RE: eth0: PHY id 0x00221619 is not supported! HELP

Hi Colin,

I worked on a project 2 years ago, and the hardware supplier decided to
change the tranceiver twice during the course of the project.  So =
instead of modifying the mii_phy_table
each time the transeiver was changed, I bypassed the id check. =20

I did this because I doubt that in the future people are going to be =
interested in updating software, and delivering a new s/w version when a =
hardware fix is performed.

It's not clean, but it's easy and convenient.
This is just a suggestion of mine for a quick fix.

Tien

-----Original Message-----
From: Colin Leroy [mailto:colin@colino.net]
Sent: Friday, May 13, 2005 3:31 PM
To: Nghiem, Tien
Cc: Garcia J=E9r=E9mie; linuxppc-dev@ozlabs.org
Subject: Re: eth0: PHY id 0x00221619 is not supported! HELP


On 13 May 2005 at 14h05, Nghiem, Tien wrote:

Hi,=20

         rc =3D reset_one_mii_phy(phy, mii_id);
        //if (rc)
        //        return -ENODEV;

Why disabling such guards?

        //for (i=3D0; (def =3D mii_phy_table[i]) !=3D NULL; i++)
        //        if ((id & def->phy_id_mask) =3D=3D def->phy_id)
        //                break;

Just an idea, I guess it'd be far more clean to add the phy id to
mii_phy_table[] (and, optionally, send the patch so you won't have to
do it everytime).

You're just asking for breakage there.
--=20
Colin