Commitab77b2d

committed

Fix incorrect comment in SetBufferCommitInfoNeedsSave().

Noah Misch spotted the fact that the old comment is in fact incorrect, dueto memory ordering hazards.

1 parente93c0b8 commitab77b2dCopy full SHA for ab77b2d

File tree

-3

lines changed

-3

lines changed

Lines changed: 5 additions & 3 deletions

Original file line number	Diff line number	Diff line change
`@@ -2376,9 +2376,11 @@ SetBufferCommitInfoNeedsSave(Buffer buffer)`
`2376`	`2376`	`* making the first scan after commit of an xact that added/deleted many`
`2377`	`2377`	`* tuples.So, be as quick as we can if the buffer is already dirty. We`
`2378`	`2378`	`* do this by not acquiring spinlock if it looks like the status bits are`
`2379`		`- * already OK.(Note it is okay if someone else clears BM_JUST_DIRTIED`
`2380`		`- * immediately after we look, because the buffer content update is already`
`2381`		`- * done and will be reflected in the I/O.)`
	`2379`	`+ * already. Since we make this test unlocked, there's a chance we might`
	`2380`	`+ * fail to notice that the flags have just been cleared, and failed to reset`
	`2381`	`+ * them, due to memory-ordering issues. But since this function is only`
	`2382`	`+ * intended to be used in cases where failing to write out the data would`
	`2383`	`+ * be harmless anyway, it doesn't really matter.`
`2382`	`2384`	`*/`
`2383`	`2385`	`if ((bufHdr->flags& (BM_DIRTY \|BM_JUST_DIRTIED))!=`
`2384`	`2386`	`(BM_DIRTY \|BM_JUST_DIRTIED))`

Comments

(0)