core_reset is not an optional property for the platforms supported in
upstream. Only for the non-upstreamed legacy platforms it is optional.
But somehow a few of the upstreamed platforms do not pass this property
by mistake.

So clarify the comment to make it clear that even though core_reset is
required, it is kept as optional to support the DTs that do not pass this
property.

Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
---
 drivers/ufs/host/ufs-qcom.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/ufs/host/ufs-qcom.c b/drivers/ufs/host/ufs-qcom.c
index 39eef470f8fa..32760506dfeb 100644
--- a/drivers/ufs/host/ufs-qcom.c
+++ b/drivers/ufs/host/ufs-qcom.c
@@ -1027,7 +1027,11 @@ static int ufs_qcom_init(struct ufs_hba *hba)
 	host->hba = hba;
 	ufshcd_set_variant(hba, host);
 
-	/* Setup the optional reset control of HCI */
+	/*
+	 * Even though core_reset is required on all platforms, some DTs never
+	 * passed this property. So we have to keep it optional for supporting
+	 * them.
+	 */
 	host->core_reset = devm_reset_control_get_optional(hba->dev, "rst");
 	if (IS_ERR(host->core_reset)) {
 		err = dev_err_probe(dev, PTR_ERR(host->core_reset),

-- 
2.25.1

On Mon, Jan 29, 2024 at 01:22:05PM +0530, Manivannan Sadhasivam wrote:
> core_reset is not an optional property for the platforms supported in
> upstream. Only for the non-upstreamed legacy platforms it is optional.
> But somehow a few of the upstreamed platforms do not pass this property
> by mistake.
> 
> So clarify the comment to make it clear that even though core_reset is
> required, it is kept as optional to support the DTs that do not pass this
> property.
> 
> Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
> ---
>  drivers/ufs/host/ufs-qcom.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/ufs/host/ufs-qcom.c b/drivers/ufs/host/ufs-qcom.c
> index 39eef470f8fa..32760506dfeb 100644
> --- a/drivers/ufs/host/ufs-qcom.c
> +++ b/drivers/ufs/host/ufs-qcom.c
> @@ -1027,7 +1027,11 @@ static int ufs_qcom_init(struct ufs_hba *hba)
>  	host->hba = hba;
>  	ufshcd_set_variant(hba, host);
>  
> -	/* Setup the optional reset control of HCI */
> +	/*
> +	 * Even though core_reset is required on all platforms, some DTs never
> +	 * passed this property. So we have to keep it optional for supporting
> +	 * them.
> +	 */

Any desire to print a warning if !host->core_reset? I'll defer to
Qualcomm to review since they can confirm the accuracy past Can's
comment, but this looks good to me for what its worth.

>  	host->core_reset = devm_reset_control_get_optional(hba->dev, "rst");
>  	if (IS_ERR(host->core_reset)) {
>  		err = dev_err_probe(dev, PTR_ERR(host->core_reset),
> 
> -- 
> 2.25.1
> 
>

On Mon, Jan 29, 2024 at 02:57:20PM -0600, Andrew Halaney wrote:
> On Mon, Jan 29, 2024 at 01:22:05PM +0530, Manivannan Sadhasivam wrote:
> > core_reset is not an optional property for the platforms supported in
> > upstream. Only for the non-upstreamed legacy platforms it is optional.
> > But somehow a few of the upstreamed platforms do not pass this property
> > by mistake.
> > 
> > So clarify the comment to make it clear that even though core_reset is
> > required, it is kept as optional to support the DTs that do not pass this
> > property.
> > 
> > Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
> > ---
> >  drivers/ufs/host/ufs-qcom.c | 6 +++++-
> >  1 file changed, 5 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/ufs/host/ufs-qcom.c b/drivers/ufs/host/ufs-qcom.c
> > index 39eef470f8fa..32760506dfeb 100644
> > --- a/drivers/ufs/host/ufs-qcom.c
> > +++ b/drivers/ufs/host/ufs-qcom.c
> > @@ -1027,7 +1027,11 @@ static int ufs_qcom_init(struct ufs_hba *hba)
> >  	host->hba = hba;
> >  	ufshcd_set_variant(hba, host);
> >  
> > -	/* Setup the optional reset control of HCI */
> > +	/*
> > +	 * Even though core_reset is required on all platforms, some DTs never
> > +	 * passed this property. So we have to keep it optional for supporting
> > +	 * them.
> > +	 */
> 
> Any desire to print a warning if !host->core_reset? I'll defer to
> Qualcomm to review since they can confirm the accuracy past Can's
> comment, but this looks good to me for what its worth.
> 

My only worry is that the existing users of the legacy DTs will get annoyed by
the warning. And I'm not sure if we can do that.

- Mani

-- 
மணிவண்ணன் சதாசிவம்