Re: LXT971 driver in the phy lib

[Andy Fleming]

On May 8, 2006, at 11:05, Uwe Zeisberger wrote:

> Hello,
>
> I try to get an network interface running that has an LXT971A[1].
>
> If I apply the following patch, the target can detect the phy.
>
> diff --git a/drivers/net/phy/lxt.c b/drivers/net/phy/lxt.c
> index bef79e4..4c66fac 100644
> --- a/drivers/net/phy/lxt.c
> +++ b/drivers/net/phy/lxt.c
> @@ -137,9 +137,9 @@ static struct phy_driver lxt970_driver =
>  };
>
>  static struct phy_driver lxt971_driver = {
> -	.phy_id		= 0x0001378e,
> +	.phy_id		= 0x001378e0,
>  	.name		= "LXT971",
> -	.phy_id_mask	= 0x0fffffff,
> +	.phy_id_mask	= 0xfffffff0,
>  	.features	= PHY_BASIC_FEATURES,
>  	.flags		= PHY_HAS_INTERRUPT,
>  	.config_aneg	= genphy_config_aneg,


This looks good.

> According to
>
> 	http://www.intel.com/design/network/products/LAN/datashts/ 
> 24941402.pdf
>
> page 90f the id registers yield 0x001378eX (with X being current
> revision ID)
>
> 	uzeisberger@io:~/gsrc/linux-2.6$ git grep -i 1378e drivers/net/
> 	drivers/net/arm/at91_ether.h:#define MII_LXT971A_ID     0x001378E0
> 	drivers/net/e1000/e1000_hw.h:#define L1LXT971A_PHY_ID   0x001378E0
> 	drivers/net/fec.c:      .id = 0x0001378e,
> 	drivers/net/fec_8xx/fec_mii.c:   .id = 0x0001378e,
> 	drivers/net/phy/lxt.c:  .phy_id         = 0x0001378e,

The fec.c and fec_mii.c drivers used a shift by 4 to chop off the end  
of the ID.  When the PHY Layer switched to using a mask, not all of  
the drivers got properly switched.

> So both variants occur more than once.  (I only took a quick glance at
> the usage of these ids, but I think they all use it in the same way.
> That is, ID1 << 16 | ID2.)

The fec-style drivers do this:

        if(phy_info[i]->id == (fep->phy_id >> 4))

Anyway, your patch is correct.

> "My" phy reports 0x001378e2 and now I wonder if there are different
> chips out there with the same name.
>
> Can anybody explain this mismatch to me?  (Or point me to the right
> query for google.)


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/