[PATCH 2/3] nvme-fabrics: fix kernel crash while shutting down controller

[Ming Lei]

On Sun, Oct 27, 2024 at 10:32:05PM +0530, Nilay Shroff wrote:
> The nvme keep-alive operation, which executes at a periodic interval,
> could potentially sneak in while shutting down a fabric controller.
> This may lead to a race between the fabric controller admin queue
> destroy code path (invoked while shutting down controller) and hw/hctx
> queue dispatcher called from the nvme keep-alive async request queuing
> operation. This race could lead to the kernel crash shown below:
> 
> Call Trace:
>     autoremove_wake_function+0x0/0xbc (unreliable)
>     __blk_mq_sched_dispatch_requests+0x114/0x24c
>     blk_mq_sched_dispatch_requests+0x44/0x84
>     blk_mq_run_hw_queue+0x140/0x220
>     nvme_keep_alive_work+0xc8/0x19c [nvme_core]
>     process_one_work+0x200/0x4e0
>     worker_thread+0x340/0x504
>     kthread+0x138/0x140
>     start_kernel_thread+0x14/0x18
> 
> While shutting down fabric controller, if nvme keep-alive request sneaks
> in then it would be flushed off. The nvme_keep_alive_end_io function is
> then invoked to handle the end of the keep-alive operation which
> decrements the admin->q_usage_counter and assuming this is the last/only
> request in the admin queue then the admin->q_usage_counter becomes zero.
> If that happens then blk-mq destroy queue operation (blk_mq_destroy_
> queue()) which could be potentially running simultaneously on another
> cpu (as this is the controller shutdown code path) would forward
> progress and deletes the admin queue. So, now from this point onward
> we are not supposed to access the admin queue resources. However the
> issue here's that the nvme keep-alive thread running hw/hctx queue
> dispatch operation hasn't yet finished its work and so it could still
> potentially access the admin queue resource while the admin queue had
> been already deleted and that causes the above crash.
> 
> The above kernel crash is regression caused due to changes implemented
> in commit a54a93d0e359 ("nvme: move stopping keep-alive into
> nvme_uninit_ctrl()"). Ideally we should stop keep-alive at the very
> beggining of the controller shutdown code path so that it wouldn't
> sneak in during the shutdown operation. However we removed the keep
> alive stop operation from the beginning of the controller shutdown
> code path in commit a54a93d0e359 ("nvme: move stopping keep-alive into
> nvme_uninit_ctrl()") and that now created the possibility of keep-alive
> sneaking in and interfering with the shutdown operation and causing
> observed kernel crash. So to fix this crash, now we're adding back the
> keep-alive stop operation at very beginning of the fabric controller
> shutdown code path so that the actual controller shutdown opeation only
> begins after it's ensured that keep-alive operation is not in-flight and
> also it can't be scheduled in future.
> 
> Fixes: a54a93d0e359 ("nvme: move stopping keep-alive into nvme_uninit_ctrl()")
> Link: https://lore.kernel.org/all/196f4013-3bbf-43ff-98b4-9cb2a96c20c2@grimberg.me/#t
> Signed-off-by: Nilay Shroff <nilay at linux.ibm.com>
> ---
>  drivers/nvme/host/core.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
> index 5016f69e9a15..865c00ea19e3 100644
> --- a/drivers/nvme/host/core.c
> +++ b/drivers/nvme/host/core.c
> @@ -4648,6 +4648,11 @@ void nvme_stop_ctrl(struct nvme_ctrl *ctrl)
>  {
>  	nvme_mpath_stop(ctrl);
>  	nvme_auth_stop(ctrl);
> +	/*
> +	 * the transport driver may be terminating the admin tagset a little
> +	 * later on, so we cannot have the keep-alive work running
> +	 */
> +	nvme_stop_keep_alive(ctrl);
>  	nvme_stop_failfast_work(ctrl);
>  	flush_work(&ctrl->async_event_work);
>  	cancel_work_sync(&ctrl->fw_act_work);

The change looks fine.

IMO the `nvme_stop_keep_alive` in nvme_uninit_ctrl() may be moved to
entry of nvme_remove_admin_tag_set(), then this one in nvme_stop_ctrl()
can be saved?


thanks,
Ming