diff mbox

nbd: set discard_alignment to the granularity

Message ID 20180605154123.3226-1-josef@toxicpanda.com (mailing list archive)
State New, archived
Headers show

Commit Message

Josef Bacik June 5, 2018, 3:41 p.m. UTC
Technically we should be able to get away with 0 as the
discard_alignment, but there's no way currently for the protocol to
indicate different alignments, and in real life most disks have
discard_alignment == discard_granularity.  Just set our alignment to our
blocksize to make sure discards will actually work properly with 4k
drives.

Signed-off-by: Josef Bacik <josef@toxicpanda.com>
---
 drivers/block/nbd.c | 3 +++
 1 file changed, 3 insertions(+)

Comments

Jens Axboe June 5, 2018, 3:51 p.m. UTC | #1
On 6/5/18 9:41 AM, Josef Bacik wrote:
> Technically we should be able to get away with 0 as the
> discard_alignment, but there's no way currently for the protocol to
> indicate different alignments, and in real life most disks have
> discard_alignment == discard_granularity.  Just set our alignment to our
> blocksize to make sure discards will actually work properly with 4k
> drives.

Applied, thanks.
Wouter Verhelst June 12, 2018, 9:22 a.m. UTC | #2
Hi Josef,

On Tue, Jun 05, 2018 at 11:41:23AM -0400, Josef Bacik wrote:
> Technically we should be able to get away with 0 as the
> discard_alignment, but there's no way currently for the protocol to
> indicate different alignments,

Actually there is, with the NBD_INFO_BLOCK_SIZE (and related)
response(s) to the NBD_OPT_GO message during negotiation, but
implementing that will probably require some more netlink messages.

(some of this is still being hashed out on the NBD mailinglist; I'm sure
your insights in that would be welcome)
diff mbox

Patch

diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
index bdfbb2a7018c..ec9b43c9f7e3 100644
--- a/drivers/block/nbd.c
+++ b/drivers/block/nbd.c
@@ -238,6 +238,7 @@  static void nbd_size_update(struct nbd_device *nbd)
 
 	if (config->flags & NBD_FLAG_SEND_TRIM) {
 		nbd->disk->queue->limits.discard_granularity = config->blksize;
+		nbd->disk->queue->limits.discard_alignment = config->blksize;
 		blk_queue_max_discard_sectors(nbd->disk->queue, UINT_MAX);
 	}
 	blk_queue_logical_block_size(nbd->disk->queue, config->blksize);
@@ -1055,6 +1056,7 @@  static void nbd_config_put(struct nbd_device *nbd)
 
 		nbd->tag_set.timeout = 0;
 		nbd->disk->queue->limits.discard_granularity = 0;
+		nbd->disk->queue->limits.discard_alignment = 0;
 		blk_queue_max_discard_sectors(nbd->disk->queue, UINT_MAX);
 		blk_queue_flag_clear(QUEUE_FLAG_DISCARD, nbd->disk->queue);
 
@@ -1509,6 +1511,7 @@  static int nbd_dev_add(int index)
 	blk_queue_flag_set(QUEUE_FLAG_NONROT, disk->queue);
 	blk_queue_flag_clear(QUEUE_FLAG_ADD_RANDOM, disk->queue);
 	disk->queue->limits.discard_granularity = 0;
+	disk->queue->limits.discard_alignment = 0;
 	blk_queue_max_discard_sectors(disk->queue, 0);
 	blk_queue_max_segment_size(disk->queue, UINT_MAX);
 	blk_queue_max_segments(disk->queue, USHRT_MAX);