[Intel-wired-lan] [PATCH v1 1/1] igc: Remove copper fiber switch control

[Neftin, Sasha <sasha.neftin@intel.com>]

On 3/3/2020 08:00, Alexander Duyck wrote:
> On Mon, Mar 2, 2020 at 4:44 PM Neftin, Sasha <sasha.neftin@intel.com> wrote:
>>
>> On 3/2/2020 13:26, Alexander Duyck wrote:
>>> On Mon, Mar 2, 2020 at 12:23 PM Sasha Neftin <sasha.neftin@intel.com> wrote:
>>>>
>>>> i225 device support copper mode only
>>>> PHY signal detect indication for copper fiber switch
>>>> not applicable to i225 part
>>>>
>>>> Signed-off-by: Sasha Neftin <sasha.neftin@intel.com>
>>>
>>> So there are a couple issues with this patch.
>>>
>>> All the changes in igc_ethtool.c are broken at this point. Once a
>>> register is defined in regs_buff you cannot change it. Otherwise you
>>> cannot debug this in the future. You would be better off just skipping
>>> the register that you were storing CONNSW and let it default to zero
>>> instead of doing all of the shifting you are doing. You can just skip
>>> over the register in the dump in ethtool assuming there is even a file
>>> for the device that hasbeen added.
>>>
>> This change not affected igc_ethtool.c behavior. I see the same behavior
>> on my setup.
>> Actually ethtool --register-dump not called (as properly).get_regs
>> callback from igc_ethtool.c. This is not related to this patch and I
>> need investigate and fix it.
>> ethtool --register-dump <adapter> show me row generic data. Data is
>> really from i225 registers, but not parcered as for other drivers.
>>
> 
> Right. It isn't implemented yet, but you don't want to break it. The
> logic for the register dump assumes the register locations are fixed.
> So once you released this the first time you cannot move the registers
> around. You can drop registers or add registers, but once they are in
> the list they are permanently in that position. If you don't do that
> you will never be able to decode the data.
> 
ok. I will re-send v2 and leave igc_ethtool.c no changes at this point.
Sasha
> - Alex
>