From: Tom Lane <tgl@sss.pgh.pa.us>
Date: Tue, 14 Feb 2006 17:20:10 +0000 (+0000)
Subject: Move btbulkdelete's vacuum_delay_point() call to a place in the loop where
X-Git-Url: http://waps.l3s.uni-hannover.de/gitweb/?a=commitdiff_plain;h=0f313afdb5ca88e1a25f74c3059e1e6f0bf54668;p=users%2Fbernd%2Fpostgres.git

Move btbulkdelete's vacuum_delay_point() call to a place in the loop where
we are not holding a buffer content lock; where it was, InterruptHoldoffCount
is positive and so we'd not respond to cancel signals as intended.  Also
add missing vacuum_delay_point() call in btvacuumcleanup.  This should fix
complaint from Evgeny Gridasov about failure to respond to SIGINT/SIGTERM
in a timely fashion (bug #2257).
---

diff --git a/src/backend/access/nbtree/nbtree.c b/src/backend/access/nbtree/nbtree.c
index 28b9775088..9d44f156f4 100644
--- a/src/backend/access/nbtree/nbtree.c
+++ b/src/backend/access/nbtree/nbtree.c
@@ -619,8 +619,6 @@ btbulkdelete(PG_FUNCTION_ARGS)
 						maxoff;
 			BlockNumber nextpage;
 
-			vacuum_delay_point();
-
 			ndeletable = 0;
 			page = BufferGetPage(buf);
 			opaque = (BTPageOpaque) PageGetSpecialPointer(page);
@@ -679,6 +677,10 @@ btbulkdelete(PG_FUNCTION_ARGS)
 			}
 			else
 				_bt_relbuf(rel, buf);
+
+			/* call vacuum_delay_point while not holding any buffer lock */
+			vacuum_delay_point();
+
 			/* And advance to next page, if any */
 			if (nextpage == P_NONE)
 				break;
@@ -772,6 +774,8 @@ btvacuumcleanup(PG_FUNCTION_ARGS)
 		Page		page;
 		BTPageOpaque opaque;
 
+		vacuum_delay_point();
+
 		buf = _bt_getbuf(rel, blkno, BT_READ);
 		page = BufferGetPage(buf);
 		opaque = (BTPageOpaque) PageGetSpecialPointer(page);