bcache/md: Use raid stripe size

Now that we've got code for raid5/6 stripe awareness, bcache just needs
to know about the stripes and when writing partial stripes is expensive
- we probably don't want to enable this optimization for raid1 or 10,
even though they have stripes. So add a flag to queue_limits.

Signed-off-by: Kent Overstreet <kmo@daterainc.com>

diff --git a/block/blk-settings.c b/block/blk-settings.c
index 05e826793e4e36b2e6c8de29802674767e3bd4d8..5d21239bc8599feb3fa12fa4906ba57aaf1c553f 100644 (file)
--- a/block/blk-settings.c
+++ b/block/blk-settings.c
@@ -592,6 +592,10 @@ int blk_stack_limits(struct queue_limits *t, struct queue_limits *b,
                ret = -1;
        }
 
+       t->raid_partial_stripes_expensive =
+               max(t->raid_partial_stripes_expensive,
+                   b->raid_partial_stripes_expensive);
+
        /* Find lowest common alignment_offset */
        t->alignment_offset = lcm(t->alignment_offset, alignment)
                & (max(t->physical_block_size, t->io_min) - 1);

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 63ebef78df4ab38dd926cccf0a9788e81d640121..e363efcf2b76b5e1bc7aaca90788fb44533ec3d3 100644 (file)
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1134,6 +1134,12 @@ static int cached_dev_init(struct cached_dev *dc, unsigned block_size)
                hlist_add_head(&io->hash, dc->io_hash + RECENT_IO);
        }
 
+       dc->disk.stripe_size = q->limits.io_opt >> 9;
+
+       if (dc->disk.stripe_size)
+               dc->partial_stripes_expensive =
+                       q->limits.raid_partial_stripes_expensive;
+
        ret = bcache_device_init(&dc->disk, block_size,
                         dc->bdev->bd_part->nr_sects - dc->sb.data_offset);
        if (ret)

diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index eea63372e4d30533b2255159c8b428b2ad90acb3..1cfb22c025b6b5b8f02c1fe907e2810adb384a73 100644 (file)
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -6101,6 +6101,7 @@ static int run(struct mddev *mddev)
                blk_queue_io_min(mddev->queue, chunk_size);
                blk_queue_io_opt(mddev->queue, chunk_size *
                                 (conf->raid_disks - conf->max_degraded));
+               mddev->queue->limits.raid_partial_stripes_expensive = 1;
                /*
                 * We can only discard a whole stripe. It doesn't make sense to
                 * discard data disk but write parity disk

diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 02cb6f0ea71d52a09c3b2147f2bbc91997904650..0375654adb28cb423911aa7afccce6eb1f015e14 100644 (file)
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -291,6 +291,7 @@ struct queue_limits {
        unsigned char           discard_misaligned;
        unsigned char           cluster;
        unsigned char           discard_zeroes_data;
+       unsigned char           raid_partial_stripes_expensive;
 };
 
 struct request_queue {