[PATCH 14/14] nvme: use blk_mq_[un]quiesce_tagset

[Sagi Grimberg]

>>> This patch is causing the following crash at the boot and reverting it
>>> fixes the issue. This is next-20221121 kernel.
>> This patch can fix it.
>> https://lore.kernel.org/linux-nvme/20221116072711.1903536-1-hch@lst.de/
> 
> Yes,  But I'm a little curious how it happened.  Shakeel, do you
> have a genuine admin controller that does not have any I/O queues
> to start with, or did we have other setup time errors?  Can youpost the
> full dmesg?

Is this after splitting reset from probe?

If not, I think that probe schedules nvme_reset, which first
disables the controller, calling nvme_stop_queues unconditionally,
i.e. before the tagset was allocated.