diff --git a/mysql-test/r/innodb.result b/mysql-test/r/innodb.result
index 0cd276852bf..c244c948211 100644
--- a/mysql-test/r/innodb.result
+++ b/mysql-test/r/innodb.result
@@ -925,7 +925,7 @@ id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
 1	SIMPLE	t1	index	NULL	b	4	NULL	#	Using index
 explain select a,b from t1;
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
-1	SIMPLE	t1	index	NULL	PRIMARY	4	NULL	#	
+1	SIMPLE	t1	index	NULL	b	4	NULL	#	Using index
 explain select a,b,c from t1;
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
 1	SIMPLE	t1	ALL	NULL	NULL	NULL	NULL	#	
@@ -1166,14 +1166,14 @@ UPDATE t1 set a=a+100 where b between 2 and 3 and a < 1000;
 SELECT * from t1;
 a	b
 1	1
+102	2
+103	3
 4	4
 5	5
 6	6
 7	7
 8	8
 9	9
-102	2
-103	3
 drop table t1;
 CREATE TABLE t1 (a int not null primary key, b int not null, key (b)) engine=innodb;
 CREATE TABLE t2 (a int not null primary key, b int not null, key (b)) engine=innodb;
@@ -1197,6 +1197,7 @@ a	b
 update t1,t2 set t1.a=t1.a+100 where t1.a=101;
 select * from t1;
 a	b
+201	1
 102	2
 103	3
 104	4
@@ -1208,11 +1209,10 @@ a	b
 110	10
 111	11
 112	12
-201	1
 update t1,t2 set t1.b=t1.b+10 where t1.b=2;
 select * from t1;
 a	b
-102	12
+201	1
 103	3
 104	4
 105	5
@@ -1222,34 +1222,34 @@ a	b
 109	9
 110	10
 111	11
+102	12
 112	12
-201	1
 update t1,t2 set t1.b=t1.b+2,t2.b=t1.b+10 where t1.b between 3 and 5 and t1.a=t2.a+100;
 select * from t1;
 a	b
-102	12
+201	1
 103	5
 104	6
-105	7
 106	6
+105	7
 107	7
 108	8
 109	9
 110	10
 111	11
+102	12
 112	12
-201	1
 select * from t2;
 a	b
 1	1
 2	2
-3	13
-4	14
-5	15
 6	6
 7	7
 8	8
 9	9
+3	13
+4	14
+5	15
 drop table t1,t2;
 CREATE TABLE t2 (   NEXT_T         BIGINT NOT NULL PRIMARY KEY) ENGINE=MyISAM;
 CREATE TABLE t1 (  B_ID           INTEGER NOT NULL PRIMARY KEY) ENGINE=InnoDB;
@@ -1300,11 +1300,11 @@ insert into t1 (id) values (null),(null),(null),(null),(null);
 update t1 set fk=69 where fk is null order by id limit 1;
 SELECT * from t1;
 id	fk
-1	69
 2	NULL
 3	NULL
 4	NULL
 5	NULL
+1	69
 drop table t1;
 create table t1 (a int not null, b int not null, key (a));
 insert into t1 values (1,1),(1,2),(1,3),(3,1),(3,2),(3,3),(3,1),(3,2),(3,3),(2,1),(2,2),(2,3);
@@ -2438,8 +2438,8 @@ insert into t1 (b) values (1);
 replace into t1 (b) values (2), (1), (3);
 select * from t1;
 a	b
-2	2
 3	1
+2	2
 4	3
 truncate table t1;
 insert into t1 (b) values (1);
@@ -2448,8 +2448,8 @@ replace into t1 (b) values (1);
 replace into t1 (b) values (3);
 select * from t1;
 a	b
-2	2
 3	1
+2	2
 4	3
 drop table t1;
 create table t1 (rowid int not null auto_increment, val int not null,primary
diff --git a/mysql-test/r/innodb_mysql.result b/mysql-test/r/innodb_mysql.result
index 8008fd50a90..ab279933853 100644
--- a/mysql-test/r/innodb_mysql.result
+++ b/mysql-test/r/innodb_mysql.result
@@ -355,13 +355,13 @@ EXPLAIN
 SELECT COUNT(*) FROM t2 LEFT JOIN t1 ON t2.fkey = t1.id
 WHERE t1.name LIKE 'A%';
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
-1	SIMPLE	t1	index	PRIMARY,name	PRIMARY	4	NULL	3	Using where
+1	SIMPLE	t1	index	PRIMARY,name	name	23	NULL	3	Using where; Using index
 1	SIMPLE	t2	ref	fkey	fkey	5	test.t1.id	1	Using where; Using index
 EXPLAIN
 SELECT COUNT(*) FROM t2 LEFT JOIN t1 ON t2.fkey = t1.id
 WHERE t1.name LIKE 'A%' OR FALSE;
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
-1	SIMPLE	t2	index	NULL	PRIMARY	4	NULL	5	
+1	SIMPLE	t2	index	NULL	fkey	5	NULL	5	Using index
 1	SIMPLE	t1	eq_ref	PRIMARY	PRIMARY	4	test.t2.fkey	1	Using where
 DROP TABLE t1,t2;
 CREATE TABLE t1 (
@@ -1268,11 +1268,11 @@ select_type	SIMPLE
 table	t1
 type	index
 possible_keys	NULL
-key	PRIMARY
-key_len	4
+key	b
+key_len	5
 ref	NULL
 rows	3
-Extra	Using filesort
+Extra	Using index; Using filesort
 SELECT * FROM t1 ORDER BY b ASC, a DESC;
 a	b
 1	1
@@ -1284,11 +1284,11 @@ select_type	SIMPLE
 table	t1
 type	index
 possible_keys	NULL
-key	PRIMARY
-key_len	4
+key	b
+key_len	5
 ref	NULL
 rows	3
-Extra	Using filesort
+Extra	Using index; Using filesort
 SELECT * FROM t1 ORDER BY b DESC, a ASC;
 a	b
 2	2
@@ -1654,3 +1654,13 @@ ALTER TABLE t1 CHANGE id id2 INT;
 DROP TABLE t2;
 DROP TABLE t1;
 End of 5.1 tests
+drop table if exists t1, t2, t3;
+create table t1(a int);
+insert into t1 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
+create table t2 (a int, b int, pk int, key(a,b), primary key(pk)) engine=innodb;
+insert into t2 select @a:=A.a+10*(B.a + 10*C.a),@a, @a from t1 A, t1 B, t1 C;
+this must use key 'a', not PRIMARY:
+explain select a from t2 where a=b;
+id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
+1	SIMPLE	t2	index	NULL	a	10	NULL	#	Using where; Using index
+drop table t1, t2;
diff --git a/mysql-test/r/join_outer_innodb.result b/mysql-test/r/join_outer_innodb.result
index 24a11b12b03..e8a2d6f668b 100644
--- a/mysql-test/r/join_outer_innodb.result
+++ b/mysql-test/r/join_outer_innodb.result
@@ -8,12 +8,12 @@ EXPLAIN
 SELECT COUNT(*) FROM t2 LEFT JOIN t1 ON t2.fkey = t1.id 
 WHERE t1.name LIKE 'A%';
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
-1	SIMPLE	t1	index	PRIMARY,name	PRIMARY	4	NULL	3	Using where
+1	SIMPLE	t1	index	PRIMARY,name	name	23	NULL	3	Using where; Using index
 1	SIMPLE	t2	ref	fkey	fkey	5	test.t1.id	1	Using where; Using index
 EXPLAIN
 SELECT COUNT(*) FROM t2 LEFT JOIN t1 ON t2.fkey = t1.id 
 WHERE t1.name LIKE 'A%' OR FALSE;
 id	select_type	table	type	possible_keys	key	key_len	ref	rows	Extra
-1	SIMPLE	t2	index	NULL	PRIMARY	4	NULL	5	
+1	SIMPLE	t2	index	NULL	fkey	5	NULL	5	Using index
 1	SIMPLE	t1	eq_ref	PRIMARY	PRIMARY	4	test.t2.fkey	1	Using where
 DROP TABLE t1,t2;
diff --git a/mysql-test/t/innodb_mysql.test b/mysql-test/t/innodb_mysql.test
index a907866f726..bcdb1328cb7 100644
--- a/mysql-test/t/innodb_mysql.test
+++ b/mysql-test/t/innodb_mysql.test
@@ -13,3 +13,19 @@ let $test_foreign_keys= 1;
 set global innodb_support_xa=default;
 set session innodb_support_xa=default;
 --source include/mix1.inc
+
+--disable_warnings
+drop table if exists t1, t2, t3;
+--enable_warnings
+#
+# BUG#35850: Performance regression in 5.1.23/5.1.24
+#
+create table t1(a int);
+insert into t1 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
+create table t2 (a int, b int, pk int, key(a,b), primary key(pk)) engine=innodb;
+insert into t2 select @a:=A.a+10*(B.a + 10*C.a),@a, @a from t1 A, t1 B, t1 C;
+--echo this must use key 'a', not PRIMARY:
+--replace_column 9 #
+explain select a from t2 where a=b;
+drop table t1, t2;
+
diff --git a/sql/sql_select.cc b/sql/sql_select.cc
index 642544a524d..fa1adf7da10 100644
--- a/sql/sql_select.cc
+++ b/sql/sql_select.cc
@@ -6506,13 +6506,16 @@ make_join_readinfo(JOIN *join, ulonglong options)
 		   !(tab->select && tab->select->quick))
 	  {					// Only read index tree
 	    /*
-	      See bug #26447: "Using the clustered index for a table scan
-	      is always faster than using a secondary index".
-	    */
+            It has turned out that the below change, while speeding things
+            up for disk-bound loads, slows them down for cases when the data
+            is in disk cache (see BUG#35850):
+	    //  See bug #26447: "Using the clustered index for a table scan
+	    //  is always faster than using a secondary index".
             if (table->s->primary_key != MAX_KEY &&
                 table->file->primary_key_is_clustered())
               tab->index= table->s->primary_key;
             else
+	    */
               tab->index=find_shortest_key(table, & table->covering_keys);
 	    tab->read_first_record= join_read_first;
 	    tab->type=JT_NEXT;		// Read with index_first / index_next