-
由 majianpeng 提交于
If a resync is aborted cleanly, ->curr_resync is a reliable record of where we got up to. If there was an error it is less reliable but we always know that ->curr_resync_completed is safe. So add a flag MD_RECOVERY_ERROR to differentiate between these cases and set recovery_cp accordingly. Signed-off-by: NJianpeng Ma <majianpeng@gmail.com> Signed-off-by: NNeilBrown <neilb@suse.de>
0a19caab