TH1520 SoC is divided into several subsystems, shipping distinct reset
controllers with similar control logic. Let's make reset signal mapping
a data structure specific to one compatible to prepare for introduction
of more reset controllers in the future.
Signed-off-by: Yao Zi <ziyao@disroot.org>
---
drivers/reset/reset-th1520.c | 42 +++++++++++++++++++++++++-----------
1 file changed, 30 insertions(+), 12 deletions(-)
diff --git a/drivers/reset/reset-th1520.c b/drivers/reset/reset-th1520.c
index 14d964a9c6b6..2b65a95ed021 100644
--- a/drivers/reset/reset-th1520.c
+++ b/drivers/reset/reset-th1520.c
@@ -29,14 +29,20 @@
#define TH1520_HDMI_SW_MAIN_RST BIT(0)
#define TH1520_HDMI_SW_PRST BIT(1)
+struct th1520_reset_map {
+ u32 bit;
+ u32 reg;
+};
+
struct th1520_reset_priv {
struct reset_controller_dev rcdev;
struct regmap *map;
+ const struct th1520_reset_map *resets;
};
-struct th1520_reset_map {
- u32 bit;
- u32 reg;
+struct th1520_reset_data {
+ const struct th1520_reset_map *resets;
+ size_t num;
};
static const struct th1520_reset_map th1520_resets[] = {
@@ -90,7 +96,7 @@ static int th1520_reset_assert(struct reset_controller_dev *rcdev,
struct th1520_reset_priv *priv = to_th1520_reset(rcdev);
const struct th1520_reset_map *reset;
- reset = &th1520_resets[id];
+ reset = &priv->resets[id];
return regmap_update_bits(priv->map, reset->reg, reset->bit, 0);
}
@@ -101,7 +107,7 @@ static int th1520_reset_deassert(struct reset_controller_dev *rcdev,
struct th1520_reset_priv *priv = to_th1520_reset(rcdev);
const struct th1520_reset_map *reset;
- reset = &th1520_resets[id];
+ reset = &priv->resets[id];
return regmap_update_bits(priv->map, reset->reg, reset->bit,
reset->bit);
@@ -120,11 +126,14 @@ static const struct regmap_config th1520_reset_regmap_config = {
static int th1520_reset_probe(struct platform_device *pdev)
{
+ const struct th1520_reset_data *data;
struct device *dev = &pdev->dev;
struct th1520_reset_priv *priv;
void __iomem *base;
int ret;
+ data = device_get_match_data(dev);
+
priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
if (!priv)
return -ENOMEM;
@@ -138,22 +147,31 @@ static int th1520_reset_probe(struct platform_device *pdev)
if (IS_ERR(priv->map))
return PTR_ERR(priv->map);
- /* Initialize GPU resets to asserted state */
- ret = regmap_update_bits(priv->map, TH1520_GPU_RST_CFG,
- TH1520_GPU_RST_CFG_MASK, 0);
- if (ret)
- return ret;
+ if (of_device_is_compatible(dev->of_node, "thead,th1520-reset")) {
+ /* Initialize GPU resets to asserted state */
+ ret = regmap_update_bits(priv->map, TH1520_GPU_RST_CFG,
+ TH1520_GPU_RST_CFG_MASK, 0);
+ if (ret)
+ return ret;
+ }
priv->rcdev.owner = THIS_MODULE;
- priv->rcdev.nr_resets = ARRAY_SIZE(th1520_resets);
+ priv->rcdev.nr_resets = data->num;
priv->rcdev.ops = &th1520_reset_ops;
priv->rcdev.of_node = dev->of_node;
+ priv->resets = data->resets;
+
return devm_reset_controller_register(dev, &priv->rcdev);
}
+static const struct th1520_reset_data th1520_reset_data = {
+ .resets = th1520_resets,
+ .num = ARRAY_SIZE(th1520_resets),
+};
+
static const struct of_device_id th1520_reset_match[] = {
- { .compatible = "thead,th1520-reset" },
+ { .compatible = "thead,th1520-reset", .data = &th1520_reset_data },
{ /* sentinel */ }
};
MODULE_DEVICE_TABLE(of, th1520_reset_match);
--
2.50.1On Tue, Oct 14, 2025 at 01:10:30PM +0000, Yao Zi wrote:
> TH1520 SoC is divided into several subsystems, shipping distinct reset
> controllers with similar control logic. Let's make reset signal mapping
> a data structure specific to one compatible to prepare for introduction
> of more reset controllers in the future.
>
> Signed-off-by: Yao Zi <ziyao@disroot.org>
> ---
> drivers/reset/reset-th1520.c | 42 +++++++++++++++++++++++++-----------
> 1 file changed, 30 insertions(+), 12 deletions(-)
>
> diff --git a/drivers/reset/reset-th1520.c b/drivers/reset/reset-th1520.c
> index 14d964a9c6b6..2b65a95ed021 100644
> --- a/drivers/reset/reset-th1520.c
> +++ b/drivers/reset/reset-th1520.c
[snip]
> @@ -138,22 +147,31 @@ static int th1520_reset_probe(struct platform_device *pdev)
> if (IS_ERR(priv->map))
> return PTR_ERR(priv->map);
>
> - /* Initialize GPU resets to asserted state */
> - ret = regmap_update_bits(priv->map, TH1520_GPU_RST_CFG,
> - TH1520_GPU_RST_CFG_MASK, 0);
> - if (ret)
> - return ret;
> + if (of_device_is_compatible(dev->of_node, "thead,th1520-reset")) {
Is there a reason that there is a now a conditional check for the
compatible here?
Thanks,
Drew
On Wed, Oct 29, 2025 at 12:54:25PM +0000, Drew Fustini wrote:
> On Tue, Oct 14, 2025 at 01:10:30PM +0000, Yao Zi wrote:
> > TH1520 SoC is divided into several subsystems, shipping distinct reset
> > controllers with similar control logic. Let's make reset signal mapping
> > a data structure specific to one compatible to prepare for introduction
> > of more reset controllers in the future.
> >
> > Signed-off-by: Yao Zi <ziyao@disroot.org>
> > ---
> > drivers/reset/reset-th1520.c | 42 +++++++++++++++++++++++++-----------
> > 1 file changed, 30 insertions(+), 12 deletions(-)
> >
> > diff --git a/drivers/reset/reset-th1520.c b/drivers/reset/reset-th1520.c
> > index 14d964a9c6b6..2b65a95ed021 100644
> > --- a/drivers/reset/reset-th1520.c
> > +++ b/drivers/reset/reset-th1520.c
> [snip]
> > @@ -138,22 +147,31 @@ static int th1520_reset_probe(struct platform_device *pdev)
> > if (IS_ERR(priv->map))
> > return PTR_ERR(priv->map);
> >
> > - /* Initialize GPU resets to asserted state */
> > - ret = regmap_update_bits(priv->map, TH1520_GPU_RST_CFG,
> > - TH1520_GPU_RST_CFG_MASK, 0);
> > - if (ret)
> > - return ret;
> > + if (of_device_is_compatible(dev->of_node, "thead,th1520-reset")) {
>
> Is there a reason that there is a now a conditional check for the
> compatible here?
Yes, this regmap operation is for initializing GPU resets and thus
modifies TH1520_GPU_RST_CFG, which only applies for the VO reset
controller (with compatible "thead,th1520-reset") but not others, or
other unrelated resets could be unexpectedly asserted.
> Thanks,
> Drew
Regards,
Yao Zi
On Wed, Oct 29, 2025 at 03:13:46PM +0000, Yao Zi wrote:
> On Wed, Oct 29, 2025 at 12:54:25PM +0000, Drew Fustini wrote:
> > On Tue, Oct 14, 2025 at 01:10:30PM +0000, Yao Zi wrote:
> > > TH1520 SoC is divided into several subsystems, shipping distinct reset
> > > controllers with similar control logic. Let's make reset signal mapping
> > > a data structure specific to one compatible to prepare for introduction
> > > of more reset controllers in the future.
> > >
> > > Signed-off-by: Yao Zi <ziyao@disroot.org>
> > > ---
> > > drivers/reset/reset-th1520.c | 42 +++++++++++++++++++++++++-----------
> > > 1 file changed, 30 insertions(+), 12 deletions(-)
> > >
> > > diff --git a/drivers/reset/reset-th1520.c b/drivers/reset/reset-th1520.c
> > > index 14d964a9c6b6..2b65a95ed021 100644
> > > --- a/drivers/reset/reset-th1520.c
> > > +++ b/drivers/reset/reset-th1520.c
> > [snip]
> > > @@ -138,22 +147,31 @@ static int th1520_reset_probe(struct platform_device *pdev)
> > > if (IS_ERR(priv->map))
> > > return PTR_ERR(priv->map);
> > >
> > > - /* Initialize GPU resets to asserted state */
> > > - ret = regmap_update_bits(priv->map, TH1520_GPU_RST_CFG,
> > > - TH1520_GPU_RST_CFG_MASK, 0);
> > > - if (ret)
> > > - return ret;
> > > + if (of_device_is_compatible(dev->of_node, "thead,th1520-reset")) {
> >
> > Is there a reason that there is a now a conditional check for the
> > compatible here?
>
> Yes, this regmap operation is for initializing GPU resets and thus
> modifies TH1520_GPU_RST_CFG, which only applies for the VO reset
> controller (with compatible "thead,th1520-reset") but not others, or
> other unrelated resets could be unexpectedly asserted.
Thanks for the explanation.
Reviewed-by: Drew Fustini <fustini@kernel.org>
© 2016 - 2025 Red Hat, Inc.