diff mbox series

[3/6] iomap: treat a write through cache the same as FUA

Message ID 20230719195417.1704513-4-axboe@kernel.dk (mailing list archive)
State New
Headers show
Series Improve async iomap DIO performance | expand

Commit Message

Jens Axboe July 19, 2023, 7:54 p.m. UTC 
Whether we have a write back cache and are using FUA or don't have
a write back cache at all is the same situation. Treat them the same.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/iomap/direct-io.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Comments

Christoph Hellwig July 20, 2023, 4:54 a.m. UTC | #1 
On Wed, Jul 19, 2023 at 01:54:14PM -0600, Jens Axboe wrote:
> Whether we have a write back cache and are using FUA or don't have
> a write back cache at all is the same situation. Treat them the same.

This looks correct, but I think the IOMAP_DIO_WRITE_FUA is rather
misnamed now which could lead to confusion.  The comment in
__iomap_dio_rw when checking the flag and clearing IOMAP_DIO_NEED_SYNC
also needs a little update to talk about writethrough semantics and
not just FUA now.
Jens Axboe July 20, 2023, 4:23 p.m. UTC | #2 
On 7/19/23 10:54?PM, Christoph Hellwig wrote:
> On Wed, Jul 19, 2023 at 01:54:14PM -0600, Jens Axboe wrote:
>> Whether we have a write back cache and are using FUA or don't have
>> a write back cache at all is the same situation. Treat them the same.
> 
> This looks correct, but I think the IOMAP_DIO_WRITE_FUA is rather
> misnamed now which could lead to confusion.  The comment in

It is - should I rename it to IOMAP_DIO_STABLE_WRITE or something like
that as part of this change?

> __iomap_dio_rw when checking the flag and clearing IOMAP_DIO_NEED_SYNC
> also needs a little update to talk about writethrough semantics and
> not just FUA now.

Will do.
diff mbox series

Patch

diff --git a/fs/iomap/direct-io.c b/fs/iomap/direct-io.c
index 6b302bf8790b..b30c3edf2ef3 100644
--- a/fs/iomap/direct-io.c
+++ b/fs/iomap/direct-io.c
@@ -280,7 +280,8 @@  static loff_t iomap_dio_bio_iter(const struct iomap_iter *iter,
 		 * cache flushes on IO completion.
 		 */
 		if (!(iomap->flags & (IOMAP_F_SHARED|IOMAP_F_DIRTY)) &&
-		    (dio->flags & IOMAP_DIO_WRITE_FUA) && bdev_fua(iomap->bdev))
+		    (dio->flags & IOMAP_DIO_WRITE_FUA) &&
+		    (bdev_fua(iomap->bdev) || !bdev_write_cache(iomap->bdev)))
 			use_fua = true;
 	}