From 7a99fb6e1377e86afb7c19eb58a1b02d49dc9294 Mon Sep 17 00:00:00 2001 From: Tom Lane Date: Fri, 15 Dec 2023 13:55:05 -0500 Subject: [PATCH] Fix bugs in manipulation of large objects. In v16 and up (since commit afbfc0298), large object ownership checking has been broken because object_ownercheck() didn't take care of the discrepancy between our object-address representation of large objects (classId == LargeObjectRelationId) and the catalog where their ownership info is actually stored (LargeObjectMetadataRelationId). This resulted in failures such as "unrecognized class ID: 2613" when trying to update blob properties as a non-superuser. Poking around for related bugs, I found that AlterObjectOwner_internal would pass the wrong classId to the PostAlterHook in the no-op code path where the large object already has the desired owner. Also, recordExtObjInitPriv checked for the wrong classId; that bug is only latent because the stanza is dead code anyway, but as long as we're carrying it around it should be less wrong. These bugs are quite old. In HEAD, we can reduce the scope for future bugs of this ilk by changing AlterObjectOwner_internal's API to let the translation happen inside that function, rather than requiring callers to know about it. A more bulletproof fix, perhaps, would be to start using LargeObjectMetadataRelationId as the dependency and object-address classId for blobs. However that has substantial risk of breaking third-party code; even within our own code, it'd create hassles for pg_dump which would have to cope with a version-dependent representation. For now, keep the status quo. Discussion: https://postgr.es/m/2650449.1702497209@sss.pgh.pa.us --- src/backend/catalog/aclchk.c | 4 ++-- src/backend/commands/alter.c | 17 ++++++++++++++++- 2 files changed, 18 insertions(+), 3 deletions(-) diff --git a/src/backend/catalog/aclchk.c b/src/backend/catalog/aclchk.c index 95d85a7e2c9..567807b0280 100644 --- a/src/backend/catalog/aclchk.c +++ b/src/backend/catalog/aclchk.c @@ -6135,9 +6135,9 @@ recordExtObjInitPriv(Oid objoid, Oid classoid) ReleaseSysCache(tuple); } - /* pg_largeobject_metadata */ - else if (classoid == LargeObjectMetadataRelationId) + else if (classoid == LargeObjectRelationId) { + /* For large objects, we must consult pg_largeobject_metadata */ Datum aclDatum; bool isNull; HeapTuple tuple; diff --git a/src/backend/commands/alter.c b/src/backend/commands/alter.c index 5456b8222ba..6f27a492d09 100644 --- a/src/backend/commands/alter.c +++ b/src/backend/commands/alter.c @@ -1046,9 +1046,14 @@ AlterObjectOwner_internal(Relation rel, Oid objectId, Oid new_ownerId) /* Perform actual update */ CatalogTupleUpdate(rel, &newtup->t_self, newtup); - /* Update owner dependency reference */ + /* + * Update owner dependency reference. When working on a large object, + * we have to translate back to the OID conventionally used for LOs' + * classId. + */ if (classId == LargeObjectMetadataRelationId) classId = LargeObjectRelationId; + changeDependencyOnOwner(classId, objectId, new_ownerId); /* Release memory */ @@ -1056,6 +1061,16 @@ AlterObjectOwner_internal(Relation rel, Oid objectId, Oid new_ownerId) pfree(nulls); pfree(replaces); } + else + { + /* + * No need to change anything. But when working on a large object, we + * have to translate back to the OID conventionally used for LOs' + * classId, or the post-alter hook (if any) will get confused. + */ + if (classId == LargeObjectMetadataRelationId) + classId = LargeObjectRelationId; + } InvokeObjectPostAlterHook(classId, objectId, 0); }