[PATCH net-next] net: stmmac: dwc-qos: calibrate tegra with mdio bus idle

[PATCH net-next] net: stmmac: dwc-qos: calibrate tegra with mdio bus idle

[Russell King (Oracle)]

Thierry states that there are prerequists for Tegra's calibration
that should be met before starting calibration - both the RGMII and
MDIO interfaces should be idle.

This commit adds the necessary MII bus locking to ensure that the MDIO
interface is idle during calibration.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
---
 .../net/ethernet/stmicro/stmmac/dwmac-dwc-qos-eth.c  | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-dwc-qos-eth.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-dwc-qos-eth.c
index fa900b4991d0..09ae16e026eb 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-dwc-qos-eth.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-dwc-qos-eth.c
@@ -136,10 +136,11 @@ static int dwc_qos_probe(struct platform_device *pdev,
 #define AUTO_CAL_STATUS 0x880c
 #define  AUTO_CAL_STATUS_ACTIVE BIT(31)
 
-static void tegra_eqos_fix_speed(void *priv, int speed, unsigned int mode)
+static void tegra_eqos_fix_speed(void *bsp_priv, int speed, unsigned int mode)
 {
-	struct tegra_eqos *eqos = priv;
+	struct tegra_eqos *eqos = bsp_priv;
 	bool needs_calibration = false;
+	struct stmmac_priv *priv;
 	u32 value;
 	int err;
 
@@ -158,6 +159,11 @@ static void tegra_eqos_fix_speed(void *priv, int speed, unsigned int mode)
 	}
 
 	if (needs_calibration) {
+		priv = netdev_priv(dev_get_drvdata(eqos->dev));
+
+		/* Calibration should be done with the MDIO bus idle */
+		mutex_lock(&priv->mii->mdio_lock);
+
 		/* calibrate */
 		value = readl(eqos->regs + SDMEMCOMPPADCTRL);
 		value |= SDMEMCOMPPADCTRL_PAD_E_INPUT_OR_E_PWRD;
@@ -191,6 +197,8 @@ static void tegra_eqos_fix_speed(void *priv, int speed, unsigned int mode)
 		value = readl(eqos->regs + SDMEMCOMPPADCTRL);
 		value &= ~SDMEMCOMPPADCTRL_PAD_E_INPUT_OR_E_PWRD;
 		writel(value, eqos->regs + SDMEMCOMPPADCTRL);
+
+		mutex_unlock(&priv->mii->mdio_lock);
 	} else {
 		value = readl(eqos->regs + AUTO_CAL_CONFIG);
 		value &= ~AUTO_CAL_CONFIG_ENABLE;
-- 
2.30.2

Re: [PATCH net-next] net: stmmac: dwc-qos: calibrate tegra with mdio bus idle

[Maxime Chevallier]

Hello Russell,

On Tue, 22 Apr 2025 15:24:55 +0100
"Russell King (Oracle)" <rmk+kernel@armlinux.org.uk> wrote:

> Thierry states that there are prerequists for Tegra's calibration
> that should be met before starting calibration - both the RGMII and
> MDIO interfaces should be idle.
> 
> This commit adds the necessary MII bus locking to ensure that the MDIO
> interface is idle during calibration.
> 
> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>

[...]
	
> -static void tegra_eqos_fix_speed(void *priv, int speed, unsigned int mode)
> +static void tegra_eqos_fix_speed(void *bsp_priv, int speed, unsigned int mode)
>  {
> -	struct tegra_eqos *eqos = priv;
> +	struct tegra_eqos *eqos = bsp_priv;
>  	bool needs_calibration = false;
> +	struct stmmac_priv *priv;
>  	u32 value;
>  	int err;
>  
> @@ -158,6 +159,11 @@ static void tegra_eqos_fix_speed(void *priv, int speed, unsigned int mode)
>  	}
>  
>  	if (needs_calibration) {
> +		priv = netdev_priv(dev_get_drvdata(eqos->dev));
> +
> +		/* Calibration should be done with the MDIO bus idle */
> +		mutex_lock(&priv->mii->mdio_lock);

Can't priv->mii be NULL, if the PHY for that MAC is connected to
another MDIO bus for instance ?

Maxime

Re: [PATCH net-next] net: stmmac: dwc-qos: calibrate tegra with mdio bus idle

[Thierry Reding]

[-- Attachment #1: Type: text/plain, Size: 1792 bytes --]

On Tue, Apr 22, 2025 at 04:42:30PM +0200, Maxime Chevallier wrote:
> Hello Russell,
> 
> On Tue, 22 Apr 2025 15:24:55 +0100
> "Russell King (Oracle)" <rmk+kernel@armlinux.org.uk> wrote:
> 
> > Thierry states that there are prerequists for Tegra's calibration
> > that should be met before starting calibration - both the RGMII and
> > MDIO interfaces should be idle.
> > 
> > This commit adds the necessary MII bus locking to ensure that the MDIO
> > interface is idle during calibration.
> > 
> > Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
> 
> [...]
> 	
> > -static void tegra_eqos_fix_speed(void *priv, int speed, unsigned int mode)
> > +static void tegra_eqos_fix_speed(void *bsp_priv, int speed, unsigned int mode)
> >  {
> > -	struct tegra_eqos *eqos = priv;
> > +	struct tegra_eqos *eqos = bsp_priv;
> >  	bool needs_calibration = false;
> > +	struct stmmac_priv *priv;
> >  	u32 value;
> >  	int err;
> >  
> > @@ -158,6 +159,11 @@ static void tegra_eqos_fix_speed(void *priv, int speed, unsigned int mode)
> >  	}
> >  
> >  	if (needs_calibration) {
> > +		priv = netdev_priv(dev_get_drvdata(eqos->dev));
> > +
> > +		/* Calibration should be done with the MDIO bus idle */
> > +		mutex_lock(&priv->mii->mdio_lock);
> 
> Can't priv->mii be NULL, if the PHY for that MAC is connected to
> another MDIO bus for instance ?

Looking at the code, priv->mii will either be valid when mdio_bus_data
is non-NULL, or the driver probe will fail. mdio_bus_data depends on the
presence of the "mdio" child node in DT (via plat->mdio_node).

As far as I can tell this is always the case on Tegra devices, so I
don't think we need to worry about this case. If it ever wasn't valid,
I'd probably consider that a bug.

Thierry

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH net-next] net: stmmac: dwc-qos: calibrate tegra with mdio bus idle

[Thierry Reding]

[-- Attachment #1: Type: text/plain, Size: 623 bytes --]

On Tue, Apr 22, 2025 at 03:24:55PM +0100, Russell King (Oracle) wrote:
> Thierry states that there are prerequists for Tegra's calibration
> that should be met before starting calibration - both the RGMII and
> MDIO interfaces should be idle.
> 
> This commit adds the necessary MII bus locking to ensure that the MDIO
> interface is idle during calibration.
> 
> Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
> ---
>  .../net/ethernet/stmicro/stmmac/dwmac-dwc-qos-eth.c  | 12 ++++++++++--
>  1 file changed, 10 insertions(+), 2 deletions(-)

Acked-by: Thierry Reding <treding@nvidia.com>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH net-next] net: stmmac: dwc-qos: calibrate tegra with mdio bus idle

[patchwork-bot+netdevbpf]

Hello:

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

On Tue, 22 Apr 2025 15:24:55 +0100 you wrote:
> Thierry states that there are prerequists for Tegra's calibration
> that should be met before starting calibration - both the RGMII and
> MDIO interfaces should be idle.
> 
> This commit adds the necessary MII bus locking to ensure that the MDIO
> interface is idle during calibration.
> 
> [...]

Here is the summary with links:
  - [net-next] net: stmmac: dwc-qos: calibrate tegra with mdio bus idle
    https://git.kernel.org/netdev/net-next/c/87f43e6f06a2

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