[v4,4/4] scsi: core: Rework the code for dropping the LLD module reference

Instead of clearing the host template module pointer if the LLD kernel
module is being unloaded, set the 'drop_module_ref' SCSI device member.
This patch prepares for constifying the SCSI host template.

Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: John Garry <john.garry@huawei.com>
Cc: Mike Christie <michael.christie@oracle.com>
Cc: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/scsi/scsi_sysfs.c  | 7 +++----
 include/scsi/scsi_device.h | 1 +
 2 files changed, 4 insertions(+), 4 deletions(-)

On 13/09/2022 20:57, Bart Van Assche wrote:
> Instead of clearing the host template module pointer if the LLD kernel
> module is being unloaded, set the 'drop_module_ref' SCSI device member.
> This patch prepares for constifying the SCSI host template.
> 
> Cc: Christoph Hellwig <hch@lst.de>
> Cc: Ming Lei <ming.lei@redhat.com>
> Cc: Hannes Reinecke <hare@suse.de>
> Cc: John Garry <john.garry@huawei.com>
> Cc: Mike Christie <michael.christie@oracle.com>
> Cc: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
> Signed-off-by: Bart Van Assche <bvanassche@acm.org>
> ---
>   drivers/scsi/scsi_sysfs.c  | 7 +++----
>   include/scsi/scsi_device.h | 1 +
>   2 files changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/scsi/scsi_sysfs.c b/drivers/scsi/scsi_sysfs.c
> index 5d61f58399dc..822ae60a64b9 100644
> --- a/drivers/scsi/scsi_sysfs.c
> +++ b/drivers/scsi/scsi_sysfs.c
> @@ -454,7 +454,7 @@ static void scsi_device_dev_release_usercontext(struct work_struct *work)
>   
>   	sdev = container_of(work, struct scsi_device, ew.work);
>   
> -	mod = sdev->host->hostt->module;
> +	mod = sdev->drop_module_ref ? sdev->host->hostt->module : NULL;

I suppose that this works.

My reservation is that there were some concerns of current module 
referencing solution, so may not be better to build directly on it:

https://lore.kernel.org/linux-scsi/Ynt0aFMX+z%2FUhGJ2@infradead.org/

Thanks,
John

>   
>   	scsi_dh_release_device(sdev);
>   
> @@ -525,9 +525,8 @@ static void scsi_device_dev_release(struct device *dev)
>   {
>   	struct scsi_device *sdp = to_scsi_device(dev);
>   
> -	/* Set module pointer as NULL in case of module unloading */
> -	if (!try_module_get(sdp->host->hostt->module))
> -		sdp->host->hostt->module = NULL;
> +	if (try_module_get(sdp->host->hostt->module))
> +		sdp->drop_module_ref = true;
>   
>   	execute_in_process_context(scsi_device_dev_release_usercontext,
>   				   &sdp->ew);
> diff --git a/include/scsi/scsi_device.h b/include/scsi/scsi_device.h
> index 2493bd65351a..b03176b69056 100644
> --- a/include/scsi/scsi_device.h
> +++ b/include/scsi/scsi_device.h
> @@ -214,6 +214,7 @@ struct scsi_device {
>   					 * creation time */
>   	unsigned ignore_media_change:1; /* Ignore MEDIA CHANGE on resume */
>   	unsigned silence_suspend:1;	/* Do not print runtime PM related messages */
> +	unsigned drop_module_ref:1;
>   
>   	unsigned int queue_stopped;	/* request queue is quiesced */
>   	bool offline_already;		/* Device offline message logged */
> 
> .

On 9/14/22 02:52, John Garry wrote:
> On 13/09/2022 20:57, Bart Van Assche wrote:
>> -    mod = sdev->host->hostt->module;
>> +    mod = sdev->drop_module_ref ? sdev->host->hostt->module : NULL;
> 
> I suppose that this works.
> 
> My reservation is that there were some concerns of current module 
> referencing solution, so may not be better to build directly on it:
> 
> https://lore.kernel.org/linux-scsi/Ynt0aFMX+z%2FUhGJ2@infradead.org/

Hi John,

The above link points at a reply from Christoph Hellwig with the 
suggestion to use __module_get(). Implementing that suggestion would 
require to modify the kernel module implementation. My conclusion from 
reading the code in kernel/module/main.c is that increasing the module 
reference count after unloading has started currently has no effect. I 
have not found any code in the delete_module() system call 
implementation that waits for the reference count to drop to zero after 
try_stop_module() succeeded. I will look into this.

Thanks,

Bart.