iomap: move folio_unlock out of iomap_writeback_folio
authorJoanne Koong <joannelkoong@gmail.com>
Thu, 10 Jul 2025 13:33:33 +0000 (15:33 +0200)
committerChristian Brauner <brauner@kernel.org>
Mon, 14 Jul 2025 08:51:32 +0000 (10:51 +0200)
commitf8b6a94a4ccafa95433798ad486c548c22624028
tree6eb843f494ca1199ce096a7d552411717d47324e
parent58f0d5a30427738118c19e195e9b8897e52afddb
iomap: move folio_unlock out of iomap_writeback_folio

Move unlocking the folio out of iomap_writeback_folio into the caller.
This means the end writeback machinery is now run with the folio locked
when no writeback happened, or writeback completed extremely fast.

Note that having the folio locked over the call to folio_end_writeback in
iomap_writeback_folio means that the dropbehind handling there will never
run because the trylock fails.  The only way this can happen is if the
writepage either never wrote back any dirty data at all, in which case
the dropbehind handling isn't needed, or if all writeback finished
instantly, which is rather unlikely.  Even in the latter case the
dropbehind handling is an optional optimization so skipping it will not
cause correctness issues.

This prepares for exporting iomap_writeback_folio for use in folio
laundering.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
[hch: split from a larger patch]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/20250710133343.399917-10-hch@lst.de
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
fs/iomap/buffered-io.c