[PATCH V2 0/4] nvme: fix two kinds of IO hang from removing NSs

[Ming Lei]

On Wed, Jun 28, 2023 at 08:35:04AM -0600, Keith Busch wrote:
> On Wed, Jun 28, 2023 at 09:30:16AM +0800, Ming Lei wrote:
> > That may not be enough:
> > 
> > - What if nvme_sysfs_delete() is called from sysfs before the 1st check in
> > nvme_reset_work()?
> > 
> > - What if one pending nvme_dev_disable()<-nvme_timeout() comes after
> > the added nvme_unquiesce_io_queues() returns?
> 
> Okay, the following will handle both:
> 
> ---
> diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c
> index b027e5e3f4acb..c9224d39195e5 100644
> --- a/drivers/nvme/host/pci.c
> +++ b/drivers/nvme/host/pci.c
> @@ -2690,7 +2690,8 @@ static void nvme_reset_work(struct work_struct *work)
>  	if (dev->ctrl.state != NVME_CTRL_RESETTING) {
>  		dev_warn(dev->ctrl.device, "ctrl state %d is not RESETTING\n",
>  			 dev->ctrl.state);
> -		return;
> +		result = -ENODEV;
> +		goto out;
>  	}
>  
>  	/*
> @@ -2777,7 +2778,9 @@ static void nvme_reset_work(struct work_struct *work)
>  		 result);
>  	nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_DELETING);
>  	nvme_dev_disable(dev, true);
> +	nvme_sync_queues(&dev->ctrl);
>  	nvme_mark_namespaces_dead(&dev->ctrl);
> +	nvme_unquiesce_io_queues(&dev->ctrl);
>  	nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_DEAD);
>  }

This one looks better, but reset may not be scheduled successfully because
of removal, such as, the removal comes exactly before changing state
to NVME_CTRL_RESETTING.

Thanks,
Ming