v1-0001-Fill-testing-gap-for-possible-referential-integri.patch
application/octet-stream
Filename: v1-0001-Fill-testing-gap-for-possible-referential-integri.patch
Type: application/octet-stream
Part: 1
Message:
Foreign key isolation tests
Patch
Format: format-patch
Series: patch v1-0001
Subject: Fill testing gap for possible referential integrity violation
| File | + | − |
|---|---|---|
| src/test/isolation/expected/fk-snapshot-2.out | 17 | 0 |
| src/test/isolation/isolation_schedule | 1 | 0 |
| src/test/isolation/specs/fk-snapshot-2.spec | 33 | 0 |
From 7573a60df631b254701f6ad0196720b6ffcfc6ed Mon Sep 17 00:00:00 2001
From: "Paul A. Jungwirth" <pj@illuminatedcomputing.com>
Date: Thu, 15 May 2025 10:41:18 -0400
Subject: [PATCH v1 1/3] Fill testing gap for possible referential integrity
violation
This commit adds a missing isolation test for (non-PERIOD) foreign
keys. With REPEATABLE READ, one transaction can insert a referencing
row while another deletes the referenced row, and both see a valid
state. But after they have committed, the table violates referential
integrity.
If the INSERT precedes the DELETE, we use a crosscheck snapshot to see
the just-added row, so that the DELETE can raise a foreign key error.
You can see the table violate referential integrity if you change
ri_restrict to pass false for detectNewRows to ri_PerformCheck.
A crosscheck snapshot is not needed when the DELETE comes first,
because the INSERT's trigger takes a FOR KEY SHARE lock that sees the
row now marked for deletion, waits for that transaction to commit, and
raises a serialization error. I added a test for that too though.
We already have a similar test (in ri-triggers.spec) for SERIALIZABLE
snapshot isolation showing that you can implement foreign keys with
just pl/pgSQL, but that test does nothing to validate ri_triggers.c. We
also have tests (in fk-snapshot.spec) for other concurrency scenarios,
but not this one: we test concurrently deleting both the referencing
and referenced row, when the constraint activates a cascade/set null
action. But those tests don't exercise ri_restrict, and the consequence
of omitting a crosscheck comparison is different: a serialization
failure, not a referential integrity violation.
---
src/test/isolation/expected/fk-snapshot-2.out | 17 ++++++++++
src/test/isolation/isolation_schedule | 1 +
src/test/isolation/specs/fk-snapshot-2.spec | 33 +++++++++++++++++++
3 files changed, 51 insertions(+)
create mode 100644 src/test/isolation/expected/fk-snapshot-2.out
create mode 100644 src/test/isolation/specs/fk-snapshot-2.spec
diff --git a/src/test/isolation/expected/fk-snapshot-2.out b/src/test/isolation/expected/fk-snapshot-2.out
new file mode 100644
index 00000000000..202d1429a5a
--- /dev/null
+++ b/src/test/isolation/expected/fk-snapshot-2.out
@@ -0,0 +1,17 @@
+Parsed test spec with 2 sessions
+
+starting permutation: s2ins s1del s2c s1c
+step s2ins: INSERT INTO child VALUES (1, 1);
+step s1del: DELETE FROM parent WHERE parent_id = 1; <waiting ...>
+step s2c: COMMIT;
+step s1del: <... completed>
+ERROR: update or delete on table "parent" violates foreign key constraint "child_parent_id_fkey" on table "child"
+step s1c: COMMIT;
+
+starting permutation: s1del s2ins s1c s2c
+step s1del: DELETE FROM parent WHERE parent_id = 1;
+step s2ins: INSERT INTO child VALUES (1, 1); <waiting ...>
+step s1c: COMMIT;
+step s2ins: <... completed>
+ERROR: could not serialize access due to concurrent update
+step s2c: COMMIT;
diff --git a/src/test/isolation/isolation_schedule b/src/test/isolation/isolation_schedule
index e3c669a29c7..12b6581d5ab 100644
--- a/src/test/isolation/isolation_schedule
+++ b/src/test/isolation/isolation_schedule
@@ -35,6 +35,7 @@ test: fk-deadlock2
test: fk-partitioned-1
test: fk-partitioned-2
test: fk-snapshot
+test: fk-snapshot-2
test: subxid-overflow
test: eval-plan-qual
test: eval-plan-qual-trigger
diff --git a/src/test/isolation/specs/fk-snapshot-2.spec b/src/test/isolation/specs/fk-snapshot-2.spec
new file mode 100644
index 00000000000..335f763ac3a
--- /dev/null
+++ b/src/test/isolation/specs/fk-snapshot-2.spec
@@ -0,0 +1,33 @@
+# RI Trigger test
+#
+# Test C-based referential integrity enforcement.
+# Under REPEATABLE READ we need some snapshot trickery in C,
+# or we would permit things that violate referential integrity.
+
+setup
+{
+ CREATE TABLE parent (parent_id SERIAL NOT NULL PRIMARY KEY);
+ CREATE TABLE child (
+ child_id SERIAL NOT NULL PRIMARY KEY,
+ parent_id INTEGER REFERENCES parent);
+ INSERT INTO parent VALUES(1);
+}
+
+teardown { DROP TABLE parent, child; }
+
+session s1
+setup { BEGIN ISOLATION LEVEL REPEATABLE READ; }
+step s1del { DELETE FROM parent WHERE parent_id = 1; }
+step s1c { COMMIT; }
+
+session s2
+setup { BEGIN ISOLATION LEVEL REPEATABLE READ; }
+step s2ins { INSERT INTO child VALUES (1, 1); }
+step s2c { COMMIT; }
+
+# Violates referential integrity unless we use an up-to-date crosscheck snapshot:
+permutation s2ins s1del s2c s1c
+
+# Raises a can't-serialize exception
+# when the INSERT trigger does SELECT FOR KEY SHARE:
+permutation s1del s2ins s1c s2c
--
2.45.0