During a running dev replace operation, all write requests to
the live filesystem are duplicated to also write to the target
drive. Therefore btrfs_map_block() is changed to duplicate
stripes that are written to the source disk of a device replace
procedure to be written to the target disk as well.
Signed-off-by: Stefan Behrens <sbehr...@giantdisaster.de>
---
fs/btrfs/volumes.c | 50 +++++++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 49 insertions(+), 1 deletion(-)
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 208cc13..164f9b1 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -4049,6 +4049,9 @@ static int __btrfs_map_block(struct btrfs_fs_info
*fs_info, int rw,
int num_stripes;
int max_errors = 0;
struct btrfs_bio *bbio = NULL;
+ struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
+ int dev_replace_is_ongoing = 0;
+ int num_alloc_stripes;
read_lock(&em_tree->lock);
em = lookup_extent_mapping(em_tree, logical, *length);
@@ -4094,6 +4097,11 @@ static int __btrfs_map_block(struct btrfs_fs_info
*fs_info, int rw,
if (!bbio_ret)
goto out;
+ btrfs_dev_replace_lock(dev_replace);
+ dev_replace_is_ongoing = btrfs_dev_replace_is_ongoing(dev_replace);
+ if (!dev_replace_is_ongoing)
+ btrfs_dev_replace_unlock(dev_replace);
+
num_stripes = 1;
stripe_index = 0;
stripe_nr_orig = stripe_nr;
@@ -4160,7 +4168,10 @@ static int __btrfs_map_block(struct btrfs_fs_info
*fs_info, int rw,
}
BUG_ON(stripe_index >= map->num_stripes);
- bbio = kzalloc(btrfs_bio_size(num_stripes), GFP_NOFS);
+ num_alloc_stripes = num_stripes;
+ if (dev_replace_is_ongoing && (rw & (REQ_WRITE | REQ_DISCARD)))
+ num_alloc_stripes <<= 1;
+ bbio = kzalloc(btrfs_bio_size(num_alloc_stripes), GFP_NOFS);
if (!bbio) {
ret = -ENOMEM;
goto out;
@@ -4255,11 +4266,48 @@ static int __btrfs_map_block(struct btrfs_fs_info
*fs_info, int rw,
}
}
+ if (dev_replace_is_ongoing && (rw & (REQ_WRITE | REQ_DISCARD)) &&
+ dev_replace->tgtdev != NULL) {
+ int index_where_to_add;
+ u64 srcdev_devid = dev_replace->srcdev->devid;
+
+ /*
+ * duplicate the write operations while the dev replace
+ * procedure is running. Since the copying of the old disk
+ * to the new disk takes place at run time while the
+ * filesystem is mounted writable, the regular write
+ * operations to the old disk have to be duplicated to go
+ * to the new disk as well.
+ * Note that device->missing is handled by the caller, and
+ * that the write to the old disk is already set up in the
+ * stripes array.
+ */
+ index_where_to_add = num_stripes;
+ for (i = 0; i < num_stripes; i++) {
+ if (bbio->stripes[i].dev->devid == srcdev_devid) {
+ /* write to new disk, too */
+ struct btrfs_bio_stripe *new =
+ bbio->stripes + index_where_to_add;
+ struct btrfs_bio_stripe *old =
+ bbio->stripes + i;
+
+ new->physical = old->physical;
+ new->length = old->length;
+ new->dev = dev_replace->tgtdev;
+ index_where_to_add++;
+ max_errors++;
+ }
+ }
+ num_stripes = index_where_to_add;
+ }
+
*bbio_ret = bbio;
bbio->num_stripes = num_stripes;
bbio->max_errors = max_errors;
bbio->mirror_num = mirror_num;
out:
+ if (dev_replace_is_ongoing)
+ btrfs_dev_replace_unlock(dev_replace);
free_extent_map(em);
return ret;
}
--
1.8.0
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
