diff mbox

[2/2,v2] MPC52xx FEC: be more conservative when setting MII_SPEED register

Message ID 1247780546-4426-2-git-send-email-wd@denx.de
State Changes Requested
Headers show

Commit Message

Wolfgang Denk July 16, 2009, 9:42 p.m. UTC
This patch adds error checking and prevents clobbering unrelated bits
(reserved bits or the DIS_PREAMBLE bit) when writing the MII_SPEED
register on MPC52xx systems.

Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: Kumar Gala <galak@kernel.crashing.org>
Cc: <netdev@vger.kernel.org>
---
 drivers/net/fec_mpc52xx.c     |    2 +-
 drivers/net/fec_mpc52xx_phy.c |   21 ++++++++++++++++++---
 2 files changed, 19 insertions(+), 4 deletions(-)

Comments

Grant Likely July 16, 2009, 10:48 p.m. UTC | #1
On Thu, Jul 16, 2009 at 3:42 PM, Wolfgang Denk<wd@denx.de> wrote:
> This patch adds error checking and prevents clobbering unrelated bits
> (reserved bits or the DIS_PREAMBLE bit) when writing the MII_SPEED
> register on MPC52xx systems.
>
> Signed-off-by: Wolfgang Denk <wd@denx.de>
> Cc: Grant Likely <grant.likely@secretlab.ca>
> Cc: Kumar Gala <galak@kernel.crashing.org>
> Cc: <netdev@vger.kernel.org>

Mostly good.  One comment below.  When it's resolved, feel free to add
my acked-by line.  Thanks for getting this done.

> @@ -105,8 +105,23 @@ static int mpc52xx_fec_mdio_probe(struct of_device *of,
>        dev_set_drvdata(dev, bus);
>
>        /* set MII speed */
> -       out_be32(&priv->regs->mii_speed,
> -               ((mpc5xxx_get_bus_frequency(of->node) >> 20) / 5) << 1);
> +       clock = mpc5xxx_get_bus_frequency(of->node);
> +       if (!clock) {
> +               dev_err(&of->dev, "could not determine IPS/IPB clock\n");
> +               goto out_unmap;
> +       }

Just thought of something.  If it cannot find the clock, then wouldn't
it be better to just use the maximum divider and print a warning
instead of bailing completely?  This goes for the other patch as well.

> +
> +       /* scale for a MII clock <= 2.5 MHz */
> +       speed = (clock + 2499999) / 2500000;
> +
> +       /* only 6 bits (25:30) available for MII speed */
> +       if (speed > 0x3F) {
> +               speed = 0x3F;
> +               dev_err(&of->dev, "MII clock (%d Hz) exceeds max (2.5 MHz)\n",
> +                       clock / speed);
> +       }
> +
> +       clrsetbits_be32(&priv->regs->mii_speed, 0x7E, speed << 1);
>
>        err = of_mdiobus_register(bus, np);
>        if (err)
> --
> 1.6.0.6
>
>
Wolfram Sang July 17, 2009, 9:47 a.m. UTC | #2
Hi Wolfgang,

On Thu, Jul 16, 2009 at 11:42:26PM +0200, Wolfgang Denk wrote:
> This patch adds error checking and prevents clobbering unrelated bits
> (reserved bits or the DIS_PREAMBLE bit) when writing the MII_SPEED
> register on MPC52xx systems.
> 
> Signed-off-by: Wolfgang Denk <wd@denx.de>
> Cc: Grant Likely <grant.likely@secretlab.ca>
> Cc: Kumar Gala <galak@kernel.crashing.org>
> Cc: <netdev@vger.kernel.org>
> ---
>  drivers/net/fec_mpc52xx.c     |    2 +-
>  drivers/net/fec_mpc52xx_phy.c |   21 ++++++++++++++++++---
>  2 files changed, 19 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/net/fec_mpc52xx.c b/drivers/net/fec_mpc52xx.c
> index cc78633..b69d440 100644
> --- a/drivers/net/fec_mpc52xx.c
> +++ b/drivers/net/fec_mpc52xx.c
> @@ -639,7 +639,7 @@ static void mpc52xx_fec_hw_init(struct net_device *dev)
>  	/* set phy speed.
>  	 * this can't be done in phy driver, since it needs to be called
>  	 * before fec stuff (even on resume) */
> -	out_be32(&fec->mii_speed, priv->mdio_speed);
> +	clrsetbits_be32(&fec->mii_speed, 0x7E, priv->mdio_speed);
>  }

In the probe-function when mdio_speed is set, there is still the old formula
used. Wouldn't that be better in sync?

>  
>  /**
> diff --git a/drivers/net/fec_mpc52xx_phy.c b/drivers/net/fec_mpc52xx_phy.c
> index 31e6d62..4c33dc5 100644
> --- a/drivers/net/fec_mpc52xx_phy.c
> +++ b/drivers/net/fec_mpc52xx_phy.c
> @@ -70,7 +70,7 @@ static int mpc52xx_fec_mdio_probe(struct of_device *of,
>  	struct mpc52xx_fec_mdio_priv *priv;
>  	struct resource res = {};
>  	int err;
> -	int i;
> +	int i, clock, speed;
>  
>  	bus = mdiobus_alloc();
>  	if (bus == NULL)
> @@ -105,8 +105,23 @@ static int mpc52xx_fec_mdio_probe(struct of_device *of,
>  	dev_set_drvdata(dev, bus);
>  
>  	/* set MII speed */
> -	out_be32(&priv->regs->mii_speed,
> -		((mpc5xxx_get_bus_frequency(of->node) >> 20) / 5) << 1);
> +	clock = mpc5xxx_get_bus_frequency(of->node);
> +	if (!clock) {
> +		dev_err(&of->dev, "could not determine IPS/IPB clock\n");
> +		goto out_unmap;
> +	}
> +
> +	/* scale for a MII clock <= 2.5 MHz */
> +	speed = (clock + 2499999) / 2500000;
> +
> +	/* only 6 bits (25:30) available for MII speed */
> +	if (speed > 0x3F) {
> +		speed = 0x3F;
> +		dev_err(&of->dev, "MII clock (%d Hz) exceeds max (2.5 MHz)\n",
> +			clock / speed);
> +	}
> +
> +	clrsetbits_be32(&priv->regs->mii_speed, 0x7E, speed << 1);
>  
>  	err = of_mdiobus_register(bus, np);
>  	if (err)
> -- 
> 1.6.0.6
> 
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev

Regards,

   Wolfram
Wolfgang Denk July 17, 2009, 12:25 p.m. UTC | #3
Dear Grant Likely,

In message <fa686aa40907161548n1658afaei8dce2d893222019c@mail.gmail.com> you wrote:
> 
> Mostly good.  One comment below.  When it's resolved, feel free to add
> my acked-by line.  Thanks for getting this done.
...
> Just thought of something.  If it cannot find the clock, then wouldn't
> it be better to just use the maximum divider and print a warning
> instead of bailing completely?  This goes for the other patch as well.

Will change this for both patches. Thanks.

Best regards,

Wolfgang Denk
Wolfgang Denk July 17, 2009, 12:35 p.m. UTC | #4
Dear Wolfram Sang,

In message <20090717094725.GC3150@pengutronix.de> you wrote:
> 
> In the probe-function when mdio_speed is set, there is still the old formula
> used. Wouldn't that be better in sync?

Good point. Next version of patch following soon.

Best regards,

Wolfgang Denk
diff mbox

Patch

diff --git a/drivers/net/fec_mpc52xx.c b/drivers/net/fec_mpc52xx.c
index cc78633..b69d440 100644
--- a/drivers/net/fec_mpc52xx.c
+++ b/drivers/net/fec_mpc52xx.c
@@ -639,7 +639,7 @@  static void mpc52xx_fec_hw_init(struct net_device *dev)
 	/* set phy speed.
 	 * this can't be done in phy driver, since it needs to be called
 	 * before fec stuff (even on resume) */
-	out_be32(&fec->mii_speed, priv->mdio_speed);
+	clrsetbits_be32(&fec->mii_speed, 0x7E, priv->mdio_speed);
 }
 
 /**
diff --git a/drivers/net/fec_mpc52xx_phy.c b/drivers/net/fec_mpc52xx_phy.c
index 31e6d62..4c33dc5 100644
--- a/drivers/net/fec_mpc52xx_phy.c
+++ b/drivers/net/fec_mpc52xx_phy.c
@@ -70,7 +70,7 @@  static int mpc52xx_fec_mdio_probe(struct of_device *of,
 	struct mpc52xx_fec_mdio_priv *priv;
 	struct resource res = {};
 	int err;
-	int i;
+	int i, clock, speed;
 
 	bus = mdiobus_alloc();
 	if (bus == NULL)
@@ -105,8 +105,23 @@  static int mpc52xx_fec_mdio_probe(struct of_device *of,
 	dev_set_drvdata(dev, bus);
 
 	/* set MII speed */
-	out_be32(&priv->regs->mii_speed,
-		((mpc5xxx_get_bus_frequency(of->node) >> 20) / 5) << 1);
+	clock = mpc5xxx_get_bus_frequency(of->node);
+	if (!clock) {
+		dev_err(&of->dev, "could not determine IPS/IPB clock\n");
+		goto out_unmap;
+	}
+
+	/* scale for a MII clock <= 2.5 MHz */
+	speed = (clock + 2499999) / 2500000;
+
+	/* only 6 bits (25:30) available for MII speed */
+	if (speed > 0x3F) {
+		speed = 0x3F;
+		dev_err(&of->dev, "MII clock (%d Hz) exceeds max (2.5 MHz)\n",
+			clock / speed);
+	}
+
+	clrsetbits_be32(&priv->regs->mii_speed, 0x7E, speed << 1);
 
 	err = of_mdiobus_register(bus, np);
 	if (err)