| commit | 18b9837ea0dc3cf844c6c4196871ce91d047bddb | [log] [tgz] |
|---|---|---|
| author | NeilBrown <neilb@suse.de> | Sun Apr 01 23:48:38 2012 +1000 |
| committer | NeilBrown <neilb@suse.de> | Tue Apr 03 15:36:17 2012 +1000 |
| tree | de8be80b767d144405aeeaf7238c723f575b4b40 | |
| parent | 5220ea1e640869e70f894837678315c878c651fd [diff] |
md/raid5: fix handling of bad blocks during recovery. 1/ We can only treat a known-bad-block like a read-error if we have the data that belongs in that block. So fix that test. 2/ If we cannot recovery a stripe due to insufficient data, don't tell "md_done_sync" that the sync failed unless we really did fail something. If we successfully record bad blocks, that is success. Reported-by: "majianpeng" <majianpeng@gmail.com> Signed-off-by: NeilBrown <neilb@suse.de>