[PATCH net-next 0/3] net: stmmac: rk: second chunk of cleanups

[PATCH net-next 0/3] net: stmmac: rk: second chunk of cleanups

[Russell King (Oracle)]

Hi,

This series is a cut-down installment of the dwmac-rk cleanups, covering
up to the point that the AI review had its first issue with the patches.
Unfortunately, this means that we introduce ->init but do not add any
users for it yet. That will be in the next round once this has got
through AI review and merged.

 drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c | 120 ++++++++++++-------------
 1 file changed, 57 insertions(+), 63 deletions(-)

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!

[PATCH net-next 1/3] net: stmmac: rk: fix missing reset_control_put()

[Russell King (Oracle)]

rk_gmac_setup() delves into the PHY's DT node to retrieve its reset
control using of_reset_control_get(). However, it never releases it
when the driver is removed. Add reset_control_put() to rk_gmac_exit()
to clean this up.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
---
 drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
index 5f8d2031b97c..bc69cbb5a7d4 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
@@ -1784,6 +1784,8 @@ static void rk_gmac_exit(struct device *dev, void *bsp_priv_)
 
 	if (priv->plat->phy_node && bsp_priv->integrated_phy)
 		clk_put(bsp_priv->clk_phy);
+
+	reset_control_put(bsp_priv->phy_reset);
 }
 
 static int rk_gmac_probe(struct platform_device *pdev)
-- 
2.47.3

Re: [PATCH net-next 1/3] net: stmmac: rk: fix missing reset_control_put()

[Philipp Zabel]

On Mi, 2026-01-28 at 10:58 +0000, Russell King (Oracle) wrote:
> rk_gmac_setup() delves into the PHY's DT node to retrieve its reset
> control using of_reset_control_get(). However, it never releases it
> when the driver is removed. Add reset_control_put() to rk_gmac_exit()
> to clean this up.
> 
> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
> ---
>  drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> index 5f8d2031b97c..bc69cbb5a7d4 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> @@ -1784,6 +1784,8 @@ static void rk_gmac_exit(struct device *dev, void *bsp_priv_)
>  
>  	if (priv->plat->phy_node && bsp_priv->integrated_phy)
>  		clk_put(bsp_priv->clk_phy);
> +
> +	reset_control_put(bsp_priv->phy_reset);
>  }
>  
>  static int rk_gmac_probe(struct platform_device *pdev)

This is fine because the driver sets plat_dat->suspend, and so
rk_gmac_exit() is never called via stmmac_pltfr_exit() during suspend.

It does look a bit sketchy to release resources in the rk_gmac_exit()
counterpart to rk_gmac_init(), which never requested the resources,
though. Maybe use devm_add_action_or_reset() to register the release of
the reset during remove?

regards
Philipp

Re: [PATCH net-next 1/3] net: stmmac: rk: fix missing reset_control_put()

[Russell King (Oracle)]

On Wed, Jan 28, 2026 at 01:04:21PM +0100, Philipp Zabel wrote:
> On Mi, 2026-01-28 at 10:58 +0000, Russell King (Oracle) wrote:
> > rk_gmac_setup() delves into the PHY's DT node to retrieve its reset
> > control using of_reset_control_get(). However, it never releases it
> > when the driver is removed. Add reset_control_put() to rk_gmac_exit()
> > to clean this up.
> > 
> > Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
> > ---
> >  drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> > index 5f8d2031b97c..bc69cbb5a7d4 100644
> > --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> > +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> > @@ -1784,6 +1784,8 @@ static void rk_gmac_exit(struct device *dev, void *bsp_priv_)
> >  
> >  	if (priv->plat->phy_node && bsp_priv->integrated_phy)
> >  		clk_put(bsp_priv->clk_phy);
> > +
> > +	reset_control_put(bsp_priv->phy_reset);
> >  }
> >  
> >  static int rk_gmac_probe(struct platform_device *pdev)
> 
> This is fine because the driver sets plat_dat->suspend, and so
> rk_gmac_exit() is never called via stmmac_pltfr_exit() during suspend.
> 
> It does look a bit sketchy to release resources in the rk_gmac_exit()
> counterpart to rk_gmac_init(), which never requested the resources,
> though. Maybe use devm_add_action_or_reset() to register the release of
> the reset during remove?

Thanks, but I think a sense of proportion is required here. This
patch is the result of introducing the ->init() method, and AI
noticing that there was no cleanup of this resource. This was the
simplest way to implement that cleanup.

However, your review commit also applies to bsp_priv->clk_phy which
has the same problem - this also isn't obtained in rk_gmac_init(),
but in rk_gmac_clk_init().

Given that, and the fact that this entire series is already
considerably big (it was 21 patches, then 22, now 23, and with this
it's going to become 24 patches) I'm going to say that this issue
can be addressed at a later time.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!

Re: [PATCH net-next 1/3] net: stmmac: rk: fix missing reset_control_put()

[Philipp Zabel]

On Mi, 2026-01-28 at 12:19 +0000, Russell King (Oracle) wrote:
> On Wed, Jan 28, 2026 at 01:04:21PM +0100, Philipp Zabel wrote:
> > On Mi, 2026-01-28 at 10:58 +0000, Russell King (Oracle) wrote:
> > > rk_gmac_setup() delves into the PHY's DT node to retrieve its reset
> > > control using of_reset_control_get(). However, it never releases it
> > > when the driver is removed. Add reset_control_put() to rk_gmac_exit()
> > > to clean this up.
> > > 
> > > Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
> > > ---
> > >  drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c | 2 ++
> > >  1 file changed, 2 insertions(+)
> > > 
> > > diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> > > index 5f8d2031b97c..bc69cbb5a7d4 100644
> > > --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> > > +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> > > @@ -1784,6 +1784,8 @@ static void rk_gmac_exit(struct device *dev, void *bsp_priv_)
> > >  
> > >  	if (priv->plat->phy_node && bsp_priv->integrated_phy)
> > >  		clk_put(bsp_priv->clk_phy);
> > > +
> > > +	reset_control_put(bsp_priv->phy_reset);
> > >  }
> > >  
> > >  static int rk_gmac_probe(struct platform_device *pdev)
> > 
> > This is fine because the driver sets plat_dat->suspend, and so
> > rk_gmac_exit() is never called via stmmac_pltfr_exit() during suspend.
> > 
> > It does look a bit sketchy to release resources in the rk_gmac_exit()
> > counterpart to rk_gmac_init(), which never requested the resources,
> > though. Maybe use devm_add_action_or_reset() to register the release of
> > the reset during remove?
> 
> Thanks, but I think a sense of proportion is required here. This
> patch is the result of introducing the ->init() method, and AI
> noticing that there was no cleanup of this resource. This was the
> simplest way to implement that cleanup.
> 
> However, your review commit also applies to bsp_priv->clk_phy which
> has the same problem - this also isn't obtained in rk_gmac_init(),
> but in rk_gmac_clk_init().
>

> Given that, and the fact that this entire series is already
> considerably big (it was 21 patches, then 22, now 23, and with this
> it's going to become 24 patches) I'm going to say that this issue
> can be addressed at a later time.

No objection,

Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>

regards
Philipp

Re: [PATCH net-next 1/3] net: stmmac: rk: fix missing reset_control_put()

[Andrew Lunn]

On Wed, Jan 28, 2026 at 10:58:39AM +0000, Russell King (Oracle) wrote:
> rk_gmac_setup() delves into the PHY's DT node to retrieve its reset
> control using of_reset_control_get(). However, it never releases it
> when the driver is removed. Add reset_control_put() to rk_gmac_exit()
> to clean this up.
> 
> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>

Reviewed-by: Andrew Lunn <andrew@lunn.ch>

    Andrew

[PATCH net-next 2/3] net: stmmac: rk: add GMAC_CLK_xx constants, simplify RGMII definitions

[Russell King (Oracle)]

All the definitions of the RGMII related xxx_GMAC_CLK_xxx definitions
use the same field values to select the clock rate. Provide common
definitions for these field values, passing them in to a single macro
for each variant that generates the appropriate values for the speed
register.

No change to produced code on aarch64.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
---
 .../net/ethernet/stmicro/stmmac/dwmac-rk.c    | 108 ++++++++----------
 1 file changed, 45 insertions(+), 63 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
index bc69cbb5a7d4..976d42fd90df 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
@@ -91,6 +91,10 @@ struct rk_priv_data {
 	struct regmap *php_grf;
 };
 
+#define GMAC_CLK_DIV1_125M		0
+#define GMAC_CLK_DIV50_2_5M		2
+#define GMAC_CLK_DIV5_25M		3
+
 static int rk_set_reg_speed(struct rk_priv_data *bsp_priv,
 			    const struct rk_reg_speed_data *rsd,
 			    unsigned int reg, phy_interface_t interface,
@@ -299,9 +303,7 @@ static const struct rk_gmac_ops px30_ops = {
 #define RK3128_GMAC_SPEED_100M         GRF_BIT(10)
 #define RK3128_GMAC_RMII_CLK_25M       GRF_BIT(11)
 #define RK3128_GMAC_RMII_CLK_2_5M      GRF_CLR_BIT(11)
-#define RK3128_GMAC_CLK_125M           GRF_FIELD_CONST(13, 12, 0)
-#define RK3128_GMAC_CLK_25M            GRF_FIELD_CONST(13, 12, 3)
-#define RK3128_GMAC_CLK_2_5M           GRF_FIELD_CONST(13, 12, 2)
+#define RK3128_GMAC_CLK(val)           GRF_FIELD_CONST(13, 12, val)
 #define RK3128_GMAC_RMII_MODE          GRF_BIT(14)
 #define RK3128_GMAC_RMII_MODE_CLR      GRF_CLR_BIT(14)
 
@@ -325,9 +327,9 @@ static void rk3128_set_to_rmii(struct rk_priv_data *bsp_priv)
 }
 
 static const struct rk_reg_speed_data rk3128_reg_speed_data = {
-	.rgmii_10 = RK3128_GMAC_CLK_2_5M,
-	.rgmii_100 = RK3128_GMAC_CLK_25M,
-	.rgmii_1000 = RK3128_GMAC_CLK_125M,
+	.rgmii_10 = RK3128_GMAC_CLK(GMAC_CLK_DIV50_2_5M),
+	.rgmii_100 = RK3128_GMAC_CLK(GMAC_CLK_DIV5_25M),
+	.rgmii_1000 = RK3128_GMAC_CLK(GMAC_CLK_DIV1_125M),
 	.rmii_10 = RK3128_GMAC_RMII_CLK_2_5M | RK3128_GMAC_SPEED_10M,
 	.rmii_100 = RK3128_GMAC_RMII_CLK_25M | RK3128_GMAC_SPEED_100M,
 };
@@ -362,9 +364,7 @@ static const struct rk_gmac_ops rk3128_ops = {
 #define RK3228_GMAC_SPEED_100M		GRF_BIT(2)
 #define RK3228_GMAC_RMII_CLK_25M	GRF_BIT(7)
 #define RK3228_GMAC_RMII_CLK_2_5M	GRF_CLR_BIT(7)
-#define RK3228_GMAC_CLK_125M		GRF_FIELD_CONST(9, 8, 0)
-#define RK3228_GMAC_CLK_25M		GRF_FIELD_CONST(9, 8, 3)
-#define RK3228_GMAC_CLK_2_5M		GRF_FIELD_CONST(9, 8, 2)
+#define RK3228_GMAC_CLK(val)		GRF_FIELD_CONST(9, 8, val)
 #define RK3228_GMAC_RMII_MODE		GRF_BIT(10)
 #define RK3228_GMAC_RMII_MODE_CLR	GRF_CLR_BIT(10)
 #define RK3228_GMAC_TXCLK_DLY_ENABLE	GRF_BIT(0)
@@ -399,9 +399,9 @@ static void rk3228_set_to_rmii(struct rk_priv_data *bsp_priv)
 }
 
 static const struct rk_reg_speed_data rk3228_reg_speed_data = {
-	.rgmii_10 = RK3228_GMAC_CLK_2_5M,
-	.rgmii_100 = RK3228_GMAC_CLK_25M,
-	.rgmii_1000 = RK3228_GMAC_CLK_125M,
+	.rgmii_10 = RK3228_GMAC_CLK(GMAC_CLK_DIV50_2_5M),
+	.rgmii_100 = RK3228_GMAC_CLK(GMAC_CLK_DIV5_25M),
+	.rgmii_1000 = RK3228_GMAC_CLK(GMAC_CLK_DIV1_125M),
 	.rmii_10 = RK3228_GMAC_RMII_CLK_2_5M | RK3228_GMAC_SPEED_10M,
 	.rmii_100 = RK3228_GMAC_RMII_CLK_25M | RK3228_GMAC_SPEED_100M,
 };
@@ -440,9 +440,7 @@ static const struct rk_gmac_ops rk3228_ops = {
 #define RK3288_GMAC_SPEED_100M		GRF_BIT(10)
 #define RK3288_GMAC_RMII_CLK_25M	GRF_BIT(11)
 #define RK3288_GMAC_RMII_CLK_2_5M	GRF_CLR_BIT(11)
-#define RK3288_GMAC_CLK_125M		GRF_FIELD_CONST(13, 12, 0)
-#define RK3288_GMAC_CLK_25M		GRF_FIELD_CONST(13, 12, 3)
-#define RK3288_GMAC_CLK_2_5M		GRF_FIELD_CONST(13, 12, 2)
+#define RK3288_GMAC_CLK(val)		GRF_FIELD_CONST(13, 12, val)
 #define RK3288_GMAC_RMII_MODE		GRF_BIT(14)
 #define RK3288_GMAC_RMII_MODE_CLR	GRF_CLR_BIT(14)
 
@@ -474,9 +472,9 @@ static void rk3288_set_to_rmii(struct rk_priv_data *bsp_priv)
 }
 
 static const struct rk_reg_speed_data rk3288_reg_speed_data = {
-	.rgmii_10 = RK3288_GMAC_CLK_2_5M,
-	.rgmii_100 = RK3288_GMAC_CLK_25M,
-	.rgmii_1000 = RK3288_GMAC_CLK_125M,
+	.rgmii_10 = RK3288_GMAC_CLK(GMAC_CLK_DIV50_2_5M),
+	.rgmii_100 = RK3288_GMAC_CLK(GMAC_CLK_DIV5_25M),
+	.rgmii_1000 = RK3288_GMAC_CLK(GMAC_CLK_DIV1_125M),
 	.rmii_10 = RK3288_GMAC_RMII_CLK_2_5M | RK3288_GMAC_SPEED_10M,
 	.rmii_100 = RK3288_GMAC_RMII_CLK_25M | RK3288_GMAC_SPEED_100M,
 };
@@ -543,9 +541,7 @@ static const struct rk_gmac_ops rk3308_ops = {
 #define RK3328_GMAC_SPEED_100M		GRF_BIT(2)
 #define RK3328_GMAC_RMII_CLK_25M	GRF_BIT(7)
 #define RK3328_GMAC_RMII_CLK_2_5M	GRF_CLR_BIT(7)
-#define RK3328_GMAC_CLK_125M		GRF_FIELD_CONST(12, 11, 0)
-#define RK3328_GMAC_CLK_25M		GRF_FIELD_CONST(12, 11, 3)
-#define RK3328_GMAC_CLK_2_5M		GRF_FIELD_CONST(12, 11, 2)
+#define RK3328_GMAC_CLK(val)		GRF_FIELD_CONST(12, 11, val)
 #define RK3328_GMAC_RMII_MODE		GRF_BIT(9)
 #define RK3328_GMAC_RMII_MODE_CLR	GRF_CLR_BIT(9)
 #define RK3328_GMAC_TXCLK_DLY_ENABLE	GRF_BIT(0)
@@ -580,9 +576,9 @@ static void rk3328_set_to_rmii(struct rk_priv_data *bsp_priv)
 }
 
 static const struct rk_reg_speed_data rk3328_reg_speed_data = {
-	.rgmii_10 = RK3328_GMAC_CLK_2_5M,
-	.rgmii_100 = RK3328_GMAC_CLK_25M,
-	.rgmii_1000 = RK3328_GMAC_CLK_125M,
+	.rgmii_10 = RK3328_GMAC_CLK(GMAC_CLK_DIV50_2_5M),
+	.rgmii_100 = RK3328_GMAC_CLK(GMAC_CLK_DIV5_25M),
+	.rgmii_1000 = RK3328_GMAC_CLK(GMAC_CLK_DIV1_125M),
 	.rmii_10 = RK3328_GMAC_RMII_CLK_2_5M | RK3328_GMAC_SPEED_10M,
 	.rmii_100 = RK3328_GMAC_RMII_CLK_25M | RK3328_GMAC_SPEED_100M,
 };
@@ -632,9 +628,7 @@ static const struct rk_gmac_ops rk3328_ops = {
 #define RK3366_GMAC_SPEED_100M		GRF_BIT(7)
 #define RK3366_GMAC_RMII_CLK_25M	GRF_BIT(3)
 #define RK3366_GMAC_RMII_CLK_2_5M	GRF_CLR_BIT(3)
-#define RK3366_GMAC_CLK_125M		GRF_FIELD_CONST(5, 4, 0)
-#define RK3366_GMAC_CLK_25M		GRF_FIELD_CONST(5, 4, 3)
-#define RK3366_GMAC_CLK_2_5M		GRF_FIELD_CONST(5, 4, 2)
+#define RK3366_GMAC_CLK(val)		GRF_FIELD_CONST(5, 4, val)
 #define RK3366_GMAC_RMII_MODE		GRF_BIT(6)
 #define RK3366_GMAC_RMII_MODE_CLR	GRF_CLR_BIT(6)
 
@@ -666,9 +660,9 @@ static void rk3366_set_to_rmii(struct rk_priv_data *bsp_priv)
 }
 
 static const struct rk_reg_speed_data rk3366_reg_speed_data = {
-	.rgmii_10 = RK3366_GMAC_CLK_2_5M,
-	.rgmii_100 = RK3366_GMAC_CLK_25M,
-	.rgmii_1000 = RK3366_GMAC_CLK_125M,
+	.rgmii_10 = RK3366_GMAC_CLK(GMAC_CLK_DIV50_2_5M),
+	.rgmii_100 = RK3366_GMAC_CLK(GMAC_CLK_DIV5_25M),
+	.rgmii_1000 = RK3366_GMAC_CLK(GMAC_CLK_DIV1_125M),
 	.rmii_10 = RK3366_GMAC_RMII_CLK_2_5M | RK3366_GMAC_SPEED_10M,
 	.rmii_100 = RK3366_GMAC_RMII_CLK_25M | RK3366_GMAC_SPEED_100M,
 };
@@ -697,9 +691,7 @@ static const struct rk_gmac_ops rk3366_ops = {
 #define RK3368_GMAC_SPEED_100M		GRF_BIT(7)
 #define RK3368_GMAC_RMII_CLK_25M	GRF_BIT(3)
 #define RK3368_GMAC_RMII_CLK_2_5M	GRF_CLR_BIT(3)
-#define RK3368_GMAC_CLK_125M		GRF_FIELD_CONST(5, 4, 0)
-#define RK3368_GMAC_CLK_25M		GRF_FIELD_CONST(5, 4, 3)
-#define RK3368_GMAC_CLK_2_5M		GRF_FIELD_CONST(5, 4, 2)
+#define RK3368_GMAC_CLK(val)		GRF_FIELD_CONST(5, 4, val)
 #define RK3368_GMAC_RMII_MODE		GRF_BIT(6)
 #define RK3368_GMAC_RMII_MODE_CLR	GRF_CLR_BIT(6)
 
@@ -731,9 +723,9 @@ static void rk3368_set_to_rmii(struct rk_priv_data *bsp_priv)
 }
 
 static const struct rk_reg_speed_data rk3368_reg_speed_data = {
-	.rgmii_10 = RK3368_GMAC_CLK_2_5M,
-	.rgmii_100 = RK3368_GMAC_CLK_25M,
-	.rgmii_1000 = RK3368_GMAC_CLK_125M,
+	.rgmii_10 = RK3368_GMAC_CLK(GMAC_CLK_DIV50_2_5M),
+	.rgmii_100 = RK3368_GMAC_CLK(GMAC_CLK_DIV5_25M),
+	.rgmii_1000 = RK3368_GMAC_CLK(GMAC_CLK_DIV1_125M),
 	.rmii_10 = RK3368_GMAC_RMII_CLK_2_5M | RK3368_GMAC_SPEED_10M,
 	.rmii_100 = RK3368_GMAC_RMII_CLK_25M | RK3368_GMAC_SPEED_100M,
 };
@@ -762,9 +754,7 @@ static const struct rk_gmac_ops rk3368_ops = {
 #define RK3399_GMAC_SPEED_100M		GRF_BIT(7)
 #define RK3399_GMAC_RMII_CLK_25M	GRF_BIT(3)
 #define RK3399_GMAC_RMII_CLK_2_5M	GRF_CLR_BIT(3)
-#define RK3399_GMAC_CLK_125M		GRF_FIELD_CONST(5, 4, 0)
-#define RK3399_GMAC_CLK_25M		GRF_FIELD_CONST(5, 4, 3)
-#define RK3399_GMAC_CLK_2_5M		GRF_FIELD_CONST(5, 4, 2)
+#define RK3399_GMAC_CLK(val)		GRF_FIELD_CONST(5, 4, val)
 #define RK3399_GMAC_RMII_MODE		GRF_BIT(6)
 #define RK3399_GMAC_RMII_MODE_CLR	GRF_CLR_BIT(6)
 
@@ -796,9 +786,9 @@ static void rk3399_set_to_rmii(struct rk_priv_data *bsp_priv)
 }
 
 static const struct rk_reg_speed_data rk3399_reg_speed_data = {
-	.rgmii_10 = RK3399_GMAC_CLK_2_5M,
-	.rgmii_100 = RK3399_GMAC_CLK_25M,
-	.rgmii_1000 = RK3399_GMAC_CLK_125M,
+	.rgmii_10 = RK3399_GMAC_CLK(GMAC_CLK_DIV50_2_5M),
+	.rgmii_100 = RK3399_GMAC_CLK(GMAC_CLK_DIV5_25M),
+	.rgmii_1000 = RK3399_GMAC_CLK(GMAC_CLK_DIV1_125M),
 	.rmii_10 = RK3399_GMAC_RMII_CLK_2_5M | RK3399_GMAC_SPEED_10M,
 	.rmii_100 = RK3399_GMAC_RMII_CLK_25M | RK3399_GMAC_SPEED_100M,
 };
@@ -905,9 +895,7 @@ static const struct rk_gmac_ops rk3506_ops = {
 #define RK3528_GMAC1_CLK_RMII_DIV2	GRF_BIT(10)
 #define RK3528_GMAC1_CLK_RMII_DIV20	GRF_CLR_BIT(10)
 
-#define RK3528_GMAC1_CLK_RGMII_DIV1	GRF_FIELD_CONST(11, 10, 0)
-#define RK3528_GMAC1_CLK_RGMII_DIV5	GRF_FIELD_CONST(11, 10, 3)
-#define RK3528_GMAC1_CLK_RGMII_DIV50	GRF_FIELD_CONST(11, 10, 2)
+#define RK3528_GMAC1_CLK_RGMII(val)	GRF_FIELD_CONST(11, 10, val)
 
 #define RK3528_GMAC0_CLK_RMII_GATE	GRF_BIT(2)
 #define RK3528_GMAC0_CLK_RMII_NOGATE	GRF_CLR_BIT(2)
@@ -945,9 +933,9 @@ static const struct rk_reg_speed_data rk3528_gmac0_reg_speed_data = {
 };
 
 static const struct rk_reg_speed_data rk3528_gmac1_reg_speed_data = {
-	.rgmii_10 = RK3528_GMAC1_CLK_RGMII_DIV50,
-	.rgmii_100 = RK3528_GMAC1_CLK_RGMII_DIV5,
-	.rgmii_1000 = RK3528_GMAC1_CLK_RGMII_DIV1,
+	.rgmii_10 = RK3528_GMAC1_CLK_RGMII(GMAC_CLK_DIV50_2_5M),
+	.rgmii_100 = RK3528_GMAC1_CLK_RGMII(GMAC_CLK_DIV5_25M),
+	.rgmii_1000 = RK3528_GMAC1_CLK_RGMII(GMAC_CLK_DIV1_125M),
 	.rmii_10 = RK3528_GMAC1_CLK_RMII_DIV20,
 	.rmii_100 = RK3528_GMAC1_CLK_RMII_DIV2,
 };
@@ -1099,9 +1087,7 @@ static const struct rk_gmac_ops rk3568_ops = {
 #define RK3576_GMAC_CLK_RMII_DIV2		GRF_BIT(5)
 #define RK3576_GMAC_CLK_RMII_DIV20		GRF_CLR_BIT(5)
 
-#define RK3576_GMAC_CLK_RGMII_DIV1		GRF_FIELD_CONST(6, 5, 0)
-#define RK3576_GMAC_CLK_RGMII_DIV5		GRF_FIELD_CONST(6, 5, 3)
-#define RK3576_GMAC_CLK_RGMII_DIV50		GRF_FIELD_CONST(6, 5, 2)
+#define RK3576_GMAC_CLK_RGMII(val)		GRF_FIELD_CONST(6, 5, val)
 
 #define RK3576_GMAC_CLK_RMII_GATE		GRF_BIT(4)
 #define RK3576_GMAC_CLK_RMII_NOGATE		GRF_CLR_BIT(4)
@@ -1145,9 +1131,9 @@ static void rk3576_set_to_rmii(struct rk_priv_data *bsp_priv)
 }
 
 static const struct rk_reg_speed_data rk3578_reg_speed_data = {
-	.rgmii_10 = RK3576_GMAC_CLK_RGMII_DIV50,
-	.rgmii_100 = RK3576_GMAC_CLK_RGMII_DIV5,
-	.rgmii_1000 = RK3576_GMAC_CLK_RGMII_DIV1,
+	.rgmii_10 = RK3576_GMAC_CLK_RGMII(GMAC_CLK_DIV50_2_5M),
+	.rgmii_100 = RK3576_GMAC_CLK_RGMII(GMAC_CLK_DIV5_25M),
+	.rgmii_1000 = RK3576_GMAC_CLK_RGMII(GMAC_CLK_DIV1_125M),
 	.rmii_10 = RK3576_GMAC_CLK_RMII_DIV20,
 	.rmii_100 = RK3576_GMAC_CLK_RMII_DIV2,
 };
@@ -1223,12 +1209,8 @@ static const struct rk_gmac_ops rk3576_ops = {
 #define RK3588_GMA_CLK_RMII_DIV2(id)		GRF_BIT(5 * (id) + 2)
 #define RK3588_GMA_CLK_RMII_DIV20(id)		GRF_CLR_BIT(5 * (id) + 2)
 
-#define RK3588_GMAC_CLK_RGMII_DIV1(id)		\
-	(GRF_FIELD_CONST(3, 2, 0) << ((id) * 5))
-#define RK3588_GMAC_CLK_RGMII_DIV5(id)		\
-	(GRF_FIELD_CONST(3, 2, 3) << ((id) * 5))
-#define RK3588_GMAC_CLK_RGMII_DIV50(id)		\
-	(GRF_FIELD_CONST(3, 2, 2) << ((id) * 5))
+#define RK3588_GMAC_CLK_RGMII(id, val)		\
+	(GRF_FIELD_CONST(3, 2, val) << ((id) * 5))
 
 #define RK3588_GMAC_CLK_RMII_GATE(id)		GRF_BIT(5 * (id) + 1)
 #define RK3588_GMAC_CLK_RMII_NOGATE(id)		GRF_CLR_BIT(5 * (id) + 1)
@@ -1275,17 +1257,17 @@ static int rk3588_set_gmac_speed(struct rk_priv_data *bsp_priv,
 		if (interface == PHY_INTERFACE_MODE_RMII)
 			val = RK3588_GMA_CLK_RMII_DIV20(id);
 		else
-			val = RK3588_GMAC_CLK_RGMII_DIV50(id);
+			val = RK3588_GMAC_CLK_RGMII(id, GMAC_CLK_DIV50_2_5M);
 		break;
 	case 100:
 		if (interface == PHY_INTERFACE_MODE_RMII)
 			val = RK3588_GMA_CLK_RMII_DIV2(id);
 		else
-			val = RK3588_GMAC_CLK_RGMII_DIV5(id);
+			val = RK3588_GMAC_CLK_RGMII(id, GMAC_CLK_DIV5_25M);
 		break;
 	case 1000:
 		if (interface != PHY_INTERFACE_MODE_RMII)
-			val = RK3588_GMAC_CLK_RGMII_DIV1(id);
+			val = RK3588_GMAC_CLK_RGMII(id, GMAC_CLK_DIV1_125M);
 		else
 			goto err;
 		break;
-- 
2.47.3

Re: [PATCH net-next 2/3] net: stmmac: rk: add GMAC_CLK_xx constants, simplify RGMII definitions

[Maxime Chevallier]

Hi,

On 28/01/2026 11:58, Russell King (Oracle) wrote:
> All the definitions of the RGMII related xxx_GMAC_CLK_xxx definitions
> use the same field values to select the clock rate. Provide common
> definitions for these field values, passing them in to a single macro
> for each variant that generates the appropriate values for the speed
> register.
> 
> No change to produced code on aarch64.
> 
> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>

Maxime

[PATCH net-next 3/3] net: stmmac: rk: add SoC specific ->init() method

[Russell King (Oracle)]

Add a SoC specific init method.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
---
 drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
index 976d42fd90df..ea01c5965cbb 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
@@ -35,6 +35,7 @@ struct rk_reg_speed_data {
 };
 
 struct rk_gmac_ops {
+	int (*init)(struct rk_priv_data *bsp_priv);
 	void (*set_to_rgmii)(struct rk_priv_data *bsp_priv,
 			     int tx_delay, int rx_delay);
 	void (*set_to_rmii)(struct rk_priv_data *bsp_priv);
@@ -1618,6 +1619,15 @@ static struct rk_priv_data *rk_gmac_setup(struct platform_device *pdev,
 
 	bsp_priv->dev = dev;
 
+	if (ops->init) {
+		ret = ops->init(bsp_priv);
+		if (ret) {
+			reset_control_put(bsp_priv->phy_reset);
+			dev_err_probe(dev, ret, "failed to init BSP\n");
+			return ERR_PTR(ret);
+		}
+	}
+
 	return bsp_priv;
 }
 
-- 
2.47.3

Re: [PATCH net-next 0/3] net: stmmac: rk: second chunk of cleanups

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Wed, 28 Jan 2026 10:58:20 +0000 you wrote:
> Hi,
> 
> This series is a cut-down installment of the dwmac-rk cleanups, covering
> up to the point that the AI review had its first issue with the patches.
> Unfortunately, this means that we introduce ->init but do not add any
> users for it yet. That will be in the next round once this has got
> through AI review and merged.
> 
> [...]

Here is the summary with links:
  - [net-next,1/3] net: stmmac: rk: fix missing reset_control_put()
    https://git.kernel.org/netdev/net-next/c/1c249fc1e19e
  - [net-next,2/3] net: stmmac: rk: add GMAC_CLK_xx constants, simplify RGMII definitions
    https://git.kernel.org/netdev/net-next/c/f7aaa6fb1193
  - [net-next,3/3] net: stmmac: rk: add SoC specific ->init() method
    https://git.kernel.org/netdev/net-next/c/1639ba67bfb5

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html