SELECT FOR UPDATE from a secondary index has no effect

[msirek]

Describe the problem
If a SELECT ... FOR UPDATE statement happens to scan a secondary index, it does not block other updates.

To Reproduce
In one SQL session:

CREATE TABLE t1 (a int PRIMARY KEY, b int, c int, INDEX c_idx(c));
INSERT INTO t1 SELECT g,g,g FROM generate_series(1,10000) g(g);
ANALYZE t1;

BEGIN;
SELECT * FROM t1 WHERE c = 1 FOR UPDATE;

In another SQL session:

UPDATE t1 SET c=-1 WHERE c = 1;
SELECT * FROM t1 WHERE c = -1;
  a | b | c
----+---+-----
  1 | 1 | -1

Expected behavior
The update statement should block until the first SQL session commits or rolls back. The predicate WHERE c = 1 causes a scan of index c_idx, but if the primary index is scanned, the update blocks.

Jira issue: CRDB-19748

[michae2]

Closing this as a duplicate of #57031. This is fixed in 23.2 when using the new SFU implementation (optimizer_use_lock_op_for_serializable).