summaryrefslogtreecommitdiff
path: root/src/backend/replication/logical/logicalfuncs.c
diff options
context:
space:
mode:
authorÁlvaro Herrera <alvherre@kurilemu.de>2025-09-12 18:47:25 +0200
committerÁlvaro Herrera <alvherre@kurilemu.de>2025-09-12 18:47:25 +0200
commit7dcea51c2a4dcf7c512bbd4f618d1d3620f9d3d7 (patch)
tree8ede9809c7008a0df2643781cdf2eb2aa456d131 /src/backend/replication/logical/logicalfuncs.c
parent20d541a200e9dfed8affef9e798ff35ca0f30b8e (diff)
Avoid unexpected changes of CurrentResourceOwner and CurrentMemoryContext
Users of logical decoding can encounter an unexpected change of CurrentResourceOwner and CurrentMemoryContext. The problem is that, unlike other call sites of RollbackAndReleaseCurrentSubTransaction(), in reorderbuffer.c we fail to restore the original values of these global variables after being clobbered by subtransaction abort. This patch saves the values prior to the call and restores them eventually. In addition, logical.c and logicalfuncs.c had a hack to restore resource owner, presumably because of lack of this restore. Remove that. Instead, because the test coverage here is not very consistent, add an Assert() to ensure that the resowner is kept identical; this would make it easy to detect other cases of bugs were we fail to restore resowner properly. This could be removed later. This is arguably an old bug, but there appears to be no reason to backpatch it and it's risky to do so, so refrain for now. Author: Antonin Houska <ah@cybertec.at> Reported-by: Mihail Nikalayeu <mihailnikalayeu@gmail.com> Reviewed-by: Euler Taveira <euler@eulerto.com> Discussion: https://postgr.es/m/119497.1756892972@localhost
Diffstat (limited to 'src/backend/replication/logical/logicalfuncs.c')
-rw-r--r--src/backend/replication/logical/logicalfuncs.c19
1 files changed, 11 insertions, 8 deletions
diff --git a/src/backend/replication/logical/logicalfuncs.c b/src/backend/replication/logical/logicalfuncs.c
index ca53caac2f2..25f890ddeed 100644
--- a/src/backend/replication/logical/logicalfuncs.c
+++ b/src/backend/replication/logical/logicalfuncs.c
@@ -107,7 +107,7 @@ pg_logical_slot_get_changes_guts(FunctionCallInfo fcinfo, bool confirm, bool bin
XLogRecPtr end_of_wal;
XLogRecPtr wait_for_wal_lsn;
LogicalDecodingContext *ctx;
- ResourceOwner old_resowner = CurrentResourceOwner;
+ ResourceOwner old_resowner PG_USED_FOR_ASSERTS_ONLY = CurrentResourceOwner;
ArrayType *arr;
Size ndim;
List *options = NIL;
@@ -263,8 +263,18 @@ pg_logical_slot_get_changes_guts(FunctionCallInfo fcinfo, bool confirm, bool bin
* store the description into our tuplestore.
*/
if (record != NULL)
+ {
LogicalDecodingProcessRecord(ctx, ctx->reader);
+ /*
+ * We used to have bugs where logical decoding would fail to
+ * preserve the resource owner. Verify that that doesn't
+ * happen anymore. XXX this could be removed once it's been
+ * battle-tested.
+ */
+ Assert(CurrentResourceOwner == old_resowner);
+ }
+
/* check limits */
if (upto_lsn != InvalidXLogRecPtr &&
upto_lsn <= ctx->reader->EndRecPtr)
@@ -276,13 +286,6 @@ pg_logical_slot_get_changes_guts(FunctionCallInfo fcinfo, bool confirm, bool bin
}
/*
- * Logical decoding could have clobbered CurrentResourceOwner during
- * transaction management, so restore the executor's value. (This is
- * a kluge, but it's not worth cleaning up right now.)
- */
- CurrentResourceOwner = old_resowner;
-
- /*
* Next time, start where we left off. (Hunting things, the family
* business..)
*/