Re: [PATCH v2 1/2] scsi: ufs: core: Add support to notify userspace of UniPro QoS events

[Bart Van Assche <bvanassche@acm.org>]

On 2/24/26 6:29 PM, Can Guo wrote:
> +What:		/sys/bus/platform/drivers/ufshcd/*/dme_qos_notification
> +What:		/sys/bus/platform/devices/*.ufs/dme_qos_notification
> +Date:		February 2026
> +Contact:	Can Guo <can.guo@oss.qualcomm.com>
> +Description:
> +		This attribute shows and clears the DME	Quality of Service
> +		notification from UFSHCI UECDME.
> +
> +		The attribute is read/write.

The above text is incomplete. It should explain that
dme_qos_notification is a bitfield, what the meaning of the bits in this
bitfield are, when this bitfield is updated, that the only value that
can be written into this bitfield is 0 and also what the effect of
writing 0 into this bitfield is.

> diff --git a/drivers/ufs/core/ufshcd-priv.h b/drivers/ufs/core/ufshcd-priv.h
> index 7d6d19361af9..14e8cb145f43 100644
> --- a/drivers/ufs/core/ufshcd-priv.h
> +++ b/drivers/ufs/core/ufshcd-priv.h
> @@ -446,4 +446,10 @@ static inline void ufs_rpmb_remove(struct ufs_hba *hba)
>   }
>   #endif
>   
> +static inline void sysfs_notify_dirent_safe(struct kernfs_node *sd)
> +{
> +	if (sd)
> +		sysfs_notify_dirent(sd);
> +}

This function is very short and is not used outside
drivers/ufs/core/ufshcd.c. Is it really needed to introduce this 
function? If this function is preserved, please consider moving it into
drivers/ufs/core/ufshcd.c.

 > @@ -11044,6 +11051,8 @@ int ufshcd_init(struct ufs_hba *hba, void 
__iomem *mmio_base, unsigned int irq)
 >   		goto out_disable;
 >
 >   	ufs_sysfs_add_nodes(hba->dev);
 > +	hba->dme_qos_sysfs_handle = sysfs_get_dirent(hba->dev->kobj.sd,
 > +						     "dme_qos_notification");
 >   	async_schedule(ufshcd_async_scan, hba);
 >
 >   	device_enable_async_suspend(dev);

Where is the sysfs_put() call that corresponds to the above 
sysfs_get_dirent() call?

> + * @dme_qos_notification: UFS host controller DME QoS notification

Please explain also here that this is a bitfield and where the meaning
of the bits in this bitfield are defined.

Thanks,

Bart.