Re: [PATCH 1/3] hwspinlock: qcom: Add support for mmio usage to sfpb-mutex

[Christian Marangi]

On Thu, Jul 07, 2022 at 07:00:15PM +0800, Baolin Wang wrote:
> On Thu, Jul 7, 2022 at 9:30 AM Christian Marangi <ansuelsmth@xxxxxxxxx> wrote:
> >
> > Allow sfpb-mutex to use mmio in addition to syscon.
> >
> > Signed-off-by: Christian Marangi <ansuelsmth@xxxxxxxxx>
> > ---
> >  drivers/hwspinlock/qcom_hwspinlock.c | 32 ++++++++++++++++++++++------
> >  1 file changed, 25 insertions(+), 7 deletions(-)
> >
> > diff --git a/drivers/hwspinlock/qcom_hwspinlock.c b/drivers/hwspinlock/qcom_hwspinlock.c
> > index 364710966665..23c913095bd0 100644
> > --- a/drivers/hwspinlock/qcom_hwspinlock.c
> > +++ b/drivers/hwspinlock/qcom_hwspinlock.c
> > @@ -19,6 +19,11 @@
> >  #define QCOM_MUTEX_APPS_PROC_ID        1
> >  #define QCOM_MUTEX_NUM_LOCKS   32
> >
> > +struct qcom_hwspinlock_of_data {
> > +       u32 offset;
> > +       u32 stride;
> > +};
> > +
> >  static int qcom_hwspinlock_trylock(struct hwspinlock *lock)
> >  {
> >         struct regmap_field *field = lock->priv;
> > @@ -63,9 +68,20 @@ static const struct hwspinlock_ops qcom_hwspinlock_ops = {
> >         .unlock         = qcom_hwspinlock_unlock,
> >  };
> >
> > +static const struct qcom_hwspinlock_of_data of_sfpb_mutex = {
> > +       .offset = 0x4,
> > +       .stride = 0x4,
> > +};
> > +
> > +/* All modern platform has offset 0 and stride of 4k */
> > +static const struct qcom_hwspinlock_of_data of_tcsr_mutex = {
> > +       .offset = 0,
> > +       .stride = 0x1000,
> > +};
> > +
> >  static const struct of_device_id qcom_hwspinlock_of_match[] = {
> > -       { .compatible = "qcom,sfpb-mutex" },
> > -       { .compatible = "qcom,tcsr-mutex" },
> > +       { .compatible = "qcom,sfpb-mutex", .data = &of_sfpb_mutex },
> > +       { .compatible = "qcom,tcsr-mutex", .data = &of_tcsr_mutex },
> >         { }
> >  };
> >  MODULE_DEVICE_TABLE(of, qcom_hwspinlock_of_match);
> > @@ -101,7 +117,7 @@ static struct regmap *qcom_hwspinlock_probe_syscon(struct platform_device *pdev,
> >         return regmap;
> >  }
> >
> > -static const struct regmap_config tcsr_mutex_config = {
> > +static const struct regmap_config qcom_hwspinlock_mmio_config = {
> >         .reg_bits               = 32,
> >         .reg_stride             = 4,
> >         .val_bits               = 32,
> > @@ -112,18 +128,20 @@ static const struct regmap_config tcsr_mutex_config = {
> >  static struct regmap *qcom_hwspinlock_probe_mmio(struct platform_device *pdev,
> >                                                  u32 *offset, u32 *stride)
> >  {
> > +       const struct qcom_hwspinlock_of_data *data;
> >         struct device *dev = &pdev->dev;
> >         void __iomem *base;
> >
> > -       /* All modern platform has offset 0 and stride of 4k */
> > -       *offset = 0;
> > -       *stride = 0x1000;
> > +       data = of_device_get_match_data(dev);
> 
> Nit: better to validate the return value though this is a rare case.
> 
> if (!data)
>          return -ENODEV;
> 

Wonder if that can actually happen?

Looking at of_device_get_match_data() it can only return the data or
NULL if the match data is not defined but considering it should ALWAYS
be defined (or the driver can't work). But yhea should be a value to
check.

-- 
	Ansuel