On 11/11/18 2:17 PM, Greg Kroah-Hartman wrote:
4.19-stable review patch. If anyone has any objections, please let me know.
------------------
From: Sagi Grimberg <sagi@xxxxxxxxxxx>
[ Upstream commit 2acf70ade79d26b97611a8df52eb22aa33814cd4 ]
Queue deletion is done asynchronous when the last reference on the queue
is dropped. Thus, in order to make sure we don't over allocate under a
connect/disconnect storm, we let queue deletion complete before making
forward progress.
However, given that we flush the system_wq from rdma_cm context which
runs from a workqueue context, we can have a circular locking complaint
[1]. Fix that by using a private workqueue for queue deletion.
Hi Greg,
You may want to drop this patch. A bug was discovered in this patch a
few days ago. I think not backporting this patch is better than
backporting it. See also the discussion at
http://lists.infradead.org/pipermail/linux-nvme/2018-November/020766.html.
A quote from that e-mail thread: "I think we need to revert 2acf70a
("nvmet-rdma: use a private workqueue for delete") altogether because it
never made any difference..".
Bart.
