database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-06-27 23:21:58 +03:00
Code Activity

tableam: Avoid relying on relation size to determine validity of tids.

Browse Source 
Instead add a tableam callback to do so. To avoid adding per
validation overhead, pass a scan to tuple_tid_valid. In heap's case
we'd otherwise incurred a RelationGetNumberOfBlocks() call for each
tid - which'd have added noticable overhead to nodeTidscan.c.

Author: Andres Freund
Reviewed-By: Ashwin Agrawal
Discussion: https://postgr.es/m/20190515185447.gno2jtqxyktylyvs@alap3.anarazel.de
This commit is contained in:
Andres Freund
2019-05-17 18:52:01 -07:00
parent 7f44ede594
commit 147e3722f7
7 changed files with 129 additions and 57 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
src/backend/access/heap/heapam.c
View File

@ -1654,8 +1654,8 @@ heap_hot_search_buffer(ItemPointer tid, Relation relation, Buffer buffer,
/*
* heap_get_latest_tid - get the latest tid of a specified tuple
*
* Actually, this gets the latest version that is visible according to
* the passed snapshot. You can pass SnapshotDirty to get the very latest,
* Actually, this gets the latest version that is visible according to the
* scan's snapshot. Create a scan using SnapshotDirty to get the very latest,
* possibly uncommitted version.
*
* *tid is both an input and an output parameter: it is updated to
@ -1663,28 +1663,20 @@ heap_hot_search_buffer(ItemPointer tid, Relation relation, Buffer buffer,
* if no version of the row passes the snapshot test.
*/
void
heap_get_latest_tid(Relation relation,
Snapshot snapshot,
heap_get_latest_tid(TableScanDesc sscan,
ItemPointer tid)
{
BlockNumber blk;
Relation relation = sscan->rs_rd;
Snapshot snapshot = sscan->rs_snapshot;
ItemPointerData ctid;
TransactionId priorXmax;
/* this is to avoid Assert failures on bad input */
if (!ItemPointerIsValid(tid))
return;
/*
* Since this can be called with user-supplied TID, don't trust the input
* too much. (RelationGetNumberOfBlocks is an expensive check, so we
* don't check t_ctid links again this way. Note that it would not do to
* call it just once and save the result, either.)
* table_get_latest_tid verified that the passed in tid is valid. Assume
* that t_ctid links are valid however - there shouldn't be invalid ones
* in the table.
*/
blk = ItemPointerGetBlockNumber(tid);
if (blk >= RelationGetNumberOfBlocks(relation))
elog(ERROR, "block number %u is out of range for relation \"%s\"",
blk, RelationGetRelationName(relation));
Assert(ItemPointerIsValid(tid));
/*
* Loop to chase down t_ctid links. At top of loop, ctid is the tuple we