The phy ids used are taken from an driver that used a right shift of 4 to chop
off the revision number. This driver does not shift, so the id and mask
values are wrong and must be left shifted by 4 to actually detect the chips.
Signed-off-by: Uwe Zeisberger <Uwe_Zeisberger@digi.com> Acked-by: Andy Fleming <afleming@freescale.com>
[akpm: this is a previously-nacked patch, but the problem is real] Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Jeff Garzik <jeff@garzik.org>