Re: [PATCH 03/10] ufs: host: Add common Hibern8 TX FSM polling helper

[Stanley Jhu]

On Wed, Aug 26, 2026 at 03:40:41PM +0200, Larisa Grigore wrote:
> [...]
> diff --git a/drivers/ufs/host/ufshcd-pltfrm.c b/drivers/ufs/host/ufshcd-pltfrm.c
> index c2dafb583cf5..034ee988924a 100644
> --- a/drivers/ufs/host/ufshcd-pltfrm.c
> +++ b/drivers/ufs/host/ufshcd-pltfrm.c
> @@ -563,6 +564,84 @@ void ufshcd_pltfrm_remove(struct platform_device *pdev)
>  }
>  EXPORT_SYMBOL_GPL(ufshcd_pltfrm_remove);
>
> +/**
> + * ufshcd_check_hibern8 - Check if all TX lanes entered Hibern8 state
> + * @hba: host controller instance
> + * @num_lanes: number of TX lanes to check
> + * @timeout_ms: timeout in milliseconds for all lanes
> + *
> + * Return: 0 on success, negative errno on failure.
> + */
> +int ufshcd_check_hibern8(struct ufs_hba *hba, unsigned int num_lanes,
> +                      unsigned int timeout_ms)

Hi Larisa,

Thanks for the great cleanup!

Would you consider moving this helper to drivers/ufs/core/ufshcd.c and
renaming it to ufshcd_dme_check_tx_hibern8()?

Since ufshcd-pltfrm.c primarily handles platform device glue and DT parsing,
polling M-PHY states via DME commands would fit more naturally in the UFS
core. This also keeps layering clean and allows non-platform controllers to
reuse it.

Additionally, the ufshcd_dme_ prefix would align nicely with existing DME
conventions in ufshcd.c while avoiding potential confusion with
ufshcd_is_link_hibern8().

Thank you,
Stanley