Fix race condition in statext_store().

Must hold some lock on the pg_statistic_ext_data catalog *before*
we look up the tuple we aim to replace.  Otherwise a concurrent
VACUUM FULL or similar operation could move it to a different TID,
leaving us trying to replace the wrong tuple.

Back-patch to v12 where this got broken.

Credit goes to Dean Rasheed; I'm just doing the clerical work.

Discussion: https://postgr.es/m/CAEZATCU0zHMDiQV0g8P2U+YSP9C1idUPrn79DajsbonwkN0xvQ@mail.gmail.com
This commit is contained in:
Tom Lane 2020-03-31 17:06:22 -04:00
parent 1fd0eee66f
commit 6c426cd437

View File

@ -322,12 +322,14 @@ statext_store(Oid statOid,
MVNDistinct *ndistinct, MVDependencies *dependencies,
MCVList *mcv, VacAttrStats **stats)
{
Relation pg_stextdata;
HeapTuple stup,
oldtup;
Datum values[Natts_pg_statistic_ext_data];
bool nulls[Natts_pg_statistic_ext_data];
bool replaces[Natts_pg_statistic_ext_data];
Relation pg_stextdata;
pg_stextdata = table_open(StatisticExtDataRelationId, RowExclusiveLock);
memset(nulls, true, sizeof(nulls));
memset(replaces, false, sizeof(replaces));
@ -371,8 +373,6 @@ statext_store(Oid statOid,
elog(ERROR, "cache lookup failed for statistics object %u", statOid);
/* replace it */
pg_stextdata = table_open(StatisticExtDataRelationId, RowExclusiveLock);
stup = heap_modify_tuple(oldtup,
RelationGetDescr(pg_stextdata),
values,