Re: [PATCH v3 09/13] net: Rework ucc_geth driver to use of_mdio infrastructure

[Grant Likely <grant.likely@secretlab.ca>]

On Sun, Apr 26, 2009 at 3:00 AM, Joakim Tjernlund
<Joakim.Tjernlund@transmode.se> wrote:
> Change in fixed link case, see inline.
>
>>
>> From: Grant Likely <grant.likely@secretlab.ca>
>>
>> This patch simplifies the driver by making use of more common code.
>>
>> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
>> Acked-by: Andy Fleming <afleming@freescale.com>
>> ---
>>
>>  drivers/net/ucc_geth.c |   47
> ++++++++++++-----------------------------------
>>  drivers/net/ucc_geth.h |    2 +-
>>  2 files changed, 13 insertions(+), 36 deletions(-)
>>
>>
>> diff --git a/drivers/net/ucc_geth.c b/drivers/net/ucc_geth.c
>> index d3f39e8..195b490 100644
>> --- a/drivers/net/ucc_geth.c
>> +++ b/drivers/net/ucc_geth.c
>> @@ -27,6 +27,7 @@
>>  #include <linux/mii.h>
>>  #include <linux/phy.h>
>>  #include <linux/workqueue.h>
>> +#include <linux/of_mdio.h>
>>  #include <linux/of_platform.h>
>>
>>  #include <asm/uaccess.h>
>> @@ -1542,12 +1543,14 @@ static int init_phy(struct net_device *dev)
>>     priv->oldspeed = 0;
>>     priv->oldduplex = -1;
>>
>> -   phydev = phy_connect(dev, ug_info->phy_bus_id, &adjust_link, 0,
>> -              priv->phy_interface);
>> +   if (!ug_info->phy_node)
>> +      return 0;
>>
>> -   if (IS_ERR(phydev)) {
>> +   phydev = of_phy_connect(dev, ug_info->phy_node, &adjust_link, 0,
>> +            priv->phy_interface);
>> +   if (!phydev) {
>>        printk("%s: Could not attach to PHY\n", dev->name);
>> -      return PTR_ERR(phydev);
>> +      return -ENODEV;
>>     }
>>
>>     phydev->supported &= (ADVERTISED_10baseT_Half |
>> @@ -3519,14 +3522,12 @@ static int ucc_geth_probe(struct of_device*
> ofdev, const struct of_device_id *ma
>>  {
>>     struct device *device = &ofdev->dev;
>>     struct device_node *np = ofdev->node;
>> -   struct device_node *mdio;
>>     struct net_device *dev = NULL;
>>     struct ucc_geth_private *ugeth = NULL;
>>     struct ucc_geth_info *ug_info;
>>     struct resource res;
>>     struct device_node *phy;
>>     int err, ucc_num, max_speed = 0;
>> -   const phandle *ph;
>>     const u32 *fixed_link;
>>     const unsigned int *prop;
>>     const char *sprop;
>> @@ -3626,45 +3627,21 @@ static int ucc_geth_probe(struct of_device*
> ofdev, const struct of_device_id *ma
>>     ug_info->uf_info.irq = irq_of_parse_and_map(np, 0);
>>     fixed_link = of_get_property(np, "fixed-link", NULL);
>>     if (fixed_link) {
>> -      snprintf(ug_info->phy_bus_id, sizeof(ug_info->phy_bus_id),
>> -          PHY_ID_FMT, "0", fixed_link[0]);
>>        phy = NULL;
>
> phy assigned to NULL ...
>
>>     } else {
>> -      char bus_name[MII_BUS_ID_SIZE];
>> -
>> -      ph = of_get_property(np, "phy-handle", NULL);
>> -      phy = of_find_node_by_phandle(*ph);
>> -
>> +      phy = of_parse_phandle(np, "phy-handle", 0);
>>        if (phy == NULL)
>>           return -ENODEV;
>> -
>> -      /* set the PHY address */
>> -      prop = of_get_property(phy, "reg", NULL);
>> -      if (prop == NULL)
>> -         return -1;
>> -
>> -      /* Set the bus id */
>> -      mdio = of_get_parent(phy);
>> -
>> -      if (mdio == NULL)
>> -         return -ENODEV;
>> -
>> -      err = of_address_to_resource(mdio, 0, &res);
>> -
>> -      if (err) {
>> -         of_node_put(mdio);
>> -         return err;
>> -      }
>> -      fsl_pq_mdio_bus_name(bus_name, mdio);
>> -      of_node_put(mdio);
>> -      snprintf(ug_info->phy_bus_id, sizeof(ug_info->phy_bus_id),
>> -         "%s:%02x", bus_name, *prop);
>>     }
>> +   ug_info->phy_node = phy;
>>
>>     /* get the phy interface type, or default to MII */
>>     prop = of_get_property(np, "phy-connection-type", NULL);
>>     if (!prop) {
>>        /* handle interface property present in old trees */
>> +      if (!phy)
>> +         return -ENODEV;
>
> .. here an error is returned. Is this intentional?

Yes, I did that intentionally because it looked like a bug to me
because of_get_property() is passed the phy value.  However, I suppose
that if phy is null then of_get_property will just return NULL, so I
think I made a mistake.  I'll post a patch to remove these two lines.

g.

>
>> +
>>        prop = of_get_property(phy, "interface", NULL);
>>        if (prop != NULL) {
>>           phy_interface = enet_to_phy_interface[*prop];
>> diff --git a/drivers/net/ucc_geth.h b/drivers/net/ucc_geth.h
>> index 2f8ee7c..bc31975 100644
>> --- a/drivers/net/ucc_geth.h
>> +++ b/drivers/net/ucc_geth.h
>> @@ -1100,7 +1100,7 @@ struct ucc_geth_info {
>>     u32 eventRegMask;
>>     u16 pausePeriod;
>>     u16 extensionField;
>> -   char phy_bus_id[BUS_ID_SIZE];
>> +   struct device_node *phy_node;
>>     u8 weightfactor[NUM_TX_QUEUES];
>>     u8 interruptcoalescingmaxvalue[NUM_RX_QUEUES];
>>     u8 l2qt[UCC_GETH_VLAN_PRIORITY_MAX];
>>
>> _______________________________________________
>> Linuxppc-dev mailing list
>> Linuxppc-dev@ozlabs.org
>> https://ozlabs.org/mailman/listinfo/linuxppc-dev
>>
>>
>
>



-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.