On Fri, May 19, 2017 at 08:52:45PM +0800, Ming Lei wrote:

But I still think it may be better to move nvme_kill_queues() into
nvme_remove_dead_ctrl() as an improvement because during this small
window page cache can be used up by write application, and no writeback
can move on meantime.

Yes, I agree that's a better placement for it. I was just concerned
about the reasoning since it would also mean we're still stuck if an
IO timeout occurs while calling del_gendisk. So I'm okay with the patch
as-is, but I'll also look into my other concern.