Stephen Hemminger wrote:

Several of the PHY registers are multiplexed; access to
register must be proceeded by setting page register.

The driver setup is safer if this is done before the access
rather than depending on the last value.

Signed-off-by: Stephen Hemminger <redacted>

NAK for 2.622 - this appears to be a highly speculative cleanup-style 
change that is not an appropriate submission so late in the Release 
Candidate cycle.

I would have expected this sort of change to have been brewing for a 
little while in -mm, via netdev#upstream perhaps.  As it is, this change 
has been zero testing exposure, which is not something we want in a -rc2 
or -rc3.

	Jeff