md-cluster: clear another node's suspend_area after the copy is finished
authorGuoqing Jiang <gqjiang@suse.com>
Mon, 2 Jul 2018 08:26:24 +0000 (16:26 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 4 Oct 2018 00:00:47 +0000 (17:00 -0700)
commite0ccd2360a474d6505c5a5fa7ea5036a2eecd665
treeceac69b7d75948286af2ed7b1b90bb037748a342
parente70f938a605a82c96760ee2afb6f9eac6d0e7088
md-cluster: clear another node's suspend_area after the copy is finished

[ Upstream commit 010228e4a932ca1e8365e3b58c8e1e44c16ff793 ]

When one node leaves cluster or stops the resyncing
(resync or recovery) array, then other nodes need to
call recover_bitmaps to continue the unfinished task.

But we need to clear suspend_area later after other
nodes copy the resync information to their bitmap
(by call bitmap_copy_from_slot). Otherwise, all nodes
could write to the suspend_area even the suspend_area
is not handled by any node, because area_resyncing
returns 0 at the beginning of raid1_write_request.
Which means one node could write suspend_area while
another node is resyncing the same area, then data
could be inconsistent.

So let's clear suspend_area later to avoid above issue
with the protection of bm lock. Also it is straightforward
to clear suspend_area after nodes have copied the resync
info to bitmap.

Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Reviewed-by: NeilBrown <neilb@suse.com>
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Sasha Levin <alexander.levin@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/md/md-cluster.c