edfd673530..658901208ef 100644
--- a/storage/innodb_plugin/btr/btr0pcur.c
+++ b/storage/innodb_plugin/btr/btr0pcur.c
@@ -219,6 +219,9 @@ btr_pcur_restore_position_func(
 	ulint		old_mode;
 	mem_heap_t*	heap;
 
+	ut_ad(mtr);
+	ut_ad(mtr->state == MTR_ACTIVE);
+
 	index = btr_cur_get_index(btr_pcur_get_btr_cur(cursor));
 
 	if (UNIV_UNLIKELY(cursor->old_stored != BTR_PCUR_OLD_STORED)
diff --git a/storage/innodb_plugin/buf/buf0buf.c b/storage/innodb_plugin/buf/buf0buf.c
index d5c14155f72..40f7925f520 100644
--- a/storage/innodb_plugin/buf/buf0buf.c
+++ b/storage/innodb_plugin/buf/buf0buf.c
@@ -2041,6 +2041,7 @@ buf_page_get_gen(
 	ulint		retries = 0;
 
 	ut_ad(mtr);
+	ut_ad(mtr->state == MTR_ACTIVE);
 	ut_ad((rw_latch == RW_S_LATCH)
 	      || (rw_latch == RW_X_LATCH)
 	      || (rw_latch == RW_NO_LATCH));
@@ -2397,7 +2398,9 @@ buf_page_optimistic_get(
 	ibool		success;
 	ulint		fix_type;
 
-	ut_ad(mtr && block);
+	ut_ad(block);
+	ut_ad(mtr);
+	ut_ad(mtr->state == MTR_ACTIVE);
 	ut_ad((rw_latch == RW_S_LATCH) || (rw_latch == RW_X_LATCH));
 
 	mutex_enter(&block->mutex);
@@ -2509,6 +2512,7 @@ buf_page_get_known_nowait(
 	ulint		fix_type;
 
 	ut_ad(mtr);
+	ut_ad(mtr->state == MTR_ACTIVE);
 	ut_ad((rw_latch == RW_S_LATCH) || (rw_latch == RW_X_LATCH));
 
 	mutex_enter(&block->mutex);
@@ -2608,6 +2612,9 @@ buf_page_try_get_func(
 	ibool		success;
 	ulint		fix_type;
 
+	ut_ad(mtr);
+	ut_ad(mtr->state == MTR_ACTIVE);
+
 	buf_pool_mutex_enter();
 	block = buf_block_hash_get(space_id, page_no);
 
@@ -2981,6 +2988,7 @@ buf_page_create(
 	ulint		time_ms		= ut_time_ms();
 
 	ut_ad(mtr);
+	ut_ad(mtr->state == MTR_ACTIVE);
 	ut_ad(space || !zip_size);
 
 	free_block = buf_LRU_get_free_block(0);
diff --git a/storage/innodb_plugin/include/mtr0mtr.ic b/storage/innodb_plugin/include/mtr0mtr.ic
index 310c7c4117f..fb6e1740198 100644
--- a/storage/innodb_plugin/include/mtr0mtr.ic
+++ b/storage/innodb_plugin/include/mtr0mtr.ic
@@ -70,6 +70,7 @@ mtr_memo_push(
 	ut_ad(type <= MTR_MEMO_X_LOCK);
 	ut_ad(mtr);
 	ut_ad(mtr->magic_n == MTR_MAGIC_N);
+	ut_ad(mtr->state == MTR_ACTIVE);
 
 	memo = &(mtr->memo);
 
@@ -92,6 +93,7 @@ mtr_set_savepoint(
 
 	ut_ad(mtr);
 	ut_ad(mtr->magic_n == MTR_MAGIC_N);
+	ut_ad(mtr->state == MTR_ACTIVE);
 
 	memo = &(mtr->memo);
 
@@ -149,6 +151,7 @@ mtr_memo_contains(
 
 	ut_ad(mtr);
 	ut_ad(mtr->magic_n == MTR_MAGIC_N);
+	ut_ad(mtr->state == MTR_ACTIVE);
 
 	memo = &(mtr->memo);
 

From 15568b946edbe7cc07fb73d1649784a98e27fc9a Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:27:34 +0400
Subject: [PATCH 097/116] Applying InnoDB snapshot

Detailed revision comments:

r6801 | marko | 2010-03-11 13:34:28 +0200 (Thu, 11 Mar 2010) | 2 lines
branches/zip: mtr_memo_contains(): Relax the assertion of r6800,
allowing mtr->state == MTR_COMMITTING.
---
 storage/innodb_plugin/include/mtr0mtr.ic | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/include/mtr0mtr.ic b/storage/innodb_plugin/include/mtr0mtr.ic
index fb6e1740198..eaf68e1b393 100644
--- a/storage/innodb_plugin/include/mtr0mtr.ic
+++ b/storage/innodb_plugin/include/mtr0mtr.ic
@@ -151,7 +151,7 @@ mtr_memo_contains(
 
 	ut_ad(mtr);
 	ut_ad(mtr->magic_n == MTR_MAGIC_N);
-	ut_ad(mtr->state == MTR_ACTIVE);
+	ut_ad(mtr->state == MTR_ACTIVE || mtr->state == MTR_COMMITTING);
 
 	memo = &(mtr->memo);
 

From 2ce59df6edfd9fe16136f63e53d0530e69070164 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:27:53 +0400
Subject: [PATCH 098/116] Applying InnoDB snapshot

Detailed revision comments:

r6805 | inaam | 2010-03-11 23:15:17 +0200 (Thu, 11 Mar 2010) | 6 lines
branches/zip issue#463

Fixed compiler warning about uninitialized variable.

Non-functional change.
---
 storage/innodb_plugin/os/os0file.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/storage/innodb_plugin/os/os0file.c b/storage/innodb_plugin/os/os0file.c
index e1effd9b177..b244e3974b3 100644
--- a/storage/innodb_plugin/os/os0file.c
+++ b/storage/innodb_plugin/os/os0file.c
@@ -3914,6 +3914,9 @@ os_aio_simulated_handle(
 	ulint		n;
 	ulint		i;
 
+	/* Fix compiler warning */
+	*consecutive_ios = NULL;
+
 	segment = os_aio_get_array_and_local_segment(&array, global_segment);
 
 restart:

From 5183fe49f0ddc381a3bacae61f860f980d74252d Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:52:14 +0400
Subject: [PATCH 099/116] Applying InnoDB snapshot

Detailed revision comments:

r6828 | calvin | 2010-03-17 17:16:38 +0200 (Wed, 17 Mar 2010) | 7 lines
branches/zip: rename IB_HAVE_PAUSE_INSTRUCTION to
HAVE_IB_PAUSE_INSTRUCTION in CMakeLists.txt.

The rename was done as r5871, but CMakeLists.txt was
forgotten. Also, add INNODB_RW_LOCKS_USE_ATOMICS to
CMake.
---
 storage/innodb_plugin/CMakeLists.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/CMakeLists.txt b/storage/innodb_plugin/CMakeLists.txt
index 25cd212a473..083fc0fe777 100644
--- a/storage/innodb_plugin/CMakeLists.txt
+++ b/storage/innodb_plugin/CMakeLists.txt
@@ -80,5 +80,5 @@ SET(INNODB_PLUGIN_SOURCES	btr/btr0btr.c btr/btr0cur.c btr/btr0pcur.c btr/btr0sea
 			usr/usr0sess.c
 			ut/ut0byte.c ut/ut0dbg.c ut/ut0mem.c ut/ut0rnd.c ut/ut0ut.c ut/ut0vec.c
 			ut/ut0list.c ut/ut0wqueue.c)
-ADD_DEFINITIONS(-DHAVE_WINDOWS_ATOMICS -DIB_HAVE_PAUSE_INSTRUCTION)
+ADD_DEFINITIONS(-DHAVE_WINDOWS_ATOMICS -DINNODB_RW_LOCKS_USE_ATOMICS -DHAVE_IB_PAUSE_INSTRUCTION)
 MYSQL_STORAGE_ENGINE(INNODB_PLUGIN)

From 74dffaa610e7b9ff59f6f703e9b04f5a45e14c3d Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:52:37 +0400
Subject: [PATCH 100/116] Applying InnoDB snapshot

Detailed revision comments:

r6830 | marko | 2010-03-18 09:48:18 +0200 (Thu, 18 Mar 2010) | 3 lines
branches/zip: buf_page_peek_if_too_old(): Use 32-bit arithmetics
when comparing the age of access_time to buf_LRU_old_threshold_ms.
This fixes a bug on 64-bit systems.
---
 storage/innodb_plugin/ChangeLog          | 7 +++++++
 storage/innodb_plugin/include/buf0buf.ic | 2 +-
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index a8b8c52908d..6aef312f7d0 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,10 @@
+2010-03-18	The InnoDB Team
+
+	* buf0buf.ic:
+	When comparing the time of the first access to a block against
+	innodb_old_blocks_time, use 32-bit arithmetics.  The comparison
+	was incorrect on 64-bit systems.
+
 2010-03-11	The InnoDB Team
 
 	* buf0buf.h, buf0buf.ic:
diff --git a/storage/innodb_plugin/include/buf0buf.ic b/storage/innodb_plugin/include/buf0buf.ic
index e5cbde73ff6..5ce9712c7fa 100644
--- a/storage/innodb_plugin/include/buf0buf.ic
+++ b/storage/innodb_plugin/include/buf0buf.ic
@@ -81,7 +81,7 @@ buf_page_peek_if_too_old(
 		unsigned	access_time = buf_page_is_accessed(bpage);
 
 		if (access_time > 0
-		    && (ut_time_ms() - access_time)
+		    && ((ib_uint32_t) (ut_time_ms() - access_time))
 		    >= buf_LRU_old_threshold_ms) {
 			return(TRUE);
 		}

From e1d20aefe14a59c41e60afbae8217ec9f37ac342 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:53:56 +0400
Subject: [PATCH 101/116] Applying InnoDB snapshot, fixes BUG#52102.

Detailed revision comments:

r6840 | calvin | 2010-03-19 00:32:23 +0200 (Fri, 19 Mar 2010) | 6 lines
branches/zip: Fix Bug #52102 InnoDB Plugin shows performance drop
comparing to builtin InnoDB (Windows only).

Disable Windows atomics by default.

Approved by: Inaam
---
 storage/innodb_plugin/CMakeLists.txt | 6 +++++-
 storage/innodb_plugin/ChangeLog      | 7 +++++++
 2 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/CMakeLists.txt b/storage/innodb_plugin/CMakeLists.txt
index 083fc0fe777..f436fe29457 100644
--- a/storage/innodb_plugin/CMakeLists.txt
+++ b/storage/innodb_plugin/CMakeLists.txt
@@ -80,5 +80,9 @@ SET(INNODB_PLUGIN_SOURCES	btr/btr0btr.c btr/btr0cur.c btr/btr0pcur.c btr/btr0sea
 			usr/usr0sess.c
 			ut/ut0byte.c ut/ut0dbg.c ut/ut0mem.c ut/ut0rnd.c ut/ut0ut.c ut/ut0vec.c
 			ut/ut0list.c ut/ut0wqueue.c)
-ADD_DEFINITIONS(-DHAVE_WINDOWS_ATOMICS -DINNODB_RW_LOCKS_USE_ATOMICS -DHAVE_IB_PAUSE_INSTRUCTION)
+# Windows atomics do not perform well. Disable Windows atomics by default.
+# See bug#52102 for details.
+#ADD_DEFINITIONS(-DHAVE_WINDOWS_ATOMICS -DINNODB_RW_LOCKS_USE_ATOMICS -DHAVE_IB_PAUSE_INSTRUCTION)
+ADD_DEFINITIONS(-DHAVE_IB_PAUSE_INSTRUCTION)
+
 MYSQL_STORAGE_ENGINE(INNODB_PLUGIN)
diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 6aef312f7d0..3dc29ce1321 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,3 +1,10 @@
+2010-03-18	The InnoDB Team
+
+	* CMakeLists.txt:
+	Fix Bug #52102 InnoDB Plugin shows performance drop comparing to
+	builtin InnoDB (Windows only).
+	Disable Windows atomics by default.
+
 2010-03-18	The InnoDB Team
 
 	* buf0buf.ic:

From 0e9c2ea16854553ef20047e78f5458cd5ed2058b Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:54:22 +0400
Subject: [PATCH 102/116] Applying InnoDB snapshot

Detailed revision comments:

r6853 | marko | 2010-03-22 13:35:29 +0200 (Mon, 22 Mar 2010) | 1 line
branches/zip: mutex_own(), rw_lock_own(): Add attribute((warn_unused_result)).
---
 storage/innodb_plugin/include/sync0rw.h   | 3 ++-
 storage/innodb_plugin/include/sync0sync.h | 3 ++-
 2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/storage/innodb_plugin/include/sync0rw.h b/storage/innodb_plugin/include/sync0rw.h
index aedfd5f3f86..630f6c30138 100644
--- a/storage/innodb_plugin/include/sync0rw.h
+++ b/storage/innodb_plugin/include/sync0rw.h
@@ -429,8 +429,9 @@ ibool
 rw_lock_own(
 /*========*/
 	rw_lock_t*	lock,		/*!< in: rw-lock */
-	ulint		lock_type);	/*!< in: lock type: RW_LOCK_SHARED,
+	ulint		lock_type)	/*!< in: lock type: RW_LOCK_SHARED,
 					RW_LOCK_EX */
+	__attribute__((warn_unused_result));
 #endif /* UNIV_SYNC_DEBUG */
 /******************************************************************//**
 Checks if somebody has locked the rw-lock in the specified mode. */
diff --git a/storage/innodb_plugin/include/sync0sync.h b/storage/innodb_plugin/include/sync0sync.h
index b93a8d1c080..d470b823fc3 100644
--- a/storage/innodb_plugin/include/sync0sync.h
+++ b/storage/innodb_plugin/include/sync0sync.h
@@ -206,7 +206,8 @@ UNIV_INTERN
 ibool
 mutex_own(
 /*======*/
-	const mutex_t*	mutex);	/*!< in: mutex */
+	const mutex_t*	mutex)	/*!< in: mutex */
+	__attribute__((warn_unused_result));
 #endif /* UNIV_DEBUG */
 #ifdef UNIV_SYNC_DEBUG
 /******************************************************************//**

From d56ea0322e0baafe718b74b51a744bde8be2e8dd Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:54:42 +0400
Subject: [PATCH 103/116] Applying InnoDB snapshot

Detailed revision comments:

r6857 | mmakela | 2010-03-23 14:07:53 +0200 (Tue, 23 Mar 2010) | 1 line
branches/zip: innodb_change_buffering: Correct the documentation.
---
 storage/innodb_plugin/handler/ha_innodb.cc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 2f61e164e42..2334d41cb39 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -10724,7 +10724,7 @@ static MYSQL_SYSVAR_BOOL(use_sys_malloc, srv_use_sys_malloc,
 static MYSQL_SYSVAR_STR(change_buffering, innobase_change_buffering,
   PLUGIN_VAR_RQCMDARG,
   "Buffer changes to reduce random access: "
-  "OFF, ON, inserting, deleting, changing, or purging.",
+  "OFF, ON, none, inserts.",
   innodb_change_buffering_validate,
   innodb_change_buffering_update, NULL);
 

From 9f0592dd38f2d268e954d69869f0a90dfdc03303 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:55:08 +0400
Subject: [PATCH 104/116] Applying InnoDB snapshot

Detailed revision comments:

r6858 | mmakela | 2010-03-23 14:09:24 +0200 (Tue, 23 Mar 2010) | 1 line
branches/zip: innodb_read_ahead_threshold: Add missing space to help string.
---
 storage/innodb_plugin/handler/ha_innodb.cc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 2334d41cb39..24a56dd3b0d 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -10730,7 +10730,7 @@ static MYSQL_SYSVAR_STR(change_buffering, innobase_change_buffering,
 
 static MYSQL_SYSVAR_ULONG(read_ahead_threshold, srv_read_ahead_threshold,
   PLUGIN_VAR_RQCMDARG,
-  "Number of pages that must be accessed sequentially for InnoDB to"
+  "Number of pages that must be accessed sequentially for InnoDB to "
   "trigger a readahead.",
   NULL, NULL, 56, 0, 64, 0);
 

From 58cb8202710f23bf178bf0dc5edef055bd751e9e Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:56:22 +0400
Subject: [PATCH 105/116] Applying InnoDB snapshot

Detailed revision comments:

r6860 | jyang | 2010-03-23 18:20:36 +0200 (Tue, 23 Mar 2010) | 5 lines
branches/zip: This is patch from Inaam that uses red-black tree
to speed up insertions into the flush_list and thus the recovery
process. The patch has been tested by Nokia.
---
 storage/innodb_plugin/CMakeLists.txt    |    2 +-
 storage/innodb_plugin/Makefile.am       |    2 +
 storage/innodb_plugin/buf/buf0buddy.c   |    2 +
 storage/innodb_plugin/buf/buf0buf.c     |   23 +-
 storage/innodb_plugin/buf/buf0flu.c     |  254 ++++-
 storage/innodb_plugin/buf/buf0lru.c     |   22 +-
 storage/innodb_plugin/buf/buf0rea.c     |    6 +-
 storage/innodb_plugin/include/buf0buf.h |   14 +
 storage/innodb_plugin/include/buf0flu.h |   30 +-
 storage/innodb_plugin/include/ut0rbt.h  |  309 ++++++
 storage/innodb_plugin/log/log0recv.c    |   19 +-
 storage/innodb_plugin/ut/ut0rbt.c       | 1249 +++++++++++++++++++++++
 12 files changed, 1880 insertions(+), 52 deletions(-)
 create mode 100644 storage/innodb_plugin/include/ut0rbt.h
 create mode 100644 storage/innodb_plugin/ut/ut0rbt.c

diff --git a/storage/innodb_plugin/CMakeLists.txt b/storage/innodb_plugin/CMakeLists.txt
index f436fe29457..87318ceec78 100644
--- a/storage/innodb_plugin/CMakeLists.txt
+++ b/storage/innodb_plugin/CMakeLists.txt
@@ -78,7 +78,7 @@ SET(INNODB_PLUGIN_SOURCES	btr/btr0btr.c btr/btr0cur.c btr/btr0pcur.c btr/btr0sea
 			trx/trx0i_s.c trx/trx0purge.c trx/trx0rec.c trx/trx0roll.c trx/trx0rseg.c
 			trx/trx0sys.c trx/trx0trx.c trx/trx0undo.c
 			usr/usr0sess.c
-			ut/ut0byte.c ut/ut0dbg.c ut/ut0mem.c ut/ut0rnd.c ut/ut0ut.c ut/ut0vec.c
+			ut/ut0byte.c ut/ut0dbg.c ut/ut0mem.c ut/ut0rbt.c ut/ut0rnd.c ut/ut0ut.c ut/ut0vec.c
 			ut/ut0list.c ut/ut0wqueue.c)
 # Windows atomics do not perform well. Disable Windows atomics by default.
 # See bug#52102 for details.
diff --git a/storage/innodb_plugin/Makefile.am b/storage/innodb_plugin/Makefile.am
index 5c71fe18d14..1d0dd936895 100644
--- a/storage/innodb_plugin/Makefile.am
+++ b/storage/innodb_plugin/Makefile.am
@@ -217,6 +217,7 @@ noinst_HEADERS=		\
 			include/ut0lst.h	\
 			include/ut0mem.h	\
 			include/ut0mem.ic	\
+			include/ut0rbt.h	\
 			include/ut0rnd.h	\
 			include/ut0rnd.ic	\
 			include/ut0sort.h	\
@@ -318,6 +319,7 @@ libinnobase_a_SOURCES=	\
 			ut/ut0dbg.c			\
 			ut/ut0list.c			\
 			ut/ut0mem.c			\
+			ut/ut0rbt.c			\
 			ut/ut0rnd.c			\
 			ut/ut0ut.c			\
 			ut/ut0vec.c			\
diff --git a/storage/innodb_plugin/buf/buf0buddy.c b/storage/innodb_plugin/buf/buf0buddy.c
index f0e1395c307..b879e97a989 100644
--- a/storage/innodb_plugin/buf/buf0buddy.c
+++ b/storage/innodb_plugin/buf/buf0buddy.c
@@ -391,6 +391,8 @@ buf_buddy_relocate_block(
 		UT_LIST_ADD_FIRST(list, buf_pool->zip_clean, dpage);
 	}
 
+	UNIV_MEM_INVALID(bpage, sizeof *bpage);
+
 	mutex_exit(&buf_pool_zip_mutex);
 	return(TRUE);
 }
diff --git a/storage/innodb_plugin/buf/buf0buf.c b/storage/innodb_plugin/buf/buf0buf.c
index 40f7925f520..d4a88565570 100644
--- a/storage/innodb_plugin/buf/buf0buf.c
+++ b/storage/innodb_plugin/buf/buf0buf.c
@@ -1191,8 +1191,6 @@ buf_relocate(
 
 	HASH_DELETE(buf_page_t, hash, buf_pool->page_hash, fold, bpage);
 	HASH_INSERT(buf_page_t, hash, buf_pool->page_hash, fold, dpage);
-
-	UNIV_MEM_INVALID(bpage, sizeof *bpage);
 }
 
 /********************************************************************//**
@@ -2224,22 +2222,8 @@ wait_until_unfixed:
 			ut_ad(!block->page.in_flush_list);
 		} else {
 			/* Relocate buf_pool->flush_list. */
-			buf_page_t*	b;
-
-			b = UT_LIST_GET_PREV(list, &block->page);
-			ut_ad(block->page.in_flush_list);
-			UT_LIST_REMOVE(list, buf_pool->flush_list,
-				       &block->page);
-
-			if (b) {
-				UT_LIST_INSERT_AFTER(
-					list, buf_pool->flush_list, b,
-					&block->page);
-			} else {
-				UT_LIST_ADD_FIRST(
-					list, buf_pool->flush_list,
-					&block->page);
-			}
+			buf_flush_relocate_on_flush_list(bpage,
+							 &block->page);
 		}
 
 		/* Buffer-fix, I/O-fix, and X-latch the block
@@ -2253,6 +2237,9 @@ wait_until_unfixed:
 		block->page.buf_fix_count = 1;
 		buf_block_set_io_fix(block, BUF_IO_READ);
 		rw_lock_x_lock(&block->lock);
+
+		UNIV_MEM_INVALID(bpage, sizeof *bpage);
+
 		mutex_exit(&block->mutex);
 		mutex_exit(&buf_pool_zip_mutex);
 		buf_pool->n_pend_unzip++;
diff --git a/storage/innodb_plugin/buf/buf0flu.c b/storage/innodb_plugin/buf/buf0flu.c
index 8b614ce90e5..86b9ced73fa 100644
--- a/storage/innodb_plugin/buf/buf0flu.c
+++ b/storage/innodb_plugin/buf/buf0flu.c
@@ -87,6 +87,138 @@ buf_flush_validate_low(void);
 /*========================*/
 #endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
 
+/********************************************************************//**
+Insert a block in the flush_rbt and returns a pointer to its
+predecessor or NULL if no predecessor. The ordering is maintained
+on the basis of the <oldest_modification, space, offset> key.
+@return pointer to the predecessor or NULL if no predecessor. */
+static
+buf_page_t*
+buf_flush_insert_in_flush_rbt(
+/*==========================*/
+	buf_page_t*	bpage)		/*!< in: bpage to be inserted. */
+{
+	buf_page_t*		prev = NULL;
+	const ib_rbt_node_t*	c_node;
+	const ib_rbt_node_t*	p_node;
+
+	ut_ad(buf_pool_mutex_own());
+
+	/* Insert this buffer into the rbt. */
+	c_node = rbt_insert(buf_pool->flush_rbt, &bpage, &bpage);
+	ut_a(c_node != NULL);
+
+	/* Get the predecessor. */
+	p_node = rbt_prev(buf_pool->flush_rbt, c_node);
+
+	if (p_node != NULL) {
+		prev = *rbt_value(buf_page_t*, p_node);
+		ut_a(prev != NULL);
+	}
+
+	return(prev);
+}
+
+/********************************************************************//**
+Delete a bpage from the flush_rbt. */
+static
+void
+buf_flush_delete_from_flush_rbt(
+/*============================*/
+	buf_page_t*	bpage)		/*!< in: bpage to be removed. */
+{
+
+	ibool	ret = FALSE;
+
+	ut_ad(buf_pool_mutex_own());
+	ret = rbt_delete(buf_pool->flush_rbt, &bpage);
+	ut_ad(ret);
+}
+
+/********************************************************************//**
+Compare two modified blocks in the buffer pool. The key for comparison
+is:
+key = <oldest_modification, space, offset>
+This comparison is used to maintian ordering of blocks in the
+buf_pool->flush_rbt.
+Note that for the purpose of flush_rbt, we only need to order blocks
+on the oldest_modification. The other two fields are used to uniquely
+identify the blocks.
+@return < 0 if b2 < b1, 0 if b2 == b1, > 0 if b2 > b1 */
+static
+int
+buf_flush_block_cmp(
+/*================*/
+	const void*	p1,		/*!< in: block1 */
+	const void*	p2)		/*!< in: block2 */
+{
+	int		ret;
+
+	ut_ad(p1 != NULL);
+	ut_ad(p2 != NULL);
+
+	const buf_page_t* b1 = *(const buf_page_t**) p1;
+	const buf_page_t* b2 = *(const buf_page_t**) p2;
+
+	ut_ad(b1 != NULL);
+	ut_ad(b2 != NULL);
+
+	ut_ad(b1->in_flush_list);
+	ut_ad(b2->in_flush_list);
+
+	if (b2->oldest_modification
+	    > b1->oldest_modification) {
+		return(1);
+	}
+
+	if (b2->oldest_modification
+	    < b1->oldest_modification) {
+		return(-1);
+	}
+
+	/* If oldest_modification is same then decide on the space. */
+	ret = (int)(b2->space - b1->space);
+
+	/* Or else decide ordering on the offset field. */
+	return(ret ? ret : (int)(b2->offset - b1->offset));
+}
+
+/********************************************************************//**
+Initialize the red-black tree to speed up insertions into the flush_list
+during recovery process. Should be called at the start of recovery
+process before any page has been read/written. */
+UNIV_INTERN
+void
+buf_flush_init_flush_rbt(void)
+/*==========================*/
+{
+	buf_pool_mutex_enter();
+
+	/* Create red black tree for speedy insertions in flush list. */
+	buf_pool->flush_rbt = rbt_create(sizeof(buf_page_t*),
+					 buf_flush_block_cmp);
+	buf_pool_mutex_exit();
+}
+
+/********************************************************************//**
+Frees up the red-black tree. */
+UNIV_INTERN
+void
+buf_flush_free_flush_rbt(void)
+/*==========================*/
+{
+	buf_pool_mutex_enter();
+
+#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
+	ut_a(buf_flush_validate_low());
+#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
+
+	rbt_free(buf_pool->flush_rbt);
+	buf_pool->flush_rbt = NULL;
+
+	buf_pool_mutex_exit();
+}
+
 /********************************************************************//**
 Inserts a modified block into the flush list. */
 UNIV_INTERN
@@ -100,6 +232,13 @@ buf_flush_insert_into_flush_list(
 	      || (UT_LIST_GET_FIRST(buf_pool->flush_list)->oldest_modification
 		  <= block->page.oldest_modification));
 
+	/* If we are in the recovery then we need to update the flush
+	red-black tree as well. */
+	if (UNIV_LIKELY_NULL(buf_pool->flush_rbt)) {
+		buf_flush_insert_sorted_into_flush_list(block);
+		return;
+	}
+
 	ut_ad(buf_block_get_state(block) == BUF_BLOCK_FILE_PAGE);
 	ut_ad(block->page.in_LRU_list);
 	ut_ad(block->page.in_page_hash);
@@ -136,12 +275,27 @@ buf_flush_insert_sorted_into_flush_list(
 	ut_d(block->page.in_flush_list = TRUE);
 
 	prev_b = NULL;
-	b = UT_LIST_GET_FIRST(buf_pool->flush_list);
 
-	while (b && b->oldest_modification > block->page.oldest_modification) {
-		ut_ad(b->in_flush_list);
-		prev_b = b;
-		b = UT_LIST_GET_NEXT(list, b);
+	/* For the most part when this function is called the flush_rbt
+	should not be NULL. In a very rare boundary case it is possible
+	that the flush_rbt has already been freed by the recovery thread
+	before the last page was hooked up in the flush_list by the
+	io-handler thread. In that case we'll  just do a simple
+	linear search in the else block. */
+	if (buf_pool->flush_rbt) {
+
+		prev_b = buf_flush_insert_in_flush_rbt(&block->page);
+
+	} else {
+
+		b = UT_LIST_GET_FIRST(buf_pool->flush_list);
+
+		while (b && b->oldest_modification
+		       > block->page.oldest_modification) {
+			ut_ad(b->in_flush_list);
+			prev_b = b;
+			b = UT_LIST_GET_NEXT(list, b);
+		}
 	}
 
 	if (prev_b == NULL) {
@@ -237,7 +391,6 @@ buf_flush_remove(
 	ut_ad(buf_pool_mutex_own());
 	ut_ad(mutex_own(buf_page_get_mutex(bpage)));
 	ut_ad(bpage->in_flush_list);
-	ut_d(bpage->in_flush_list = FALSE);
 
 	switch (buf_page_get_state(bpage)) {
 	case BUF_BLOCK_ZIP_PAGE:
@@ -259,12 +412,78 @@ buf_flush_remove(
 		break;
 	}
 
+	/* If the flush_rbt is active then delete from it as well. */
+	if (UNIV_LIKELY_NULL(buf_pool->flush_rbt)) {
+		buf_flush_delete_from_flush_rbt(bpage);
+	}
+
+	/* Must be done after we have removed it from the flush_rbt
+	because we assert on in_flush_list in comparison function. */
+	ut_d(bpage->in_flush_list = FALSE);
+
 	bpage->oldest_modification = 0;
 
 	ut_d(UT_LIST_VALIDATE(list, buf_page_t, buf_pool->flush_list,
 			      ut_ad(ut_list_node_313->in_flush_list)));
 }
 
+/********************************************************************//**
+Relocates a buffer control block on the flush_list.
+Note that it is assumed that the contents of bpage has already been
+copied to dpage. */
+UNIV_INTERN
+void
+buf_flush_relocate_on_flush_list(
+/*=============================*/
+	buf_page_t*	bpage,	/*!< in/out: control block being moved */
+	buf_page_t*	dpage)	/*!< in/out: destination block */
+{
+	buf_page_t* prev;
+	buf_page_t* prev_b = NULL;
+
+	ut_ad(buf_pool_mutex_own());
+
+	ut_ad(mutex_own(buf_page_get_mutex(bpage)));
+
+	ut_ad(bpage->in_flush_list);
+	ut_ad(dpage->in_flush_list);
+
+	/* If recovery is active we must swap the control blocks in
+	the flush_rbt as well. */
+	if (UNIV_LIKELY_NULL(buf_pool->flush_rbt)) {
+		buf_flush_delete_from_flush_rbt(bpage);
+		prev_b = buf_flush_insert_in_flush_rbt(dpage);
+	}
+
+	/* Must be done after we have removed it from the flush_rbt
+	because we assert on in_flush_list in comparison function. */
+	ut_d(bpage->in_flush_list = FALSE);
+
+	prev = UT_LIST_GET_PREV(list, bpage);
+	UT_LIST_REMOVE(list, buf_pool->flush_list, bpage);
+
+	if (prev) {
+		ut_ad(prev->in_flush_list);
+		UT_LIST_INSERT_AFTER(
+			list,
+			buf_pool->flush_list,
+			prev, dpage);
+	} else {
+		UT_LIST_ADD_FIRST(
+			list,
+			buf_pool->flush_list,
+			dpage);
+	}
+
+	/* Just an extra check. Previous in flush_list
+	should be the same control block as in flush_rbt. */
+	ut_a(!buf_pool->flush_rbt || prev_b == prev);
+
+#if defined UNIV_DEBUG || defined UNIV_BUF_DEBUG
+	ut_a(buf_flush_validate_low());
+#endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
+}
+
 /********************************************************************//**
 Updates the flush system data structures when a write is completed. */
 UNIV_INTERN
@@ -1367,24 +1586,45 @@ ibool
 buf_flush_validate_low(void)
 /*========================*/
 {
-	buf_page_t*	bpage;
+	buf_page_t*		bpage;
+	const ib_rbt_node_t*	rnode = NULL;
 
 	UT_LIST_VALIDATE(list, buf_page_t, buf_pool->flush_list,
 			 ut_ad(ut_list_node_313->in_flush_list));
 
 	bpage = UT_LIST_GET_FIRST(buf_pool->flush_list);
 
+	/* If we are in recovery mode i.e.: flush_rbt != NULL
+	then each block in the flush_list must also be present
+	in the flush_rbt. */
+	if (UNIV_LIKELY_NULL(buf_pool->flush_rbt)) {
+		rnode = rbt_first(buf_pool->flush_rbt);
+	}
+
 	while (bpage != NULL) {
 		const ib_uint64_t om = bpage->oldest_modification;
 		ut_ad(bpage->in_flush_list);
 		ut_a(buf_page_in_file(bpage));
 		ut_a(om > 0);
 
+		if (UNIV_LIKELY_NULL(buf_pool->flush_rbt)) {
+			ut_a(rnode);
+			buf_page_t* rpage = *rbt_value(buf_page_t*,
+						       rnode);
+			ut_a(rpage);
+			ut_a(rpage == bpage);
+			rnode = rbt_next(buf_pool->flush_rbt, rnode);
+		}
+
 		bpage = UT_LIST_GET_NEXT(list, bpage);
 
 		ut_a(!bpage || om >= bpage->oldest_modification);
 	}
 
+	/* By this time we must have exhausted the traversal of
+	flush_rbt (if active) as well. */
+	ut_a(rnode == NULL);
+
 	return(TRUE);
 }
 
diff --git a/storage/innodb_plugin/buf/buf0lru.c b/storage/innodb_plugin/buf/buf0lru.c
index 09878f789d4..9cfa02ba3ac 100644
--- a/storage/innodb_plugin/buf/buf0lru.c
+++ b/storage/innodb_plugin/buf/buf0lru.c
@@ -1530,26 +1530,8 @@ alloc:
 			if (b->state == BUF_BLOCK_ZIP_PAGE) {
 				buf_LRU_insert_zip_clean(b);
 			} else {
-				buf_page_t* prev;
-
-				ut_ad(b->in_flush_list);
-				ut_d(bpage->in_flush_list = FALSE);
-
-				prev = UT_LIST_GET_PREV(list, b);
-				UT_LIST_REMOVE(list, buf_pool->flush_list, b);
-
-				if (prev) {
-					ut_ad(prev->in_flush_list);
-					UT_LIST_INSERT_AFTER(
-						list,
-						buf_pool->flush_list,
-						prev, b);
-				} else {
-					UT_LIST_ADD_FIRST(
-						list,
-						buf_pool->flush_list,
-						b);
-				}
+				/* Relocate on buf_pool->flush_list. */
+				buf_flush_relocate_on_flush_list(bpage, b);
 			}
 
 			bpage->zip.data = NULL;
diff --git a/storage/innodb_plugin/buf/buf0rea.c b/storage/innodb_plugin/buf/buf0rea.c
index dd98ea17eb5..a973b1b2d26 100644
--- a/storage/innodb_plugin/buf/buf0rea.c
+++ b/storage/innodb_plugin/buf/buf0rea.c
@@ -608,14 +608,14 @@ buf_read_recv_pages(
 		while (buf_pool->n_pend_reads >= recv_n_pool_free_frames / 2) {
 
 			os_aio_simulated_wake_handler_threads();
-			os_thread_sleep(500000);
+			os_thread_sleep(10000);
 
 			count++;
 
-			if (count > 100) {
+			if (count > 1000) {
 				fprintf(stderr,
 					"InnoDB: Error: InnoDB has waited for"
-					" 50 seconds for pending\n"
+					" 10 seconds for pending\n"
 					"InnoDB: reads to the buffer pool to"
 					" be finished.\n"
 					"InnoDB: Number of pending reads %lu,"
diff --git a/storage/innodb_plugin/include/buf0buf.h b/storage/innodb_plugin/include/buf0buf.h
index b5ee2bfdaf6..cd4ee5906f0 100644
--- a/storage/innodb_plugin/include/buf0buf.h
+++ b/storage/innodb_plugin/include/buf0buf.h
@@ -33,6 +33,7 @@ Created 11/5/1995 Heikki Tuuri
 #include "hash0hash.h"
 #include "ut0byte.h"
 #include "page0types.h"
+#include "ut0rbt.h"
 #ifndef UNIV_HOTBACKUP
 #include "os0proc.h"
 
@@ -1359,6 +1360,19 @@ struct buf_pool_struct{
 					/*!< this is in the set state
 					when there is no flush batch
 					of the given type running */
+	ib_rbt_t*	flush_rbt;	/* !< a red-black tree is used
+					exclusively during recovery to
+					speed up insertions in the
+					flush_list. This tree contains
+					blocks in order of
+					oldest_modification LSN and is
+					kept in sync with the
+					flush_list.
+					Each member of the tree MUST
+					also be on the flush_list.
+					This tree is relevant only in
+					recovery and is set to NULL
+					once the recovery is over. */
 	ulint		freed_page_clock;/*!< a sequence number used
 					to count the number of buffer
 					blocks removed from the end of
diff --git a/storage/innodb_plugin/include/buf0flu.h b/storage/innodb_plugin/include/buf0flu.h
index 6c751852f54..fc462ceeba0 100644
--- a/storage/innodb_plugin/include/buf0flu.h
+++ b/storage/innodb_plugin/include/buf0flu.h
@@ -40,6 +40,16 @@ buf_flush_remove(
 /*=============*/
 	buf_page_t*	bpage);	/*!< in: pointer to the block in question */
 /********************************************************************//**
+Relocates a buffer control block on the flush_list.
+Note that it is assumed that the contents of bpage has already been
+copied to dpage. */
+UNIV_INTERN
+void
+buf_flush_relocate_on_flush_list(
+/*=============================*/
+	buf_page_t*	bpage,	/*!< in/out: control block being moved */
+	buf_page_t*	dpage);	/*!< in/out: destination block */
+/********************************************************************//**
 Updates the flush system data structures when a write is completed. */
 UNIV_INTERN
 void
@@ -139,8 +149,8 @@ how much redo the workload is generating and at what rate. */
 
 struct buf_flush_stat_struct
 {
-	ib_uint64_t	redo;		/**< amount of redo generated. */
-	ulint		n_flushed;	/**< number of pages flushed. */
+	ib_uint64_t	redo;		/*!< amount of redo generated. */
+	ulint		n_flushed;	/*!< number of pages flushed. */
 };
 
 /** Statistics for selecting flush rate of dirty pages. */
@@ -175,6 +185,22 @@ buf_flush_validate(void);
 /*====================*/
 #endif /* UNIV_DEBUG || UNIV_BUF_DEBUG */
 
+/******************************************************************//**
+Initialize the red-black tree to speed up insertions into the flush_list
+during recovery process. Should be called at the start of recovery
+process before any page has been read/written. */
+UNIV_INTERN
+void
+buf_flush_init_flush_rbt(void);
+/*==========================*/
+
+/******************************************************************//**
+Frees up the red-black tree. */
+UNIV_INTERN
+void
+buf_flush_free_flush_rbt(void);
+/*==========================*/
+
 /** When buf_flush_free_margin is called, it tries to make this many blocks
 available to replacement in the free list and at the end of the LRU list (to
 make sure that a read-ahead batch can be read efficiently in a single
diff --git a/storage/innodb_plugin/include/ut0rbt.h b/storage/innodb_plugin/include/ut0rbt.h
new file mode 100644
index 00000000000..6fd050acfe7
--- /dev/null
+++ b/storage/innodb_plugin/include/ut0rbt.h
@@ -0,0 +1,309 @@
+/*****************************************************************************
+Copyright (c) 2006, 2009, Innobase Oy. All Rights Reserved.
+
+This program is free software; you can redistribute it and/or modify it under
+the terms of the GNU General Public License as published by the Free Software
+Foundation; version 2 of the License.
+
+This program is distributed in the hope that it will be useful, but WITHOUT
+ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License along with
+this program; if not, write to the Free Software Foundation, Inc., 59 Temple
+Place, Suite 330, Boston, MA 02111-1307 USA
+
+*****************************************************************************/
+
+/*******************************************************************//**
+@file include/ut0rbt.h
+Red-Black tree implementation.
+
+Created 2007-03-20 Sunny Bains
+************************************************************************/
+
+#ifndef INNOBASE_UT0RBT_H
+#define INNOBASE_UT0RBT_H
+
+#if !defined(IB_RBT_TESTING)
+#include "univ.i"
+#include "ut0mem.h"
+#else
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <assert.h>
+
+#define	ut_malloc	malloc
+#define	ut_free		free
+#define	ulint		unsigned long
+#define	ut_a(c)		assert(c)
+#define ut_error	assert(0)
+#define	ibool		unsigned int
+#define	TRUE		1
+#define	FALSE		0
+#endif
+
+/* Red black tree typedefs */
+typedef struct ib_rbt_struct ib_rbt_t;
+typedef struct ib_rbt_node_struct ib_rbt_node_t;
+/* FIXME: Iterator is a better name than _bound_ */
+typedef struct ib_rbt_bound_struct ib_rbt_bound_t;
+typedef void (*ib_rbt_print_node)(const ib_rbt_node_t* node);
+typedef int (*ib_rbt_compare)(const void* p1, const void* p2);
+
+/* Red black tree color types */
+enum ib_rbt_color_enum {
+	IB_RBT_RED,
+	IB_RBT_BLACK
+};
+
+typedef enum ib_rbt_color_enum ib_rbt_color_t;
+
+/* Red black tree node */
+struct ib_rbt_node_struct {
+	ib_rbt_color_t	color;			/* color of this node */
+
+	ib_rbt_node_t*	left;			/* points left child */
+	ib_rbt_node_t*	right;			/* points right child */
+	ib_rbt_node_t*	parent;			/* points parent node */
+
+	char		value[1];		/* Data value */
+};
+
+/* Red black tree instance.*/
+struct	ib_rbt_struct {
+	ib_rbt_node_t*	nil;			/* Black colored node that is
+						used as a sentinel. This is
+						pre-allocated too.*/
+
+	ib_rbt_node_t*	root;			/* Root of the tree, this is
+						pre-allocated and the first
+						data node is the left child.*/
+
+	ulint		n_nodes;		/* Total number of data nodes */
+
+	ib_rbt_compare	compare;		/* Fn. to use for comparison */
+	ulint		sizeof_value;		/* Sizeof the item in bytes */
+};
+
+/* The result of searching for a key in the tree, this is useful for
+a speedy lookup and insert if key doesn't exist.*/
+struct ib_rbt_bound_struct {
+	const ib_rbt_node_t*
+			last;			/* Last node visited */
+
+	int		result;			/* Result of comparing with
+						the last non-nil node that
+						was visited */
+};
+
+/* Size in elements (t is an rb tree instance) */
+#define rbt_size(t)	(t->n_nodes)
+
+/* Check whether the rb tree is empty (t is an rb tree instance) */
+#define rbt_empty(t)	(rbt_size(t) == 0)
+
+/* Get data value (t is the data type, n is an rb tree node instance) */
+#define rbt_value(t, n) ((t*) &n->value[0])
+
+/* Compare a key with the node value (t is tree, k is key, n is node)*/
+#define rbt_compare(t, k, n) (t->compare(k, n->value))
+
+/****************************************************************//**
+Free an instance of  a red black tree */
+UNIV_INTERN
+void
+rbt_free(
+/*=====*/
+	ib_rbt_t*	tree);		/*!< in: rb tree to free */
+/****************************************************************//**
+Create an instance of a red black tree
+@return	rb tree instance */
+UNIV_INTERN
+ib_rbt_t*
+rbt_create(
+/*=======*/
+	size_t		sizeof_value,	/*!< in: size in bytes */
+	ib_rbt_compare	compare);	/*!< in: comparator */
+/****************************************************************//**
+Delete a node from the red black tree, identified by key.
+@return TRUE if success FALSE if not found */
+UNIV_INTERN
+ibool
+rbt_delete(
+/*=======*/
+	ib_rbt_t*	tree,		/*!< in: rb tree */
+	const void*	key);		/*!< in: key to delete */
+/****************************************************************//**
+Remove a node from the rb tree, the node is not free'd, that is the
+callers responsibility.
+@return	the deleted node with the const. */
+UNIV_INTERN
+ib_rbt_node_t*
+rbt_remove_node(
+/*============*/
+	ib_rbt_t*	tree,		/*!< in: rb tree */
+	const ib_rbt_node_t*
+			node);		/*!< in: node to delete, this
+					is a fudge and declared const
+					because the caller has access
+					only to const nodes.*/
+/****************************************************************//**
+Find a matching node in the rb tree.
+@return	node if found else return NULL */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_lookup(
+/*=======*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree to search */
+	const void*	key);		/*!< in: key to lookup */
+/****************************************************************//**
+Generic insert of a value in the rb tree.
+@return	inserted node */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_insert(
+/*=======*/
+	ib_rbt_t*	tree,		/*!< in: rb tree */
+	const void*	key,		/*!< in: key for ordering */
+	const void*	value);		/*!< in: data that will be
+					copied to the node.*/
+/****************************************************************//**
+Add a new node to the tree, useful for data that is pre-sorted.
+@return	appended node */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_add_node(
+/*=========*/
+	ib_rbt_t*	tree,		/*!< in: rb tree */
+	ib_rbt_bound_t*	parent,		/*!< in: parent */
+	const void*	value);		/*!< in: this value is copied
+					to the node */
+/****************************************************************//**
+Return the left most data node in the tree
+@return	left most node */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_first(
+/*======*/
+	const ib_rbt_t*	tree);		/*!< in: rb tree */
+/****************************************************************//**
+Return the right most data node in the tree
+@return	right most node */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_last(
+/*=====*/
+	const ib_rbt_t*	tree);		/*!< in: rb tree */
+/****************************************************************//**
+Return the next node from current.
+@return	successor node to current that is passed in. */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_next(
+/*=====*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	const ib_rbt_node_t*		/*!< in: current node */
+			current);
+/****************************************************************//**
+Return the prev node from current.
+@return	precedessor node to current that is passed in */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_prev(
+/*=====*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	const ib_rbt_node_t*		/*!< in: current node */
+			current);
+/****************************************************************//**
+Find the node that has the lowest key that is >= key.
+@return	node that satisfies the lower bound constraint or NULL */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_lower_bound(
+/*============*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	const void*	key);		/*!< in: key to search */
+/****************************************************************//**
+Find the node that has the greatest key that is <= key.
+@return	node that satisifies the upper bound constraint or NULL */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_upper_bound(
+/*============*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	const void*	key);		/*!< in: key to search */
+/****************************************************************//**
+Search for the key, a node will be retuned in parent.last, whether it
+was found or not. If not found then parent.last will contain the
+parent node for the possibly new key otherwise the matching node.
+@return	result of last comparison */
+UNIV_INTERN
+int
+rbt_search(
+/*=======*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	ib_rbt_bound_t*	parent,		/*!< in: search bounds */
+	const void*	key);		/*!< in: key to search */
+/****************************************************************//**
+Search for the key, a node will be retuned in parent.last, whether it
+was found or not. If not found then parent.last will contain the
+parent node for the possibly new key otherwise the matching node.
+@return	result of last comparison */
+UNIV_INTERN
+int
+rbt_search_cmp(
+/*===========*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	ib_rbt_bound_t*	parent,		/*!< in: search bounds */
+	const void*	key,		/*!< in: key to search */
+	ib_rbt_compare	compare);	/*!< in: comparator */
+/****************************************************************//**
+Clear the tree, deletes (and free's) all the nodes. */
+UNIV_INTERN
+void
+rbt_clear(
+/*======*/
+	ib_rbt_t*	tree);		/*!< in: rb tree */
+/****************************************************************//**
+Merge the node from dst into src. Return the number of nodes merged.
+@return	no. of recs merged */
+UNIV_INTERN
+ulint
+rbt_merge_uniq(
+/*===========*/
+	ib_rbt_t*	dst,		/*!< in: dst rb tree */
+	const ib_rbt_t*	src);		/*!< in: src rb tree */
+/****************************************************************//**
+Merge the node from dst into src. Return the number of nodes merged.
+Delete the nodes from src after copying node to dst. As a side effect
+the duplicates will be left untouched in the src, since we don't support
+duplicates (yet). NOTE: src and dst must be similar, the function doesn't
+check for this condition (yet).
+@return	no. of recs merged */
+UNIV_INTERN
+ulint
+rbt_merge_uniq_destructive(
+/*=======================*/
+	ib_rbt_t*	dst,		/*!< in: dst rb tree */
+	ib_rbt_t*	src);		/*!< in: src rb tree */
+/****************************************************************//**
+Verify the integrity of the RB tree. For debugging. 0 failure else height
+of tree (in count of black nodes).
+@return	TRUE if OK FALSE if tree invalid. */
+UNIV_INTERN
+ibool
+rbt_validate(
+/*=========*/
+	const ib_rbt_t*	tree);		/*!< in: tree to validate */
+/****************************************************************//**
+Iterate over the tree in depth first order. */
+UNIV_INTERN
+void
+rbt_print(
+/*======*/
+	const ib_rbt_t*		tree,	/*!< in: tree to traverse */
+	ib_rbt_print_node	print);	/*!< in: print function */
+
+#endif /* INNOBASE_UT0RBT_H */
diff --git a/storage/innodb_plugin/log/log0recv.c b/storage/innodb_plugin/log/log0recv.c
index e5a67f28de6..7f5b2df7882 100644
--- a/storage/innodb_plugin/log/log0recv.c
+++ b/storage/innodb_plugin/log/log0recv.c
@@ -138,7 +138,9 @@ UNIV_INTERN ulint	recv_max_parsed_page_no;
 /** This many frames must be left free in the buffer pool when we scan
 the log and store the scanned log records in the buffer pool: we will
 use these free frames to read in pages when we start applying the
-log records to the database. */
+log records to the database.
+This is the default value. If the actual size of the buffer pool is
+larger than 10 MB we'll set this value to 512. */
 UNIV_INTERN ulint	recv_n_pool_free_frames;
 
 /** The maximum lsn we see for a page during the recovery process. If this
@@ -294,6 +296,12 @@ recv_sys_init(
 		return;
 	}
 
+	/* Initialize red-black tree for fast insertions into the
+	flush_list during recovery process.
+	As this initialization is done while holding the buffer pool
+	mutex we perform it before acquiring recv_sys->mutex. */
+	buf_flush_init_flush_rbt();
+
 	mutex_enter(&(recv_sys->mutex));
 
 #ifndef UNIV_HOTBACKUP
@@ -303,6 +311,12 @@ recv_sys_init(
 	recv_is_from_backup = TRUE;
 #endif /* !UNIV_HOTBACKUP */
 
+	/* Set appropriate value of recv_n_pool_free_frames. */
+	if (buf_pool_get_curr_size() >= (10 * 1024 * 1024)) {
+		/* Buffer pool of size greater than 10 MB. */
+		recv_n_pool_free_frames = 512;
+	}
+
 	recv_sys->buf = ut_malloc(RECV_PARSING_BUF_SIZE);
 	recv_sys->len = 0;
 	recv_sys->recovered_offset = 0;
@@ -372,6 +386,9 @@ recv_sys_debug_free(void)
 	recv_sys->last_block_buf_start = NULL;
 
 	mutex_exit(&(recv_sys->mutex));
+
+	/* Free up the flush_rbt. */
+	buf_flush_free_flush_rbt();
 }
 # endif /* UNIV_LOG_DEBUG */
 
diff --git a/storage/innodb_plugin/ut/ut0rbt.c b/storage/innodb_plugin/ut/ut0rbt.c
new file mode 100644
index 00000000000..3d7bc91e714
--- /dev/null
+++ b/storage/innodb_plugin/ut/ut0rbt.c
@@ -0,0 +1,1249 @@
+/*****************************************************************************
+
+Copyright (c) 2006, 2009, Innobase Oy. All Rights Reserved.
+
+This program is free software; you can redistribute it and/or modify it under
+the terms of the GNU General Public License as published by the Free Software
+Foundation; version 2 of the License.
+
+This program is distributed in the hope that it will be useful, but WITHOUT
+ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License along with
+this program; if not, write to the Free Software Foundation, Inc., 59 Temple
+Place, Suite 330, Boston, MA 02111-1307 USA
+
+*****************************************************************************/
+
+/*******************************************************************//**
+@file ut/ut0rbt.c
+Red-Black tree implementation
+
+Created 2007-03-20 Sunny Bains
+***********************************************************************/
+
+#include "ut0rbt.h"
+
+/************************************************************************
+Definition of a red-black tree
+==============================
+
+A red-black tree is a binary search tree which has the following
+red-black properties:
+
+   1. Every node is either red or black.
+   2. Every leaf (NULL - in our case tree->nil) is black.
+   3. If a node is red, then both its children are black.
+   4. Every simple path from a node to a descendant leaf contains the
+      same number of black nodes.
+
+   from (3) above, the implication is that on any path from the root
+   to a leaf, red nodes must not be adjacent.
+
+   However, any number of black nodes may appear in a sequence. */
+
+#if	defined(IB_RBT_TESTING)
+#warning "Testing enabled!"
+#endif
+
+#define ROOT(t)		(t->root->left)
+#define	SIZEOF_NODE(t)	((sizeof(ib_rbt_node_t) + t->sizeof_value) - 1)
+
+/****************************************************************//**
+Print out the sub-tree recursively. */
+static
+void
+rbt_print_subtree(
+/*==============*/
+	const ib_rbt_t*		tree,	/*!< in: tree to traverse */
+	const ib_rbt_node_t*	node,	/*!< in: node to print */
+	ib_rbt_print_node	print)	/*!< in: print key function */
+{
+	/* FIXME: Doesn't do anything yet */
+	if (node != tree->nil) {
+		print(node);
+		rbt_print_subtree(tree, node->left, print);
+		rbt_print_subtree(tree, node->right, print);
+	}
+}
+
+/****************************************************************//**
+Verify that the keys are in order.
+@return	TRUE of OK. FALSE if not ordered */
+static
+ibool
+rbt_check_ordering(
+/*===============*/
+	const ib_rbt_t*		tree)	/*!< in: tree to verfify */
+{
+	const ib_rbt_node_t*	node;
+	const ib_rbt_node_t*	prev = NULL;
+
+	/* Iterate over all the nodes, comparing each node with the prev */
+	for (node = rbt_first(tree); node; node = rbt_next(tree, prev)) {
+
+		if (prev && tree->compare(prev->value, node->value) >= 0) {
+			return(FALSE);
+		}
+
+		prev = node;
+	}
+
+	return(TRUE);
+}
+
+/****************************************************************//**
+Check that every path from the root to the leaves has the same count.
+Count is expressed in the number of black nodes.
+@return	0 on failure else black height of the subtree */
+static
+ibool
+rbt_count_black_nodes(
+/*==================*/
+	const ib_rbt_t*		tree,	/*!< in: tree to verify */
+	const ib_rbt_node_t*	node)	/*!< in: start of sub-tree */
+{
+	ulint	result;
+
+	if (node != tree->nil) {
+		ulint	left_height = rbt_count_black_nodes(tree, node->left);
+
+		ulint	right_height = rbt_count_black_nodes(tree, node->right);
+
+		if (left_height == 0
+		    || right_height == 0
+		    || left_height != right_height) {
+
+			result = 0;
+		} else if (node->color == IB_RBT_RED) {
+
+			/* Case 3 */
+			if (node->left->color != IB_RBT_BLACK
+			    || node->right->color != IB_RBT_BLACK) {
+
+				result = 0;
+			} else {
+				result = left_height;
+			}
+		/* Check if it's anything other than RED or BLACK. */
+		} else if (node->color != IB_RBT_BLACK) {
+
+			result = 0;
+		} else {
+
+			result = right_height + 1;
+		}
+	} else {
+		result = 1;
+	}
+
+	return(result);
+}
+
+/****************************************************************//**
+Turn the node's right child's left sub-tree into node's right sub-tree.
+This will also make node's right child it's parent. */
+static
+void
+rbt_rotate_left(
+/*============*/
+	const ib_rbt_node_t*	nil,	/*!< in: nil node of the tree */
+	ib_rbt_node_t*		node)	/*!< in: node to rotate */
+{
+	ib_rbt_node_t*	right = node->right;
+
+	node->right = right->left;
+
+	if (right->left != nil) {
+		right->left->parent = node;
+	}
+
+	/* Right's new parent was node's parent. */
+	right->parent = node->parent;
+
+	/* Since root's parent is tree->nil and root->parent->left points
+	back to root, we can avoid the check. */
+	if (node == node->parent->left) {
+		/* Node was on the left of its parent. */
+		node->parent->left = right;
+	} else {
+		/* Node must have been on the right. */
+		node->parent->right = right;
+	}
+
+	/* Finally, put node on right's left. */
+	right->left = node;
+	node->parent = right;
+}
+
+/****************************************************************//**
+Turn the node's left child's right sub-tree into node's left sub-tree.
+This also make node's left child it's parent. */
+static
+void
+rbt_rotate_right(
+/*=============*/
+	const ib_rbt_node_t*	nil,	/*!< in: nil node of tree */
+	ib_rbt_node_t*		node)	/*!< in: node to rotate */
+{
+	ib_rbt_node_t*	left = node->left;
+
+	node->left = left->right;
+
+	if (left->right != nil) {
+		left->right->parent = node;
+	}
+
+	/* Left's new parent was node's parent. */
+	left->parent = node->parent;
+
+	/* Since root's parent is tree->nil and root->parent->left points
+	back to root, we can avoid the check. */
+	if (node == node->parent->right) {
+	    /* Node was on the left of its parent. */
+            node->parent->right = left;
+	} else {
+	    /* Node must have been on the left. */
+            node->parent->left = left;
+	}
+
+	/* Finally, put node on left's right. */
+	left->right = node;
+	node->parent = left;
+}
+
+/****************************************************************//**
+Append a node to the tree.
+@return inserted node */
+static
+ib_rbt_node_t*
+rbt_tree_add_child(
+/*===============*/
+	const ib_rbt_t*	tree,		/*!< in: rbt tree */
+	ib_rbt_bound_t*	parent,		/*!< in: node's parent */
+	ib_rbt_node_t*	node)		/*!< in: node to add */
+{
+	/* Cast away the const. */
+	ib_rbt_node_t*	last = (ib_rbt_node_t*) parent->last;
+
+	if (last == tree->root || parent->result < 0) {
+		last->left = node;
+	} else {
+		/* FIXME: We don't handle duplicates (yet)! */
+		ut_a(parent->result != 0);
+
+		last->right = node;
+	}
+
+	node->parent = last;
+
+	return(node);
+}
+
+/****************************************************************//**
+Generic binary tree insert
+@return inserted node */
+static
+ib_rbt_node_t*
+rbt_tree_insert(
+/*============*/
+	ib_rbt_t*	tree,		/*!< in: rb tree */
+	const void*	key,		/*!< in: key for ordering */
+	ib_rbt_node_t*	node)		/*!< in: node hold the insert value */
+{
+	ib_rbt_bound_t	parent;
+	ib_rbt_node_t*	current = ROOT(tree);
+
+	parent.result = 0;
+	parent.last = tree->root;
+
+	/* Regular binary search. */
+	while (current != tree->nil) {
+
+		parent.last = current;
+		parent.result = tree->compare(key, current->value);
+
+		if (parent.result < 0) {
+			current = current->left;
+		} else {
+			current = current->right;
+		}
+	}
+
+	ut_a(current == tree->nil);
+
+	rbt_tree_add_child(tree, &parent, node);
+
+	return(node);
+}
+
+/****************************************************************//**
+Balance a tree after inserting a node. */
+static
+void
+rbt_balance_tree(
+/*=============*/
+	const ib_rbt_t*	tree,		/*!< in: tree to balance */
+	ib_rbt_node_t*	node)		/*!< in: node that was inserted */
+{
+	const ib_rbt_node_t*	nil = tree->nil;
+	ib_rbt_node_t*		parent = node->parent;
+
+	/* Restore the red-black property. */
+	node->color = IB_RBT_RED;
+
+	while (node != ROOT(tree) && parent->color == IB_RBT_RED) {
+		ib_rbt_node_t*	grand_parent = parent->parent;
+
+		if (parent == grand_parent->left) {
+			ib_rbt_node_t*	uncle = grand_parent->right;
+
+			if (uncle->color == IB_RBT_RED) {
+
+				/* Case 1 - change the colors. */
+				uncle->color = IB_RBT_BLACK;
+				parent->color = IB_RBT_BLACK;
+				grand_parent->color = IB_RBT_RED;
+
+				/* Move node up the tree. */
+				node = grand_parent;
+
+			} else {
+
+				if (node == parent->right) {
+					/* Right is a black node and node is
+					to the right, case 2 - move node
+					up and rotate. */
+					node = parent;
+					rbt_rotate_left(nil, node);
+				}
+
+				grand_parent = node->parent->parent;
+
+				/* Case 3. */
+				node->parent->color = IB_RBT_BLACK;
+				grand_parent->color = IB_RBT_RED;
+
+				rbt_rotate_right(nil, grand_parent);
+			}
+
+		} else {
+			ib_rbt_node_t*	uncle = grand_parent->left;
+
+			if (uncle->color == IB_RBT_RED) {
+
+				/* Case 1 - change the colors. */
+				uncle->color = IB_RBT_BLACK;
+				parent->color = IB_RBT_BLACK;
+				grand_parent->color = IB_RBT_RED;
+
+				/* Move node up the tree. */
+				node = grand_parent;
+
+			} else {
+
+				if (node == parent->left) {
+					/* Left is a black node and node is to
+					the right, case 2 - move node up and
+					rotate. */
+					node = parent;
+					rbt_rotate_right(nil, node);
+				}
+
+				grand_parent = node->parent->parent;
+
+				/* Case 3. */
+				node->parent->color = IB_RBT_BLACK;
+				grand_parent->color = IB_RBT_RED;
+
+				rbt_rotate_left(nil, grand_parent);
+			}
+		}
+
+		parent = node->parent;
+	}
+
+	/* Color the root black. */
+	ROOT(tree)->color = IB_RBT_BLACK;
+}
+
+/****************************************************************//**
+Find the given node's successor.
+@return	successor node or NULL if no successor */
+static
+ib_rbt_node_t*
+rbt_find_successor(
+/*===============*/
+	const ib_rbt_t*		tree,	/*!< in: rb tree */
+	const ib_rbt_node_t*	current)/*!< in: this is declared const
+					because it can be called via
+					rbt_next() */
+{
+	const ib_rbt_node_t*	nil = tree->nil;
+	ib_rbt_node_t*		next = current->right;
+
+	/* Is there a sub-tree to the right that we can follow. */
+	if (next != nil) {
+
+		/* Follow the left most links of the current right child. */
+		while (next->left != nil) {
+			next = next->left;
+		}
+
+	} else { /* We will have to go up the tree to find the successor. */
+		ib_rbt_node_t*	parent = current->parent;
+
+		/* Cast away the const. */
+		next = (ib_rbt_node_t*) current;
+
+		while (parent != tree->root && next == parent->right) {
+			next = parent;
+			parent = next->parent;
+		}
+
+		next = (parent == tree->root) ? NULL : parent;
+	}
+
+	return(next);
+}
+
+/****************************************************************//**
+Find the given node's precedecessor.
+@return	predecessor node or NULL if no predecesor */
+static
+ib_rbt_node_t*
+rbt_find_predecessor(
+/*=================*/
+	const ib_rbt_t*		tree,		/*!< in: rb tree */
+	const ib_rbt_node_t*	current)	/*!< in: this is declared const
+						because it can be called via
+						rbt_prev() */
+{
+	const ib_rbt_node_t*	nil = tree->nil;
+	ib_rbt_node_t*		prev = current->left;
+
+	/* Is there a sub-tree to the left that we can follow. */
+	if (prev != nil) {
+
+		/* Follow the right most links of the current left child. */
+		while (prev->right != nil) {
+			prev = prev->right;
+		}
+
+	} else { /* We will have to go up the tree to find the precedecessor. */
+		ib_rbt_node_t*	parent = current->parent;
+
+		/* Cast away the const. */
+		prev = (ib_rbt_node_t*)current;
+
+		while (parent != tree->root && prev == parent->left) {
+			prev = parent;
+			parent = prev->parent;
+		}
+
+		prev = (parent == tree->root) ? NULL : parent;
+	}
+
+	return(prev);
+}
+
+/****************************************************************//**
+Replace node with child. After applying transformations eject becomes
+an orphan. */
+static
+void
+rbt_eject_node(
+/*===========*/
+	ib_rbt_node_t*	eject,		/*!< in: node to eject */
+	ib_rbt_node_t*	node)		/*!< in: node to replace with */
+{
+	/* Update the to be ejected node's parent's child pointers. */
+	if (eject->parent->left == eject) {
+		eject->parent->left = node;
+	} else if (eject->parent->right == eject) {
+		eject->parent->right = node;
+	} else {
+		ut_a(0);
+	}
+	/* eject is now an orphan but otherwise its pointers
+	and color are left intact. */
+
+	node->parent = eject->parent;
+}
+
+/****************************************************************//**
+Replace a node with another node. */
+static
+void
+rbt_replace_node(
+/*=============*/
+	ib_rbt_node_t*	replace,	/*!< in: node to replace */
+	ib_rbt_node_t*	node)		/*!< in: node to replace with */
+{
+	ib_rbt_color_t	color = node->color;
+
+	/* Update the node pointers. */
+	node->left = replace->left;
+	node->right = replace->right;
+
+	/* Update the child node pointers. */
+	node->left->parent = node;
+	node->right->parent = node;
+
+	/* Make the parent of replace point to node. */
+	rbt_eject_node(replace, node);
+
+	/* Swap the colors. */
+	node->color = replace->color;
+	replace->color = color;
+}
+
+/****************************************************************//**
+Detach node from the tree replacing it with one of it's children.
+@return	the child node that now occupies the position of the detached node */
+static
+ib_rbt_node_t*
+rbt_detach_node(
+/*============*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	ib_rbt_node_t*	node)		/*!< in: node to detach */
+{
+	ib_rbt_node_t*		child;
+	const ib_rbt_node_t*	nil = tree->nil;
+
+	if (node->left != nil && node->right != nil) {
+		/* Case where the node to be deleted has two children. */
+		ib_rbt_node_t*	successor = rbt_find_successor(tree, node);
+
+		ut_a(successor != nil);
+		ut_a(successor->parent != nil);
+		ut_a(successor->left == nil);
+
+		child = successor->right;
+
+		/* Remove the successor node and replace with its child. */
+		rbt_eject_node(successor, child);
+
+		/* Replace the node to delete with its successor node. */
+		rbt_replace_node(node, successor);
+	} else {
+		ut_a(node->left == nil || node->right == nil);
+
+		child = (node->left != nil) ? node->left : node->right;
+
+		/* Replace the node to delete with one of it's children. */
+		rbt_eject_node(node, child);
+	}
+
+	/* Reset the node links. */
+	node->parent = node->right = node->left = tree->nil;
+
+	return(child);
+}
+
+/****************************************************************//**
+Rebalance the right sub-tree after deletion.
+@return	node to rebalance if more rebalancing required else NULL */
+static
+ib_rbt_node_t*
+rbt_balance_right(
+/*==============*/
+	const ib_rbt_node_t*	nil,	/*!< in: rb tree nil node */
+	ib_rbt_node_t*		parent,	/*!< in: parent node */
+	ib_rbt_node_t*		sibling)/*!< in: sibling node */
+{
+	ib_rbt_node_t*		node = NULL;
+
+	ut_a(sibling != nil);
+
+	/* Case 3. */
+	if (sibling->color == IB_RBT_RED) {
+
+		parent->color = IB_RBT_RED;
+		sibling->color = IB_RBT_BLACK;
+
+		rbt_rotate_left(nil, parent);
+
+		sibling = parent->right;
+
+		ut_a(sibling != nil);
+	}
+
+	/* Since this will violate case 3 because of the change above. */
+	if (sibling->left->color == IB_RBT_BLACK
+	    && sibling->right->color == IB_RBT_BLACK) {
+
+		node = parent; /* Parent needs to be rebalanced too. */
+		sibling->color = IB_RBT_RED;
+
+	} else {
+		if (sibling->right->color == IB_RBT_BLACK) {
+
+			ut_a(sibling->left->color == IB_RBT_RED);
+
+			sibling->color = IB_RBT_RED;
+			sibling->left->color = IB_RBT_BLACK;
+
+			rbt_rotate_right(nil, sibling);
+
+			sibling = parent->right;
+			ut_a(sibling != nil);
+		}
+
+		sibling->color = parent->color;
+		sibling->right->color = IB_RBT_BLACK;
+
+		parent->color = IB_RBT_BLACK;
+
+		rbt_rotate_left(nil, parent);
+	}
+
+	return(node);
+}
+
+/****************************************************************//**
+Rebalance the left sub-tree after deletion.
+@return	node to rebalance if more rebalancing required else NULL */
+static
+ib_rbt_node_t*
+rbt_balance_left(
+/*=============*/
+	const ib_rbt_node_t*	nil,	/*!< in: rb tree nil node */
+	ib_rbt_node_t*		parent,	/*!< in: parent node */
+	ib_rbt_node_t*		sibling)/*!< in: sibling node */
+{
+	ib_rbt_node_t*	node = NULL;
+
+	ut_a(sibling != nil);
+
+	/* Case 3. */
+	if (sibling->color == IB_RBT_RED) {
+
+		parent->color = IB_RBT_RED;
+		sibling->color = IB_RBT_BLACK;
+
+		rbt_rotate_right(nil, parent);
+		sibling = parent->left;
+
+		ut_a(sibling != nil);
+	}
+
+	/* Since this will violate case 3 because of the change above. */
+	if (sibling->right->color == IB_RBT_BLACK
+	    && sibling->left->color == IB_RBT_BLACK) {
+
+		node = parent; /* Parent needs to be rebalanced too. */
+		sibling->color = IB_RBT_RED;
+
+	} else {
+		if (sibling->left->color == IB_RBT_BLACK) {
+
+			ut_a(sibling->right->color == IB_RBT_RED);
+
+			sibling->color = IB_RBT_RED;
+			sibling->right->color = IB_RBT_BLACK;
+
+			rbt_rotate_left(nil, sibling);
+
+			sibling = parent->left;
+
+			ut_a(sibling != nil);
+		}
+
+		sibling->color = parent->color;
+		sibling->left->color = IB_RBT_BLACK;
+
+		parent->color = IB_RBT_BLACK;
+
+		rbt_rotate_right(nil, parent);
+	}
+
+	return(node);
+}
+
+/****************************************************************//**
+Delete the node and rebalance the tree if necessary */
+static
+void
+rbt_remove_node_and_rebalance(
+/*==========================*/
+	ib_rbt_t*	tree,		/*!< in: rb tree */
+	ib_rbt_node_t*	node)		/*!< in: node to remove */
+{
+	/* Detach node and get the node that will be used
+	as rebalance start. */
+	ib_rbt_node_t*	child = rbt_detach_node(tree, node);
+
+	if (node->color == IB_RBT_BLACK) {
+		ib_rbt_node_t*	last = child;
+
+		ROOT(tree)->color = IB_RBT_RED;
+
+		while (child && child->color == IB_RBT_BLACK) {
+			ib_rbt_node_t*	parent = child->parent;
+
+			/* Did the deletion cause an imbalance in the
+			parents left sub-tree. */
+			if (parent->left == child) {
+
+				child = rbt_balance_right(
+					tree->nil, parent, parent->right);
+
+			} else if (parent->right == child) {
+
+				child = rbt_balance_left(
+					tree->nil, parent, parent->left);
+
+			} else {
+				ut_error;
+			}
+
+			if (child) {
+				last = child;
+			}
+		}
+
+		ut_a(last);
+
+		last->color = IB_RBT_BLACK;
+		ROOT(tree)->color = IB_RBT_BLACK;
+	}
+
+	/* Note that we have removed a node from the tree. */
+	--tree->n_nodes;
+}
+
+/****************************************************************//**
+Recursively free the nodes. */
+static
+void
+rbt_free_node(
+/*==========*/
+	ib_rbt_node_t*	node,		/*!< in: node to free */
+	ib_rbt_node_t*	nil)		/*!< in: rb tree nil node */
+{
+	if (node != nil) {
+		rbt_free_node(node->left, nil);
+		rbt_free_node(node->right, nil);
+
+		ut_free(node);
+	}
+}
+
+/****************************************************************//**
+Free all the nodes and free the tree. */
+UNIV_INTERN
+void
+rbt_free(
+/*=====*/
+	ib_rbt_t*	tree)		/*!< in: rb tree to free */
+{
+	rbt_free_node(tree->root, tree->nil);
+	ut_free(tree->nil);
+	ut_free(tree);
+}
+
+/****************************************************************//**
+Create an instance of a red black tree.
+@return	an empty rb tree */
+UNIV_INTERN
+ib_rbt_t*
+rbt_create(
+/*=======*/
+	size_t		sizeof_value,	/*!< in: sizeof data item */
+	ib_rbt_compare	compare)	/*!< in: fn to compare items */
+{
+	ib_rbt_t*	tree;
+	ib_rbt_node_t*	node;
+
+	tree = (ib_rbt_t*) ut_malloc(sizeof(*tree));
+	memset(tree, 0, sizeof(*tree));
+
+	tree->sizeof_value = sizeof_value;
+
+	/* Create the sentinel (NIL) node. */
+	node = tree->nil = (ib_rbt_node_t*) ut_malloc(sizeof(*node));
+	memset(node, 0, sizeof(*node));
+
+	node->color = IB_RBT_BLACK;
+	node->parent = node->left = node->right = node;
+
+	/* Create the "fake" root, the real root node will be the
+	left child of this node. */
+	node = tree->root = (ib_rbt_node_t*) ut_malloc(sizeof(*node));
+	memset(node, 0, sizeof(*node));
+
+	node->color = IB_RBT_BLACK;
+	node->parent = node->left = node->right = tree->nil;
+
+	tree->compare = compare;
+
+	return(tree);
+}
+
+/****************************************************************//**
+Generic insert of a value in the rb tree.
+@return	inserted node */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_insert(
+/*=======*/
+	ib_rbt_t*	tree,		/*!< in: rb tree */
+	const void*	key,		/*!< in: key for ordering */
+	const void*	value)		/*!< in: value of key, this value
+					is copied to the node */
+{
+	ib_rbt_node_t*	node;
+
+	/* Create the node that will hold the value data. */
+	node = (ib_rbt_node_t*) ut_malloc(SIZEOF_NODE(tree));
+
+	memcpy(node->value, value, tree->sizeof_value);
+	node->parent = node->left = node->right = tree->nil;
+
+	/* Insert in the tree in the usual way. */
+	rbt_tree_insert(tree, key, node);
+	rbt_balance_tree(tree, node);
+
+	++tree->n_nodes;
+
+	return(node);
+}
+
+/****************************************************************//**
+Add a new node to the tree, useful for data that is pre-sorted.
+@return	appended node */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_add_node(
+/*=========*/
+	ib_rbt_t*	tree,		/*!< in: rb tree */
+	ib_rbt_bound_t*	parent,		/*!< in: bounds */
+	const void*	value)		/*!< in: this value is copied
+					to the node */
+{
+	ib_rbt_node_t*	node;
+
+	/* Create the node that will hold the value data */
+	node = (ib_rbt_node_t*) ut_malloc(SIZEOF_NODE(tree));
+
+	memcpy(node->value, value, tree->sizeof_value);
+	node->parent = node->left = node->right = tree->nil;
+
+	/* If tree is empty */
+	if (parent->last == NULL) {
+		parent->last = tree->root;
+	}
+
+	/* Append the node, the hope here is that the caller knows
+	what s/he is doing. */
+	rbt_tree_add_child(tree, parent, node);
+	rbt_balance_tree(tree, node);
+
+	++tree->n_nodes;
+
+#if	defined(IB_RBT_TESTING)
+	ut_a(rbt_validate(tree));
+#endif
+	return(node);
+}
+
+/****************************************************************//**
+Find a matching node in the rb tree.
+@return	NULL if not found else the node where key was found */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_lookup(
+/*=======*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	const void*	key)		/*!< in: key to use for search */
+{
+	const ib_rbt_node_t*	current = ROOT(tree);
+
+	/* Regular binary search. */
+	while (current != tree->nil) {
+		int	result = tree->compare(key, current->value);
+
+		if (result < 0) {
+			current = current->left;
+		} else if (result > 0) {
+			current = current->right;
+		} else {
+			break;
+		}
+	}
+
+	return(current != tree->nil ? current : NULL);
+}
+
+/****************************************************************//**
+Delete a node from the red black tree, identified by key.
+@return	TRUE if success FALSE if not found */
+UNIV_INTERN
+ibool
+rbt_delete(
+/*=======*/
+	ib_rbt_t*	tree,		/*!< in: rb tree */
+	const void*	key)		/*!< in: key to delete */
+{
+	ibool		deleted = FALSE;
+	ib_rbt_node_t*	node = (ib_rbt_node_t*) rbt_lookup(tree, key);
+
+	if (node) {
+		rbt_remove_node_and_rebalance(tree, node);
+
+		ut_free(node);
+		deleted = TRUE;
+	}
+
+	return(deleted);
+}
+
+/****************************************************************//**
+Remove a node from the rb tree, the node is not free'd, that is the
+callers responsibility.
+@return	deleted node but without the const */
+UNIV_INTERN
+ib_rbt_node_t*
+rbt_remove_node(
+/*============*/
+	ib_rbt_t*		tree,		/*!< in: rb tree */
+	const ib_rbt_node_t*	const_node)	/*!< in: node to delete, this
+						is a fudge and declared const
+						because the caller can access
+						only const nodes */
+{
+	/* Cast away the const. */
+	rbt_remove_node_and_rebalance(tree, (ib_rbt_node_t*) const_node);
+
+	/* This is to make it easier to do something like this:
+		ut_free(rbt_remove_node(node));
+	*/
+
+	return((ib_rbt_node_t*) const_node);
+}
+
+/****************************************************************//**
+Find the node that has the lowest key that is >= key.
+@return	node satisfying the lower bound constraint or NULL */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_lower_bound(
+/*============*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	const void*	key)		/*!< in: key to search */
+{
+	ib_rbt_node_t*	lb_node = NULL;
+	ib_rbt_node_t*	current = ROOT(tree);
+
+	while (current != tree->nil) {
+		int result = tree->compare(key, current->value);
+
+		if (result > 0) {
+
+			current = current->right;
+
+		} else if (result < 0) {
+
+			lb_node = current;
+			current = current->left;
+
+		} else {
+			lb_node = current;
+			break;
+		}
+	}
+
+	return(lb_node);
+}
+
+/****************************************************************//**
+Find the node that has the greatest key that is <= key.
+@return	node satisfying the upper bound constraint or NULL */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_upper_bound(
+/*============*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	const void*	key)		/*!< in: key to search */
+{
+	ib_rbt_node_t*	ub_node = NULL;
+	ib_rbt_node_t*	current = ROOT(tree);
+
+	while (current != tree->nil) {
+		int result = tree->compare(key, current->value);
+
+		if (result > 0) {
+
+			ub_node = current;
+			current = current->right;
+
+		} else if (result < 0) {
+
+			current = current->left;
+
+		} else {
+			ub_node = current;
+			break;
+		}
+	}
+
+	return(ub_node);
+}
+
+/****************************************************************//**
+Find the node that has the greatest key that is <= key.
+@return	value of result */
+UNIV_INTERN
+int
+rbt_search(
+/*=======*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	ib_rbt_bound_t*	parent,		/*!< in: search bounds */
+	const void*	key)		/*!< in: key to search */
+{
+	ib_rbt_node_t*	current = ROOT(tree);
+
+	/* Every thing is greater than the NULL root. */
+	parent->result = 1;
+	parent->last = NULL;
+
+	while (current != tree->nil) {
+
+		parent->last = current;
+		parent->result = tree->compare(key, current->value);
+
+		if (parent->result > 0) {
+			current = current->right;
+		} else if (parent->result < 0) {
+			current = current->left;
+		} else {
+			break;
+		}
+	}
+
+	return(parent->result);
+}
+
+/****************************************************************//**
+Find the node that has the greatest key that is <= key. But use the
+supplied comparison function.
+@return	value of result */
+UNIV_INTERN
+int
+rbt_search_cmp(
+/*===========*/
+	const ib_rbt_t*	tree,		/*!< in: rb tree */
+	ib_rbt_bound_t*	parent,		/*!< in: search bounds */
+	const void*	key,		/*!< in: key to search */
+	ib_rbt_compare	compare)	/*!< in: fn to compare items */
+{
+	ib_rbt_node_t*	current = ROOT(tree);
+
+	/* Every thing is greater than the NULL root. */
+	parent->result = 1;
+	parent->last = NULL;
+
+	while (current != tree->nil) {
+
+		parent->last = current;
+		parent->result = compare(key, current->value);
+
+		if (parent->result > 0) {
+			current = current->right;
+		} else if (parent->result < 0) {
+			current = current->left;
+		} else {
+			break;
+		}
+	}
+
+	return(parent->result);
+}
+
+/****************************************************************//**
+Get the leftmost node.
+Return the left most node in the tree. */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_first(
+/*======*/
+	const ib_rbt_t*	tree)		/* in: rb tree */
+{
+	ib_rbt_node_t*	first = NULL;
+	ib_rbt_node_t*	current = ROOT(tree);
+
+	while (current != tree->nil) {
+		first = current;
+		current = current->left;
+	}
+
+	return(first);
+}
+
+/****************************************************************//**
+Return the right most node in the tree.
+@return	the rightmost node or NULL */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_last(
+/*=====*/
+	const ib_rbt_t*	tree)		/*!< in: rb tree */
+{
+	ib_rbt_node_t*	last = NULL;
+	ib_rbt_node_t*	current = ROOT(tree);
+
+	while (current != tree->nil) {
+		last = current;
+		current = current->right;
+	}
+
+	return(last);
+}
+
+/****************************************************************//**
+Return the next node.
+@return	node next from current */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_next(
+/*=====*/
+	const ib_rbt_t*		tree,	/*!< in: rb tree */
+	const ib_rbt_node_t*	current)/*!< in: current node */
+{
+	return(current ? rbt_find_successor(tree, current) : NULL);
+}
+
+/****************************************************************//**
+Return the previous node.
+@return	node prev from current */
+UNIV_INTERN
+const ib_rbt_node_t*
+rbt_prev(
+/*=====*/
+	const ib_rbt_t*		tree,	/*!< in: rb tree */
+	const ib_rbt_node_t*	current)/*!< in: current node */
+{
+	return(current ? rbt_find_predecessor(tree, current) : NULL);
+}
+
+/****************************************************************//**
+Reset the tree. Delete all the nodes. */
+UNIV_INTERN
+void
+rbt_clear(
+/*======*/
+	ib_rbt_t*	tree)		/*!< in: rb tree */
+{
+	rbt_free_node(ROOT(tree), tree->nil);
+
+	tree->n_nodes = 0;
+	tree->root->left = tree->root->right = tree->nil;
+}
+
+/****************************************************************//**
+Merge the node from dst into src. Return the number of nodes merged.
+@return	no. of recs merged */
+UNIV_INTERN
+ulint
+rbt_merge_uniq(
+/*===========*/
+	ib_rbt_t*	dst,		/*!< in: dst rb tree */
+	const ib_rbt_t*	src)		/*!< in: src rb tree */
+{
+	ib_rbt_bound_t		parent;
+	ulint			n_merged = 0;
+	const	ib_rbt_node_t*	src_node = rbt_first(src);
+
+	if (rbt_empty(src) || dst == src) {
+		return(0);
+	}
+
+	for (/* No op */; src_node; src_node = rbt_next(src, src_node)) {
+
+		if (rbt_search(dst, &parent, src_node->value) != 0) {
+			rbt_add_node(dst, &parent, src_node->value);
+			++n_merged;
+		}
+	}
+
+	return(n_merged);
+}
+
+/****************************************************************//**
+Merge the node from dst into src. Return the number of nodes merged.
+Delete the nodes from src after copying node to dst. As a side effect
+the duplicates will be left untouched in the src.
+@return	no. of recs merged */
+UNIV_INTERN
+ulint
+rbt_merge_uniq_destructive(
+/*=======================*/
+	ib_rbt_t*	dst,		/*!< in: dst rb tree */
+	ib_rbt_t*	src)		/*!< in: src rb tree */
+{
+	ib_rbt_bound_t	parent;
+	ib_rbt_node_t*	src_node;
+	ulint		old_size = rbt_size(dst);
+
+	if (rbt_empty(src) || dst == src) {
+		return(0);
+	}
+
+	for (src_node = (ib_rbt_node_t*) rbt_first(src); src_node; /* */) {
+		ib_rbt_node_t*	prev = src_node;
+
+		src_node = (ib_rbt_node_t*)rbt_next(src, prev);
+
+		/* Skip duplicates. */
+		if (rbt_search(dst, &parent, prev->value) != 0) {
+
+			/* Remove and reset the node but preserve
+			the node (data) value. */
+			rbt_remove_node_and_rebalance(src, prev);
+
+			/* The nil should be taken from the dst tree. */
+			prev->parent = prev->left = prev->right = dst->nil;
+			rbt_tree_add_child(dst, &parent, prev);
+			rbt_balance_tree(dst, prev);
+
+			++dst->n_nodes;
+		}
+	}
+
+#if	defined(IB_RBT_TESTING)
+	ut_a(rbt_validate(dst));
+	ut_a(rbt_validate(src));
+#endif
+	return(rbt_size(dst) - old_size);
+}
+
+/****************************************************************//**
+Check that every path from the root to the leaves has the same count and
+the tree nodes are in order.
+@return	TRUE if OK FALSE otherwise */
+UNIV_INTERN
+ibool
+rbt_validate(
+/*=========*/
+	const ib_rbt_t*	tree)		/*!< in: RB tree to validate */
+{
+	if (rbt_count_black_nodes(tree, ROOT(tree)) > 0) {
+		return(rbt_check_ordering(tree));
+	}
+
+	return(FALSE);
+}
+
+/****************************************************************//**
+Iterate over the tree in depth first order. */
+UNIV_INTERN
+void
+rbt_print(
+/*======*/
+	const ib_rbt_t*		tree,	/*!< in: tree to traverse */
+	ib_rbt_print_node	print)	/*!< in: print function */
+{
+	rbt_print_subtree(tree, ROOT(tree), print);
+}

From 64d2ade65db9766b18d14ff1f993603368586d87 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:57:08 +0400
Subject: [PATCH 106/116] Applying InnoDB snapshot

Detailed revision comments:

r6861 | vdimov | 2010-03-23 19:31:02 +0200 (Tue, 23 Mar 2010) | 36 lines
branches/zip:

Merge joerg@mysql.com-20100322150231-vdq0afbqtmbs6phy from BZR,

Including univ.i before mysql/plugin.h is needed to avoid this
compiler error:

o  This is how gcc puts it:
o  > > ccache /usr/local/gcc-4.3.2/bin/gcc -static-libgcc -DHAVE_CONFIG_H -I. -I../../include -I../../include -I../../include -I../../regex -I./include -I../../sql -I. -I../../zlib    -g -O3 -march=i686   -DUNIV_LINUX -MT libinnobase_a-trx0i_s.o -MD -MP -MF .deps/libinnobase_a-trx0i_s.Tpo -c -o libinnobase_a-trx0i_s.o `test -f 'trx/trx0i_s.c' || echo './'`trx/trx0i_s.c
o  > > In file included from ./include/univ.i:114,
o  > >                  from trx/trx0i_s.c:36:
o  > > ../../include/my_pthread.h:628: error: expected ')' before '*' token
o  > > In file included from ../../include/my_pthread.h:732,
o  > >                  from ./include/univ.i:114,
o  > >                  from trx/trx0i_s.c:36:
o  > > ../../include/mysql/psi/mysql_thread.h:100: error: expected specifier-qualifier-list before 'pthread_rwlock_t'
o  > > ../../include/mysql/psi/mysql_thread.h:116: error: expected specifier-qualifier-list before 'pthread_rwlock_t'
o  > > ../../include/mysql/psi/mysql_thread.h: In function 'inline_mysql_rwlock_init':
o  > > ../../include/mysql/psi/mysql_thread.h:711: error: 'mysql_rwlock_t' has no member named 'm_psi'
o  > > ../../include/mysql/psi/mysql_thread.h:716: error: 'mysql_rwlock_t' has no member named 'm_rwlock'
o  > > .... ((continued))
o
o  Intel's icc gives slightly clearer messages:
o  > > icc -static-intel -static-libgcc -DHAVE_CONFIG_H -I. -I../../include -I../../include -I../../include -I../../regex -I./include -I../../sql -I. -I../../zlib    -O3 -g -unroll2 -ip -mp -restrict -no-ftz -no-prefetch   -DUNIV_LINUX -MT libinnobase_a-trx0i_s.o -MD -MP -MF .deps/libinnobase_a-trx0i_s.Tpo -c -o libinnobase_a-trx0i_s.o `test -f 'trx/trx0i_s.c' || echo './'`trx/trx0i_s.c
o  > > ../../include/my_pthread.h(628): error: identifier "pthread_rwlock_t" is undefined
o  > >   extern int rw_pr_init(rw_pr_lock_t *);
o  > >                         ^
o  > >
o  > > ../../include/mysql/psi/mysql_thread.h(100): error: identifier "pthread_rwlock_t" is undefined
o  > >     rw_lock_t m_rwlock;
o  > >     ^
o  > >
o  > > ../../include/mysql/psi/mysql_thread.h(116): error: identifier "pthread_rwlock_t" is undefined
o  > >     rw_pr_lock_t m_prlock;
o  > >     ^
---
 storage/innodb_plugin/trx/trx0i_s.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/trx/trx0i_s.c b/storage/innodb_plugin/trx/trx0i_s.c
index 1b20eaabf42..a3f73e0c9e8 100644
--- a/storage/innodb_plugin/trx/trx0i_s.c
+++ b/storage/innodb_plugin/trx/trx0i_s.c
@@ -28,11 +28,18 @@ table cache" for later retrieval.
 Created July 17, 2007 Vasil Dimov
 *******************************************************/
 
+/* Found during the build of 5.5.3 on Linux 2.4 and early 2.6 kernels:
+   The includes "univ.i" -> "my_global.h" cause a different path
+   to be taken further down with pthread functions and types,
+   so they must come first.
+   From the symptoms, this is related to bug#46587 in the MySQL bug DB.
+*/
+#include "univ.i"
+
 #include <mysql/plugin.h>
 
 #include "mysql_addons.h"
 
-#include "univ.i"
 #include "buf0buf.h"
 #include "dict0dict.h"
 #include "ha0storage.h"

From 538d3f2a75e54b7ff4d0f1868bcd0e1c75187cc4 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:57:39 +0400
Subject: [PATCH 107/116] Applying InnoDB snapshot

Detailed revision comments:

r6864 | mmakela | 2010-03-24 14:05:53 +0200 (Wed, 24 Mar 2010) | 1 line
branches/zip: dtype_new_store_for_order_and_null_size(): Add ut_ad() on mtype.
---
 storage/innodb_plugin/include/data0type.ic | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/storage/innodb_plugin/include/data0type.ic b/storage/innodb_plugin/include/data0type.ic
index 240b4288f39..5c042f1028d 100644
--- a/storage/innodb_plugin/include/data0type.ic
+++ b/storage/innodb_plugin/include/data0type.ic
@@ -285,6 +285,10 @@ dtype_new_store_for_order_and_null_size(
 #endif
 	ulint	len;
 
+	ut_ad(type);
+	ut_ad(type->mtype >= DATA_VARCHAR);
+	ut_ad(type->mtype <= DATA_MYSQL);
+
 	buf[0] = (byte)(type->mtype & 0xFFUL);
 
 	if (type->prtype & DATA_BINARY_TYPE) {

From f2415eb426d4fa824291bae7031fb5018bab067c Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:58:02 +0400
Subject: [PATCH 108/116] Applying InnoDB snapshot

Detailed revision comments:

r6868 | mmakela | 2010-03-25 13:03:08 +0200 (Thu, 25 Mar 2010) | 1 line
branches/zip: page_validate(): Check the buf[] bounds.
---
 storage/innodb_plugin/page/page0page.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/storage/innodb_plugin/page/page0page.c b/storage/innodb_plugin/page/page0page.c
index 1068a413e0c..10008f9ac25 100644
--- a/storage/innodb_plugin/page/page0page.c
+++ b/storage/innodb_plugin/page/page0page.c
@@ -2414,8 +2414,13 @@ page_validate(
 		}
 
 		offs = page_offset(rec_get_start(rec, offsets));
+		i = rec_offs_size(offsets);
+		if (UNIV_UNLIKELY(offs + i >= UNIV_PAGE_SIZE)) {
+			fputs("InnoDB: record offset out of bounds\n", stderr);
+			goto func_exit;
+		}
 
-		for (i = rec_offs_size(offsets); i--; ) {
+		while (i--) {
 			if (UNIV_UNLIKELY(buf[offs + i])) {
 				/* No other record may overlap this */
 
@@ -2523,8 +2528,13 @@ n_owned_zero:
 
 		count++;
 		offs = page_offset(rec_get_start(rec, offsets));
+		i = rec_offs_size(offsets);
+		if (UNIV_UNLIKELY(offs + i >= UNIV_PAGE_SIZE)) {
+			fputs("InnoDB: record offset out of bounds\n", stderr);
+			goto func_exit;
+		}
 
-		for (i = rec_offs_size(offsets); i--; ) {
+		while (i--) {
 
 			if (UNIV_UNLIKELY(buf[offs + i])) {
 				fputs("InnoDB: Record overlaps another"

From 0a5b19e5afa3bd4fc0304724c71fb5bffe154f8c Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:58:34 +0400
Subject: [PATCH 109/116] Applying InnoDB snapshot

Detailed revision comments:

r6871 | vdimov | 2010-03-25 16:39:44 +0200 (Thu, 25 Mar 2010) | 4 lines
branches/zip:

Whitespace fixup to be consistent with the rest of the file.
---
 storage/innodb_plugin/ChangeLog | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 3dc29ce1321..70cc1cbda38 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -9,7 +9,7 @@
 
 	* buf0buf.ic:
 	When comparing the time of the first access to a block against
-	innodb_old_blocks_time, use 32-bit arithmetics.  The comparison
+	innodb_old_blocks_time, use 32-bit arithmetics. The comparison
 	was incorrect on 64-bit systems.
 
 2010-03-11	The InnoDB Team
@@ -48,7 +48,7 @@
 
 	* log/log0recv.c:
 	Remove a bogus assertion about page numbers exceeding 0x90000000
-	in the redo log.  Abort when encountering a corrupted redo log
+	in the redo log. Abort when encountering a corrupted redo log
 	record, unless innodb_force_recovery is set.
 
 2010-03-09	The InnoDB Team
@@ -138,7 +138,7 @@
 
 	* row/row0sel.c:
 	On the READ UNCOMMITTED isolation level, do not attempt to access
-	a clustered index record that has been marked for deletion.  The
+	a clustered index record that has been marked for deletion. The
 	built-in InnoDB in MySQL 5.1 and earlier would attempt to retrieve
 	a previous version of the record in this case.
 
@@ -167,7 +167,7 @@
 	Display the zlib version number at startup.
 	InnoDB compressed tables use zlib, and the implementation depends
 	on the zlib function compressBound(), whose definition was slightly
-	changed in zlib version 1.2.3.1 in 2006.  MySQL bundles zlib 1.2.3
+	changed in zlib version 1.2.3.1 in 2006. MySQL bundles zlib 1.2.3
 	from 2005, but some installations use a more recent zlib.
 
 2009-11-30	The InnoDB Team

From 7b4858c1589f095275eea7ea2a334c3d7f00f7dd Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 16:59:41 +0400
Subject: [PATCH 110/116] Applying InnoDB snapshot

Detailed revision comments:

r6872 | vdimov | 2010-03-25 17:03:17 +0200 (Thu, 25 Mar 2010) | 4 lines
branches/zip:

Fix ChangeLog - write only the bug title in bugs.mysql.com-related entires.
---
 storage/innodb_plugin/ChangeLog | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 70cc1cbda38..7adc0a4ed2e 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,9 +1,8 @@
 2010-03-18	The InnoDB Team
 
 	* CMakeLists.txt:
-	Fix Bug #52102 InnoDB Plugin shows performance drop comparing to
-	builtin InnoDB (Windows only).
-	Disable Windows atomics by default.
+	Fix Bug #52102 InnoDB Plugin shows performance drop compared to
+	InnoDB (Windows)
 
 2010-03-18	The InnoDB Team
 

From ca7e28039d0759b22f3a326135001de6364759e3 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 17:00:03 +0400
Subject: [PATCH 111/116] Applying InnoDB snapshot

Detailed revision comments:

r6873 | vdimov | 2010-03-25 17:06:56 +0200 (Thu, 25 Mar 2010) | 4 lines
branches/zip:

Use Bug#N instead of Bug #N to be consistent with the rest of the fil.
---
 storage/innodb_plugin/ChangeLog | 22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 7adc0a4ed2e..0e2dd05b184 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -1,7 +1,7 @@
 2010-03-18	The InnoDB Team
 
 	* CMakeLists.txt:
-	Fix Bug #52102 InnoDB Plugin shows performance drop compared to
+	Fix Bug#52102 InnoDB Plugin shows performance drop compared to
 	InnoDB (Windows)
 
 2010-03-18	The InnoDB Team
@@ -21,26 +21,26 @@
 2010-03-10	The InnoDB Team
 
 	* trx/trx0sys.c:
-	Fix Bug #51653 outdated reference to set-variable
+	Fix Bug#51653 outdated reference to set-variable
 
 2010-03-10	The InnoDB Team
 
 	* handler/ha_innodb.cc, mysql-test/innodb_bug21704.result,
 	mysql-test/innodb_bug47621.result, mysql-test/innodb_bug47621.test:
-	Fix Bug #47621 MySQL and InnoDB data dictionaries will become
+	Fix Bug#47621 MySQL and InnoDB data dictionaries will become
 	out of sync when renaming columns
 
 2010-03-10	The InnoDB Team
 
 	* handler/ha_innodb.cc:
-	Fix Bug #51356 Many Valgrind errors in error messages
+	Fix Bug#51356 Many Valgrind errors in error messages
 	with concurrent DDL
 
 2010-03-10	The InnoDB Team
 
 	* handler/ha_innodb.cc, handler/handler0alter.cc,
 	mysql-test/innodb_bug51378.result, mysql-test/innodb_bug51378.test:
-	Fix Bug #51378 Init 'ref_length' to correct value, in case an out
+	Fix Bug#51378 Init 'ref_length' to correct value, in case an out
 	of bound MySQL primary_key
 
 2010-03-10	The InnoDB Team
@@ -70,13 +70,13 @@
 2010-02-11	The InnoDB Team
 
 	* include/mem0mem.h, include/mem0mem.ic, mem/mem0mem.c:
-	Fix Bug #49535 Available memory check slows down crash
+	Fix Bug#49535 Available memory check slows down crash
 	recovery tens of times
 
 2010-02-09	The InnoDB Team
 
 	* buf/buf0buf.c:
-	Fix Bug #38901 InnoDB logs error repeatedly when trying to load
+	Fix Bug#38901 InnoDB logs error repeatedly when trying to load
 	page into buffer pool
 
 2010-02-09	The InnoDB Team
@@ -151,7 +151,7 @@
 2010-01-12	The InnoDB Team
 
 	* handler/ha_innodb.cc, handler/ha_innodb.h:
-	Fix Bug #46193 crash when accessing tables after enabling
+	Fix Bug#46193 crash when accessing tables after enabling
 	innodb_force_recovery option
 
 2010-01-12	The InnoDB Team
@@ -264,8 +264,8 @@
 	sync/sync0arr.c, sync/sync0sync.c, thr/thr0loc.c, trx/trx0i_s.c,
 	trx/trx0purge.c, trx/trx0rseg.c, trx/trx0sys.c, trx/trx0undo.c,
 	usr/usr0sess.c, ut/ut0mem.c:
-	Fix Bug #45992 innodb memory not freed after shutdown
-	Fix Bug #46656 InnoDB plugin: memory leaks (Valgrind)
+	Fix Bug#45992 innodb memory not freed after shutdown
+	Fix Bug#46656 InnoDB plugin: memory leaks (Valgrind)
 
 2009-10-29	The InnoDB Team
 
@@ -607,7 +607,7 @@
 	* dict/dict0dict.c:
 	When an index column cannot be found in the table during index
 	creation, display additional diagnostic before an assertion failure.
-	This does NOT fix Bug #44571 InnoDB Plugin crashes on ADD INDEX,
+	This does NOT fix Bug#44571 InnoDB Plugin crashes on ADD INDEX,
 	but it helps understand the reason of the crash.
 
 2009-06-17	The InnoDB Team

From a244b1d34f132edbbf02f1f9eda94a86519847a9 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 17:00:23 +0400
Subject: [PATCH 112/116] Applying InnoDB snapshot

Detailed revision comments:

r6874 | vdimov | 2010-03-25 17:17:52 +0200 (Thu, 25 Mar 2010) | 4 lines
branches/zip:

Wrap ChangeLog at 78th column
---
 storage/innodb_plugin/ChangeLog | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 0e2dd05b184..2a3da77d4d3 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -27,8 +27,8 @@
 
 	* handler/ha_innodb.cc, mysql-test/innodb_bug21704.result,
 	mysql-test/innodb_bug47621.result, mysql-test/innodb_bug47621.test:
-	Fix Bug#47621 MySQL and InnoDB data dictionaries will become
-	out of sync when renaming columns
+	Fix Bug#47621 MySQL and InnoDB data dictionaries will become out of
+	sync when renaming columns
 
 2010-03-10	The InnoDB Team
 

From 905c3d993e4f3e350bb0ef14dcb742bd4895fea8 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 17:00:43 +0400
Subject: [PATCH 113/116] Applying InnoDB snapshot

Detailed revision comments:

r6875 | vdimov | 2010-03-25 18:18:15 +0200 (Thu, 25 Mar 2010) | 4 lines
branches/zip:

Wrap line at 78 column in ChangeLog.
---
 storage/innodb_plugin/ChangeLog | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/storage/innodb_plugin/ChangeLog b/storage/innodb_plugin/ChangeLog
index 2a3da77d4d3..58e476716d8 100644
--- a/storage/innodb_plugin/ChangeLog
+++ b/storage/innodb_plugin/ChangeLog
@@ -8,8 +8,8 @@
 
 	* buf0buf.ic:
 	When comparing the time of the first access to a block against
-	innodb_old_blocks_time, use 32-bit arithmetics. The comparison
-	was incorrect on 64-bit systems.
+	innodb_old_blocks_time, use 32-bit arithmetics. The comparison was
+	incorrect on 64-bit systems.
 
 2010-03-11	The InnoDB Team
 

From a47cb8f0555fd34314bbc418b288779fae3bbdee Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 17:01:22 +0400
Subject: [PATCH 114/116] Applying InnoDB snapshot

Detailed revision comments:

r6891 | vdimov | 2010-03-26 16:19:01 +0200 (Fri, 26 Mar 2010) | 5 lines
Non-functional change: update copyright year to 2010 of the files
that have been modified after 2010-01-01 according to svn.

for f in $(svn log -v -r{2010-01-01}:HEAD |grep "^   M " |cut -b 16- |sort -u) ; do sed -i "" -E 's/(Copyright \(c\) [0-9]{4},) [0-9]{4}, (.*Innobase Oy.+All Rights Reserved)/\1 2010, \2/' $f ; done
---
 storage/innodb_plugin/buf/buf0buddy.c      | 2 +-
 storage/innodb_plugin/buf/buf0flu.c        | 2 +-
 storage/innodb_plugin/buf/buf0rea.c        | 2 +-
 storage/innodb_plugin/fil/fil0fil.c        | 2 +-
 storage/innodb_plugin/include/btr0btr.ic   | 2 +-
 storage/innodb_plugin/include/buf0buf.ic   | 2 +-
 storage/innodb_plugin/include/buf0flu.h    | 2 +-
 storage/innodb_plugin/include/data0type.ic | 2 +-
 storage/innodb_plugin/include/mtr0mtr.ic   | 2 +-
 storage/innodb_plugin/include/row0sel.h    | 2 +-
 storage/innodb_plugin/include/sync0rw.h    | 2 +-
 storage/innodb_plugin/plug.in              | 2 +-
 storage/innodb_plugin/row/row0row.c        | 2 +-
 storage/innodb_plugin/trx/trx0i_s.c        | 2 +-
 storage/innodb_plugin/trx/trx0rec.c        | 2 +-
 storage/innodb_plugin/trx/trx0sys.c        | 2 +-
 16 files changed, 16 insertions(+), 16 deletions(-)

diff --git a/storage/innodb_plugin/buf/buf0buddy.c b/storage/innodb_plugin/buf/buf0buddy.c
index b879e97a989..66d802f8a36 100644
--- a/storage/innodb_plugin/buf/buf0buddy.c
+++ b/storage/innodb_plugin/buf/buf0buddy.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 2006, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 2006, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/buf/buf0flu.c b/storage/innodb_plugin/buf/buf0flu.c
index 86b9ced73fa..ac62a7e9537 100644
--- a/storage/innodb_plugin/buf/buf0flu.c
+++ b/storage/innodb_plugin/buf/buf0flu.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/buf/buf0rea.c b/storage/innodb_plugin/buf/buf0rea.c
index a973b1b2d26..81f788baac2 100644
--- a/storage/innodb_plugin/buf/buf0rea.c
+++ b/storage/innodb_plugin/buf/buf0rea.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/fil/fil0fil.c b/storage/innodb_plugin/fil/fil0fil.c
index 29c9c2e5b6d..963e306c00c 100644
--- a/storage/innodb_plugin/fil/fil0fil.c
+++ b/storage/innodb_plugin/fil/fil0fil.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/btr0btr.ic b/storage/innodb_plugin/include/btr0btr.ic
index 4ec27117d85..97944cc2e26 100644
--- a/storage/innodb_plugin/include/btr0btr.ic
+++ b/storage/innodb_plugin/include/btr0btr.ic
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1994, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1994, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/buf0buf.ic b/storage/innodb_plugin/include/buf0buf.ic
index 5ce9712c7fa..378c3590181 100644
--- a/storage/innodb_plugin/include/buf0buf.ic
+++ b/storage/innodb_plugin/include/buf0buf.ic
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
diff --git a/storage/innodb_plugin/include/buf0flu.h b/storage/innodb_plugin/include/buf0flu.h
index fc462ceeba0..c996f6eaab4 100644
--- a/storage/innodb_plugin/include/buf0flu.h
+++ b/storage/innodb_plugin/include/buf0flu.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/data0type.ic b/storage/innodb_plugin/include/data0type.ic
index 5c042f1028d..2bf67a941bd 100644
--- a/storage/innodb_plugin/include/data0type.ic
+++ b/storage/innodb_plugin/include/data0type.ic
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/mtr0mtr.ic b/storage/innodb_plugin/include/mtr0mtr.ic
index eaf68e1b393..18f8e87b3cf 100644
--- a/storage/innodb_plugin/include/mtr0mtr.ic
+++ b/storage/innodb_plugin/include/mtr0mtr.ic
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/row0sel.h b/storage/innodb_plugin/include/row0sel.h
index 430493e4cde..8544b9d08ba 100644
--- a/storage/innodb_plugin/include/row0sel.h
+++ b/storage/innodb_plugin/include/row0sel.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1997, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1997, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/include/sync0rw.h b/storage/innodb_plugin/include/sync0rw.h
index 630f6c30138..6f7e13220c1 100644
--- a/storage/innodb_plugin/include/sync0rw.h
+++ b/storage/innodb_plugin/include/sync0rw.h
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1995, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1995, 2010, Innobase Oy. All Rights Reserved.
 Copyright (c) 2008, Google Inc.
 
 Portions of this file contain modifications contributed and copyrighted by
diff --git a/storage/innodb_plugin/plug.in b/storage/innodb_plugin/plug.in
index 94a2c969694..e638332d74a 100644
--- a/storage/innodb_plugin/plug.in
+++ b/storage/innodb_plugin/plug.in
@@ -1,5 +1,5 @@
 #
-# Copyright (c) 2006, 2009, Innobase Oy. All Rights Reserved.
+# Copyright (c) 2006, 2010, Innobase Oy. All Rights Reserved.
 # 
 # This program is free software; you can redistribute it and/or modify it under
 # the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/row/row0row.c b/storage/innodb_plugin/row/row0row.c
index 3103d95bfb2..cb7dfa2b7c9 100644
--- a/storage/innodb_plugin/row/row0row.c
+++ b/storage/innodb_plugin/row/row0row.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/trx/trx0i_s.c b/storage/innodb_plugin/trx/trx0i_s.c
index a3f73e0c9e8..c160eb2942a 100644
--- a/storage/innodb_plugin/trx/trx0i_s.c
+++ b/storage/innodb_plugin/trx/trx0i_s.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 2007, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 2007, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/trx/trx0rec.c b/storage/innodb_plugin/trx/trx0rec.c
index 38a0e4f0f44..bcc1f81381e 100644
--- a/storage/innodb_plugin/trx/trx0rec.c
+++ b/storage/innodb_plugin/trx/trx0rec.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
diff --git a/storage/innodb_plugin/trx/trx0sys.c b/storage/innodb_plugin/trx/trx0sys.c
index ba25662c8fb..410c55f132d 100644
--- a/storage/innodb_plugin/trx/trx0sys.c
+++ b/storage/innodb_plugin/trx/trx0sys.c
@@ -1,6 +1,6 @@
 /*****************************************************************************
 
-Copyright (c) 1996, 2009, Innobase Oy. All Rights Reserved.
+Copyright (c) 1996, 2010, Innobase Oy. All Rights Reserved.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software

From 7324f7628e89092a4c47243036bd010f285e48ce Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 17:01:43 +0400
Subject: [PATCH 115/116] Applying InnoDB snapshot

Detailed revision comments:

r6897 | mmakela | 2010-03-29 11:36:19 +0300 (Mon, 29 Mar 2010) | 3 lines
branches/zip: innodb_mutex_show_status(): Fix a condition
that was accidentally negated in r6781, making SHOW ENGINE INNODB MUTEX STATUS
display only locks with no OS waits.
---
 storage/innodb_plugin/handler/ha_innodb.cc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/storage/innodb_plugin/handler/ha_innodb.cc b/storage/innodb_plugin/handler/ha_innodb.cc
index 24a56dd3b0d..fbbebbce8fd 100644
--- a/storage/innodb_plugin/handler/ha_innodb.cc
+++ b/storage/innodb_plugin/handler/ha_innodb.cc
@@ -8831,7 +8831,7 @@ innodb_mutex_show_status(
 
 	for (lock = UT_LIST_GET_FIRST(rw_lock_list); lock != NULL;
 	     lock = UT_LIST_GET_NEXT(list, lock)) {
-		if (lock->count_os_wait) {
+		if (lock->count_os_wait == 0) {
 			continue;
 		}
 

From f770db518523a28e32715fc18363f23c74b8e204 Mon Sep 17 00:00:00 2001
From: Sergey Vojtovich <svoj@sun.com>
Date: Thu, 1 Apr 2010 17:02:01 +0400
Subject: [PATCH 116/116] Applying InnoDB snapshot

Detailed revision comments:

r6900 | mmakela | 2010-03-29 13:54:57 +0300 (Mon, 29 Mar 2010) | 5 lines
branches/zip: Merge c6899 from branches/innodb+:

Add debug assertions to track down Bug #52360.
hash_table_t::magic_n: Add HASH_TABLE_MAGIC_N checks, which were fully absent.
ut_hash_ulint(): Assert table_size > 0 before division.
---
 storage/innodb_plugin/ha/ha0ha.c           | 16 +++++++++++++++-
 storage/innodb_plugin/ha/hash0hash.c       |  6 +++++-
 storage/innodb_plugin/include/hash0hash.h  |  5 +++--
 storage/innodb_plugin/include/hash0hash.ic | 20 ++++++++++++++++++++
 storage/innodb_plugin/include/ut0rnd.ic    |  1 +
 5 files changed, 44 insertions(+), 4 deletions(-)

diff --git a/storage/innodb_plugin/ha/ha0ha.c b/storage/innodb_plugin/ha/ha0ha.c
index cb5e541b55d..db85288298d 100644
--- a/storage/innodb_plugin/ha/ha0ha.c
+++ b/storage/innodb_plugin/ha/ha0ha.c
@@ -101,6 +101,8 @@ ha_clear(
 	ulint	i;
 	ulint	n;
 
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 #ifdef UNIV_SYNC_DEBUG
 	ut_ad(rw_lock_own(&btr_search_latch, RW_LOCK_EXCLUSIVE));
 #endif /* UNIV_SYNC_DEBUG */
@@ -146,7 +148,9 @@ ha_insert_for_fold_func(
 	ha_node_t*	prev_node;
 	ulint		hash;
 
-	ut_ad(table && data);
+	ut_ad(data);
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 #if defined UNIV_AHI_DEBUG || defined UNIV_DEBUG
 	ut_a(block->frame == page_align(data));
 #endif /* UNIV_AHI_DEBUG || UNIV_DEBUG */
@@ -237,6 +241,8 @@ ha_delete_hash_node(
 	hash_table_t*	table,		/*!< in: hash table */
 	ha_node_t*	del_node)	/*!< in: node to be deleted */
 {
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 #if defined UNIV_AHI_DEBUG || defined UNIV_DEBUG
 # ifndef UNIV_HOTBACKUP
 	if (table->adaptive) {
@@ -267,6 +273,8 @@ ha_search_and_update_if_found_func(
 {
 	ha_node_t*	node;
 
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	ASSERT_HASH_MUTEX_OWN(table, fold);
 #if defined UNIV_AHI_DEBUG || defined UNIV_DEBUG
 	ut_a(new_block->frame == page_align(new_data));
@@ -304,6 +312,8 @@ ha_remove_all_nodes_to_page(
 {
 	ha_node_t*	node;
 
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	ASSERT_HASH_MUTEX_OWN(table, fold);
 
 	node = ha_chain_get_first(table, fold);
@@ -353,6 +363,8 @@ ha_validate(
 	ibool		ok	= TRUE;
 	ulint		i;
 
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	ut_a(start_index <= end_index);
 	ut_a(start_index < hash_get_n_cells(table));
 	ut_a(end_index < hash_get_n_cells(table));
@@ -391,6 +403,8 @@ ha_print_info(
 	FILE*		file,	/*!< in: file where to print */
 	hash_table_t*	table)	/*!< in: hash table */
 {
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 #ifdef UNIV_DEBUG
 /* Some of the code here is disabled for performance reasons in production
 builds, see http://bugs.mysql.com/36941 */
diff --git a/storage/innodb_plugin/ha/hash0hash.c b/storage/innodb_plugin/ha/hash0hash.c
index 2800d7793f8..30c304dafcd 100644
--- a/storage/innodb_plugin/ha/hash0hash.c
+++ b/storage/innodb_plugin/ha/hash0hash.c
@@ -119,7 +119,7 @@ hash_create(
 	table->heaps = NULL;
 #endif /* !UNIV_HOTBACKUP */
 	table->heap = NULL;
-	table->magic_n = HASH_TABLE_MAGIC_N;
+	ut_d(table->magic_n = HASH_TABLE_MAGIC_N);
 
 	/* Initialize the cell array */
 	hash_table_clear(table);
@@ -135,6 +135,8 @@ hash_table_free(
 /*============*/
 	hash_table_t*	table)	/*!< in, own: hash table */
 {
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 #ifndef UNIV_HOTBACKUP
 	ut_a(table->mutexes == NULL);
 #endif /* !UNIV_HOTBACKUP */
@@ -160,6 +162,8 @@ hash_create_mutexes_func(
 {
 	ulint	i;
 
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	ut_a(n_mutexes > 0);
 	ut_a(ut_is_2pow(n_mutexes));
 
diff --git a/storage/innodb_plugin/include/hash0hash.h b/storage/innodb_plugin/include/hash0hash.h
index 977cb829f35..b17c21a45ef 100644
--- a/storage/innodb_plugin/include/hash0hash.h
+++ b/storage/innodb_plugin/include/hash0hash.h
@@ -434,11 +434,12 @@ struct hash_table_struct {
 				these heaps */
 #endif /* !UNIV_HOTBACKUP */
 	mem_heap_t*	heap;
+#ifdef UNIV_DEBUG
 	ulint		magic_n;
+# define HASH_TABLE_MAGIC_N	76561114
+#endif /* UNIV_DEBUG */
 };
 
-#define HASH_TABLE_MAGIC_N	76561114
-
 #ifndef UNIV_NONINL
 #include "hash0hash.ic"
 #endif
diff --git a/storage/innodb_plugin/include/hash0hash.ic b/storage/innodb_plugin/include/hash0hash.ic
index 19da2d50701..0b437894e2e 100644
--- a/storage/innodb_plugin/include/hash0hash.ic
+++ b/storage/innodb_plugin/include/hash0hash.ic
@@ -35,6 +35,8 @@ hash_get_nth_cell(
 	hash_table_t*	table,	/*!< in: hash table */
 	ulint		n)	/*!< in: cell index */
 {
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	ut_ad(n < table->n_cells);
 
 	return(table->array + n);
@@ -48,6 +50,8 @@ hash_table_clear(
 /*=============*/
 	hash_table_t*	table)	/*!< in/out: hash table */
 {
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	memset(table->array, 0x0,
 	       table->n_cells * sizeof(*table->array));
 }
@@ -61,6 +65,8 @@ hash_get_n_cells(
 /*=============*/
 	hash_table_t*	table)	/*!< in: table */
 {
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	return(table->n_cells);
 }
 
@@ -74,6 +80,8 @@ hash_calc_hash(
 	ulint		fold,	/*!< in: folded value */
 	hash_table_t*	table)	/*!< in: hash table */
 {
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	return(ut_hash_ulint(fold, table->n_cells));
 }
 
@@ -88,6 +96,8 @@ hash_get_mutex_no(
 	hash_table_t*	table,	/*!< in: hash table */
 	ulint		fold)	/*!< in: fold */
 {
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	ut_ad(ut_is_2pow(table->n_mutexes));
 	return(ut_2pow_remainder(hash_calc_hash(fold, table),
 				 table->n_mutexes));
@@ -103,6 +113,8 @@ hash_get_nth_heap(
 	hash_table_t*	table,	/*!< in: hash table */
 	ulint		i)	/*!< in: index of the heap */
 {
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	ut_ad(i < table->n_mutexes);
 
 	return(table->heaps[i]);
@@ -120,6 +132,9 @@ hash_get_heap(
 {
 	ulint	i;
 
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
+
 	if (table->heap) {
 		return(table->heap);
 	}
@@ -139,6 +154,8 @@ hash_get_nth_mutex(
 	hash_table_t*	table,	/*!< in: hash table */
 	ulint		i)	/*!< in: index of the mutex */
 {
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
 	ut_ad(i < table->n_mutexes);
 
 	return(table->mutexes + i);
@@ -156,6 +173,9 @@ hash_get_mutex(
 {
 	ulint	i;
 
+	ut_ad(table);
+	ut_ad(table->magic_n == HASH_TABLE_MAGIC_N);
+
 	i = hash_get_mutex_no(table, fold);
 
 	return(hash_get_nth_mutex(table, i));
diff --git a/storage/innodb_plugin/include/ut0rnd.ic b/storage/innodb_plugin/include/ut0rnd.ic
index 763469142ec..c3dbd86923c 100644
--- a/storage/innodb_plugin/include/ut0rnd.ic
+++ b/storage/innodb_plugin/include/ut0rnd.ic
@@ -152,6 +152,7 @@ ut_hash_ulint(
 	ulint	 key,		/*!< in: value to be hashed */
 	ulint	 table_size)	/*!< in: hash table size */
 {
+	ut_ad(table_size);
 	key = key ^ UT_HASH_RANDOM_MASK2;
 
 	return(key % table_size);