diff mbox series

[net-next,v4,2/2] phy: mscc: Add support for RGMII delay configuration

Message ID 20230522122829.24945-3-harini.katakam@amd.com
State New
Headers
Series Add support for VSC85xx DT RGMII delays |

Commit Message

Harini Katakam May 22, 2023, 12:28 p.m. UTC 
  Add support for optional rx/tx-internal-delay-ps from devicetree.
- When rx/tx-internal-delay-ps is/are specified, these take priority
- When either is absent,
1) use 2ns for respective settings if rgmii-id/rxid/txid is/are present
2) use 0.2ns for respective settings if mode is rgmii

Signed-off-by: Harini Katakam <harini.katakam@amd.com>
---
v4:
Fix type of rx_delay and tx_delay
Reported by Simon Horman and
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202305140248.lh4LUw2j-lkp@intel.com/
v3 - Patch split:
- Use rx/tx-internal-delay-ps with phy_get_internal_delay
- Change RGMII delay selection precedence
- Update commit description and subject everywhere to say RGMII delays
instead of RGMII tuning.
 drivers/net/phy/mscc/mscc.h      |  2 ++
 drivers/net/phy/mscc/mscc_main.c | 35 +++++++++++++++++++++++++-------
 2 files changed, 30 insertions(+), 7 deletions(-)

Comments

Jakub Kicinski May 24, 2023, 3:32 a.m. UTC | #1 
On Mon, 22 May 2023 17:58:29 +0530 Harini Katakam wrote:
> Add support for optional rx/tx-internal-delay-ps from devicetree.
> - When rx/tx-internal-delay-ps is/are specified, these take priority
> - When either is absent,
> 1) use 2ns for respective settings if rgmii-id/rxid/txid is/are present
> 2) use 0.2ns for respective settings if mode is rgmii
> 
> Signed-off-by: Harini Katakam <harini.katakam@amd.com>

PHY folks, looks good?
Vladimir Oltean May 24, 2023, 10:09 a.m. UTC | #2 
On Mon, May 22, 2023 at 05:58:29PM +0530, Harini Katakam wrote:
> diff --git a/drivers/net/phy/mscc/mscc_main.c b/drivers/net/phy/mscc/mscc_main.c
> index 91010524e03d..9e856231e464 100644
> --- a/drivers/net/phy/mscc/mscc_main.c
> +++ b/drivers/net/phy/mscc/mscc_main.c
> @@ -525,17 +525,14 @@ static int vsc85xx_rgmii_set_skews(struct phy_device *phydev, u32 rgmii_cntl,
>  {
>  	u16 rgmii_rx_delay_pos = ffs(rgmii_rx_delay_mask) - 1;
>  	u16 rgmii_tx_delay_pos = ffs(rgmii_tx_delay_mask) - 1;
> +	struct vsc8531_private *vsc8531 = phydev->priv;
>  	u16 reg_val = 0;
>  	int rc;
>  
>  	mutex_lock(&phydev->lock);
>  
> -	if (phydev->interface == PHY_INTERFACE_MODE_RGMII_RXID ||
> -	    phydev->interface == PHY_INTERFACE_MODE_RGMII_ID)
> -		reg_val |= RGMII_CLK_DELAY_2_0_NS << rgmii_rx_delay_pos;
> -	if (phydev->interface == PHY_INTERFACE_MODE_RGMII_TXID ||
> -	    phydev->interface == PHY_INTERFACE_MODE_RGMII_ID)
> -		reg_val |= RGMII_CLK_DELAY_2_0_NS << rgmii_tx_delay_pos;
> +	reg_val |= vsc8531->rx_delay << rgmii_rx_delay_pos;
> +	reg_val |= vsc8531->tx_delay << rgmii_tx_delay_pos;

What about vsc8584_config_init() -> vsc85xx_rgmii_set_skews()? Who will
have configured rx_delay and tx_delay for that call path?

Isn't it better to absorb the device tree parsing logic into
vsc85xx_rgmii_set_skews(), I wonder?

And if we do that, is it still necessary to use struct vsc8531_private
as temporary storage space from vsc85xx_config_init() to vsc85xx_rgmii_set_skews(),
or will two local variables (rx_delay, tx_delay) serve that purpose just fine?

>  
>  	rc = phy_modify_paged(phydev, MSCC_PHY_PAGE_EXTENDED_2,
>  			      rgmii_cntl,
> @@ -1808,10 +1805,34 @@ static irqreturn_t vsc8584_handle_interrupt(struct phy_device *phydev)
>  	return IRQ_HANDLED;
>  }
>  
> +static const int vsc8531_internal_delay[] = {200, 800, 1100, 1700, 2000, 2300,
> +					     2600, 3400};

Could you please avoid intermingling this with the function code, and
move it at the top of mscc_main.c?

Also, vsc8531_internal_delay[] or vsc85xx_internal_delay[]? The values
are also good for VSC8572, the other caller of vsc85xx_rgmii_set_skews().

>  static int vsc85xx_config_init(struct phy_device *phydev)
>  {
> -	int rc, i, phy_id;
> +	int delay_size = ARRAY_SIZE(vsc8531_internal_delay);
>  	struct vsc8531_private *vsc8531 = phydev->priv;
> +	struct device *dev = &phydev->mdio.dev;
> +	int rc, i, phy_id;
> +
> +	vsc8531->rx_delay = phy_get_internal_delay(phydev, dev, &vsc8531_internal_delay[0],

You can just write "x" rather than "&x[0]".

> +						   delay_size, true);
> +	if (vsc8531->rx_delay < 0) {
> +		if (phydev->interface == PHY_INTERFACE_MODE_RGMII_RXID ||
> +		    phydev->interface == PHY_INTERFACE_MODE_RGMII_ID)
> +			vsc8531->rx_delay = RGMII_CLK_DELAY_2_0_NS;
> +		else
> +			vsc8531->rx_delay = RGMII_CLK_DELAY_0_2_NS;
> +	}
> +
> +	vsc8531->tx_delay = phy_get_internal_delay(phydev, dev, &vsc8531_internal_delay[0],
> +						   delay_size, false);
> +	if (vsc8531->tx_delay < 0) {
> +		if (phydev->interface == PHY_INTERFACE_MODE_RGMII_TXID ||
> +		    phydev->interface == PHY_INTERFACE_MODE_RGMII_ID)
> +			vsc8531->rx_delay = RGMII_CLK_DELAY_2_0_NS;
> +		else
> +			vsc8531->rx_delay = RGMII_CLK_DELAY_0_2_NS;
> +	}
>  
>  	rc = vsc85xx_default_config(phydev);
>  	if (rc)
> -- 
> 2.17.1
>
Harini Katakam May 24, 2023, 11:38 a.m. UTC | #3 
Hi Vladimir,

> -----Original Message-----
> From: Vladimir Oltean <vladimir.oltean@nxp.com>
> Sent: Wednesday, May 24, 2023 3:39 PM
> To: Katakam, Harini <harini.katakam@amd.com>
> Cc: andrew@lunn.ch; hkallweit1@gmail.com; linux@armlinux.org.uk;
> davem@davemloft.net; kuba@kernel.org; edumazet@google.com;
> pabeni@redhat.com; wsa+renesas@sang-engineering.com;
> simon.horman@corigine.com; netdev@vger.kernel.org; linux-
> kernel@vger.kernel.org; harinikatakamlinux@gmail.com; Simek, Michal
> <michal.simek@amd.com>; Pandey, Radhey Shyam
> <radhey.shyam.pandey@amd.com>
> Subject: Re: [PATCH net-next v4 2/2] phy: mscc: Add support for RGMII delay
> configuration
> 
> On Mon, May 22, 2023 at 05:58:29PM +0530, Harini Katakam wrote:
> > diff --git a/drivers/net/phy/mscc/mscc_main.c
> b/drivers/net/phy/mscc/mscc_main.c
> > index 91010524e03d..9e856231e464 100644
> > --- a/drivers/net/phy/mscc/mscc_main.c
> > +++ b/drivers/net/phy/mscc/mscc_main.c
> > @@ -525,17 +525,14 @@ static int vsc85xx_rgmii_set_skews(struct
> phy_device *phydev, u32 rgmii_cntl,
> >  {
> >  	u16 rgmii_rx_delay_pos = ffs(rgmii_rx_delay_mask) - 1;
> >  	u16 rgmii_tx_delay_pos = ffs(rgmii_tx_delay_mask) - 1;
> > +	struct vsc8531_private *vsc8531 = phydev->priv;
> >  	u16 reg_val = 0;
> >  	int rc;
> >
> >  	mutex_lock(&phydev->lock);
> >
> > -	if (phydev->interface == PHY_INTERFACE_MODE_RGMII_RXID ||
> > -	    phydev->interface == PHY_INTERFACE_MODE_RGMII_ID)
> > -		reg_val |= RGMII_CLK_DELAY_2_0_NS << rgmii_rx_delay_pos;
> > -	if (phydev->interface == PHY_INTERFACE_MODE_RGMII_TXID ||
> > -	    phydev->interface == PHY_INTERFACE_MODE_RGMII_ID)
> > -		reg_val |= RGMII_CLK_DELAY_2_0_NS << rgmii_tx_delay_pos;
> > +	reg_val |= vsc8531->rx_delay << rgmii_rx_delay_pos;
> > +	reg_val |= vsc8531->tx_delay << rgmii_tx_delay_pos;
> 
> What about vsc8584_config_init() -> vsc85xx_rgmii_set_skews()? Who will
> have configured rx_delay and tx_delay for that call path?

In my current series "rx_delay" and "tx_delay" would end up 0 and a
delay of 0.2 ns will be programmed (default for that field).
I guess if the phy-mode is RGMII/-ID on VSC8584, 2ns will not be programmed.
This will be a problem for any device not using vsc85xx_config_init

> 
> Isn't it better to absorb the device tree parsing logic into
> vsc85xx_rgmii_set_skews(), I wonder?

Yes, that is possible, let me respin. That will ensure existing values are not broken
for any VSC85xx user, if they do not have delay properties in the DT.

> 
> And if we do that, is it still necessary to use struct vsc8531_private
> as temporary storage space from vsc85xx_config_init() to
> vsc85xx_rgmii_set_skews(),
> or will two local variables (rx_delay, tx_delay) serve that purpose just fine?

No need of the structure member, local variables will do.

> 
> >
> >  	rc = phy_modify_paged(phydev, MSCC_PHY_PAGE_EXTENDED_2,
> >  			      rgmii_cntl,
> > @@ -1808,10 +1805,34 @@ static irqreturn_t
> vsc8584_handle_interrupt(struct phy_device *phydev)
> >  	return IRQ_HANDLED;
> >  }
> >
> > +static const int vsc8531_internal_delay[] = {200, 800, 1100, 1700, 2000,
> 2300,
> > +					     2600, 3400};
> 
> Could you please avoid intermingling this with the function code, and
> move it at the top of mscc_main.c?
> 
> Also, vsc8531_internal_delay[] or vsc85xx_internal_delay[]? The values
> are also good for VSC8572, the other caller of vsc85xx_rgmii_set_skews().

vsc85xx_internal_delay is more appropriate, will change.

Thanks for the review.

Regards,
Harini
diff mbox series

Patch

diff --git a/drivers/net/phy/mscc/mscc.h b/drivers/net/phy/mscc/mscc.h
index 9acee8759105..900bf37db9de 100644
--- a/drivers/net/phy/mscc/mscc.h
+++ b/drivers/net/phy/mscc/mscc.h
@@ -374,6 +374,8 @@  struct vsc8531_private {
 	 * package.
 	 */
 	unsigned int base_addr;
+	s32 rx_delay;
+	s32 tx_delay;
 
 #if IS_ENABLED(CONFIG_MACSEC)
 	/* MACsec fields:
diff --git a/drivers/net/phy/mscc/mscc_main.c b/drivers/net/phy/mscc/mscc_main.c
index 91010524e03d..9e856231e464 100644
--- a/drivers/net/phy/mscc/mscc_main.c
+++ b/drivers/net/phy/mscc/mscc_main.c
@@ -525,17 +525,14 @@  static int vsc85xx_rgmii_set_skews(struct phy_device *phydev, u32 rgmii_cntl,
 {
 	u16 rgmii_rx_delay_pos = ffs(rgmii_rx_delay_mask) - 1;
 	u16 rgmii_tx_delay_pos = ffs(rgmii_tx_delay_mask) - 1;
+	struct vsc8531_private *vsc8531 = phydev->priv;
 	u16 reg_val = 0;
 	int rc;
 
 	mutex_lock(&phydev->lock);
 
-	if (phydev->interface == PHY_INTERFACE_MODE_RGMII_RXID ||
-	    phydev->interface == PHY_INTERFACE_MODE_RGMII_ID)
-		reg_val |= RGMII_CLK_DELAY_2_0_NS << rgmii_rx_delay_pos;
-	if (phydev->interface == PHY_INTERFACE_MODE_RGMII_TXID ||
-	    phydev->interface == PHY_INTERFACE_MODE_RGMII_ID)
-		reg_val |= RGMII_CLK_DELAY_2_0_NS << rgmii_tx_delay_pos;
+	reg_val |= vsc8531->rx_delay << rgmii_rx_delay_pos;
+	reg_val |= vsc8531->tx_delay << rgmii_tx_delay_pos;
 
 	rc = phy_modify_paged(phydev, MSCC_PHY_PAGE_EXTENDED_2,
 			      rgmii_cntl,
@@ -1808,10 +1805,34 @@  static irqreturn_t vsc8584_handle_interrupt(struct phy_device *phydev)
 	return IRQ_HANDLED;
 }
 
+static const int vsc8531_internal_delay[] = {200, 800, 1100, 1700, 2000, 2300,
+					     2600, 3400};
 static int vsc85xx_config_init(struct phy_device *phydev)
 {
-	int rc, i, phy_id;
+	int delay_size = ARRAY_SIZE(vsc8531_internal_delay);
 	struct vsc8531_private *vsc8531 = phydev->priv;
+	struct device *dev = &phydev->mdio.dev;
+	int rc, i, phy_id;
+
+	vsc8531->rx_delay = phy_get_internal_delay(phydev, dev, &vsc8531_internal_delay[0],
+						   delay_size, true);
+	if (vsc8531->rx_delay < 0) {
+		if (phydev->interface == PHY_INTERFACE_MODE_RGMII_RXID ||
+		    phydev->interface == PHY_INTERFACE_MODE_RGMII_ID)
+			vsc8531->rx_delay = RGMII_CLK_DELAY_2_0_NS;
+		else
+			vsc8531->rx_delay = RGMII_CLK_DELAY_0_2_NS;
+	}
+
+	vsc8531->tx_delay = phy_get_internal_delay(phydev, dev, &vsc8531_internal_delay[0],
+						   delay_size, false);
+	if (vsc8531->tx_delay < 0) {
+		if (phydev->interface == PHY_INTERFACE_MODE_RGMII_TXID ||
+		    phydev->interface == PHY_INTERFACE_MODE_RGMII_ID)
+			vsc8531->rx_delay = RGMII_CLK_DELAY_2_0_NS;
+		else
+			vsc8531->rx_delay = RGMII_CLK_DELAY_0_2_NS;
+	}
 
 	rc = vsc85xx_default_config(phydev);
 	if (rc)