Re: [patch 2.6.14-rc3 2/3] sundance: probe PHYs from MII address 0

[Jeff Garzik <jgarzik@pobox.com>]

John W. Linville wrote:
> Probe for PHYs starting at MII address 0 instead of MII address 1.
> This covers the entire range of MII addresses.
> 
> Signed-off-by: John W. Linville <linville@tuxdriver.com>
> ---
> 
>  drivers/net/sundance.c |    2 +-
>  1 files changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/net/sundance.c b/drivers/net/sundance.c
> --- a/drivers/net/sundance.c
> +++ b/drivers/net/sundance.c
> @@ -608,7 +608,7 @@ static int __devinit sundance_probe1 (st
>  
>  	np->phys[0] = 1;		/* Default setting */
>  	np->mii_preamble_required++;
> -	for (phy = 1; phy < 32 && phy_idx < MII_CNT; phy++) {
> +	for (phy = 0; phy < 32 && phy_idx < MII_CNT; phy++) {

NAK.  MII address 0 should be scanned _last_, after all other addresses. 
  In some phys, it is a ghost, mirroring another address.

Take a look at some of the original Becker MII scan code from
ftp://ftp.scyld.com/pub/network/ to see an elegant method for this.

Becker's scan code would utilize a mask to keep the loop nice and 
elegant, eliminating an "if (phy == 32) phy = 0;" test.

	Jeff