failing device needs. Replacing a healthy device (e.g. with a larger
one) wants the opposite: an uncorrectable sector means data was already
lost, and the source should not be removed before that is understood.
-Today that user has to use device add + remove, about 30 times slower.
-With the flag, the replace is canceled with -EIO at the end of the
-block group where the first uncorrectable sector was found. The flag
-is persisted in the dev_replace item, so a resumed replace still
-aborts. Kernels without this patch reject the value with -EINVAL, so
-the option fails closed; unknown flags get -EOPNOTSUPP. A btrfs-progs
-patch adds "btrfs replace start --abort-on-uncorrectable".
+Today that user has to use device add + remove, about 30 times slower,
+to get the equivalent error-checking behavior. With the flag, the
+replace is canceled with -EIO at the end of the block group where the
+first uncorrectable sector was found. The flag is persisted in the
+dev_replace item, so a resumed replace still aborts. Kernels without
+this patch reject the value with -EINVAL, so the option fails closed;
+unknown flags get -EOPNOTSUPP. A btrfs-progs patch adds "btrfs replace
+start --abort-on-uncorrectable".
Patch 9 lets a single device filesystem with a replace in progress
mount again. Since v6.7 a scan does not register a device whose
sector by sector against a device-identifying prefill; KASAN. Patches
6, 9 and 10: an unmount mid-chunk with a subvolume drop running (single,
raid1, and a single device filesystem), with idle-unmount and no-unmount
-controls, and a scrub and a replace after a forced abort (unfixed: the
-scrub went through 5 more block groups, 1235 MiB in 76 s, and reported
-success; fixed: -EROFS within a second).
+controls, and a scrub and a replace after a forced abort.
Zygo Blaxell (10):
btrfs: commit the canceled state when a device replace fails or is