qed: Implement .bdrv_drain
authorFam Zheng <famz@redhat.com>
Thu, 27 Oct 2016 10:48:54 +0000 (12:48 +0200)
committerFam Zheng <famz@redhat.com>
Fri, 28 Oct 2016 13:50:18 +0000 (21:50 +0800)
commit6653a73d12fbac8c9b5c22bb21aef22759b28c1c
treee09b5811f6b18447ed60c0b0d21fdc8c79a50ff2
parentd42cf28837801cd1f835089fe9db2a42a1af55cd
qed: Implement .bdrv_drain

The "need_check_timer" is used to clear the "NEED_CHECK" flag in the
image header after a grace period once metadata update has finished. To
comply with the bdrv_drain semantics, we should make sure it remains
deleted once .bdrv_drain is called.

The change to qed_need_check_timer_cb is needed because bdrv_qed_drain
is called after s->bs has been drained, and should not operate on it;
instead it should operate on the BdrvChild-ren exclusively.  Doing so
is easy because QED does not have a bdrv_co_flush_to_os callback, hence
all that is needed to flush it is to ensure writes have reached the disk.

Based on commit df9a681dc9a (which however included some unrelated
hunks, possibly due to a merge failure or an overlooked squash).
The patch was reverted because at the time bdrv_qed_drain could call
qed_plug_allocating_write_reqs while an allocating write was queued.
This however is not possible anymore after the previous patch;
.bdrv_drain is only called after all writes have completed at the
QED level, and its purpose is to trigger metadata writes in bs->file.

Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <1477565348-5458-7-git-send-email-pbonzini@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
block/qed.c