Avoid trying to write an empty WAL record in log_newpage_range().

tglsfdc · tglsfdc · commit 72a914e9fb73 · 2023-04-17T14:22:06.000-04:00
If the last few pages in the specified range are empty (all zero), then log_newpage_range() could try to emit an empty WAL record containing no FPIs. This at least upsets an Assert in ReserveXLogInsertLocation, and might perhaps have bad real-world consequences in non-assert builds. This has been broken since log_newpage_range() was introduced, but the case was hard if not impossible to hit before commit 3d6a984 decided it was okay to leave VM and FSM pages intentionally zero. Nonetheless, it seems prudent to back-patch. log_newpage_range() was added in v12 but later back-patched, so this affects all supported branches. Matthias van de Meent, per report from Justin Pryzby Discussion: https://postgr.es/m/ZD1daibg4RF50IOj@telsasoft.com
diff --git a/src/backend/access/transam/xloginsert.c b/src/backend/access/transam/xloginsert.c
@@ -1171,6 +1171,10 @@ log_newpage_range(Relation rel, ForkNumber forkNum,
 			blkno++;
 		}
 
+		/* Nothing more to do if all remaining blocks were empty. */
+		if (nbufs == 0)
+			break;
+
 		/* Write WAL record for this batch. */
 		XLogBeginInsert();
 

Original file line number	Diff line number	Diff line change
`@@ -1171,6 +1171,10 @@ log_newpage_range(Relation rel, ForkNumber forkNum,`
`1171`	`1171`	`blkno++;`
`1172`	`1172`	`}`
`1173`	`1173`
	`1174`	`+ /* Nothing more to do if all remaining blocks were empty. */`
	`1175`	`+ if (nbufs == 0)`
	`1176`	`+ break;`
	`1177`	`+`
`1174`	`1178`	`/* Write WAL record for this batch. */`
`1175`	`1179`	`XLogBeginInsert();`
`1176`	`1180`