Yes, I sent a patch. Then he sent another patch which breaks things differe=
ntly. I have not yet submitted my fix. My fix is to revert his patch, and t=
hen modify your updated device trees to automatically set the tbi to someth=
ing.

On Nov 24, 2011, at 1:51, "Kumar Gala" [off-list ref] wrote:

=20
On Nov 9, 2011, at 2:10 PM, Andy Fleming wrote:
=20

quoted

quoted

Fix this by moving the of_mdiobus_register() call earlier.
=20
Cc: Andy Fleming <redacted>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
---
drivers/net/ethernet/freescale/fsl_pq_mdio.c |   14 +++++++-------
1 files changed, 7 insertions(+), 7 deletions(-)
=20

diff --git a/drivers/net/ethernet/freescale/fsl_pq_mdio.c b/drivers/net=

/ethernet/freescale/fsl_pq_mdio.c

quoted

quoted

index 52f4e8a..e17fd2f 100644

--- a/drivers/net/ethernet/freescale/fsl_pq_mdio.c
+++ b/drivers/net/ethernet/freescale/fsl_pq_mdio.c

@@ -385,6 +385,13 @@ static int fsl_pq_mdio_probe(struct platform_devic=

e *ofdev)

quoted

quoted

                      tbiaddr =3D *prop;
      }
=20
+       err =3D of_mdiobus_register(new_bus, np);
+       if (err) {
+               printk (KERN_ERR "%s: Cannot register as MDIO bus\n",
+                               new_bus->name);
+               goto err_free_irqs;
+       }
+

=20
=20
This fix totally breaks the point of setting tbipa beforehand.
mdiobus_register will cause the bus to be scanned, and if any of the
PHYs are at the default address for tbipa, they won't be found. I have
a different fix which I will (re)submit today.

=20
What happened here, did you send a patch?
=20
- k