Hi, Andrew
On 2017/6/27 21:29, Andrew Lunn wrote:
>>>> - phy_write(phy_dev, COPPER_CONTROL_REG, val);
>>>> + err = phy_resume(phy_dev);
>>>
>>> Maybe this was discussed with an earlier version of these patches. Why
>>> are using phy_resume() and phy_suspend()?
>> When self_test is invoked with ETH_TEST_FL_OFFLINE option, hns mac driver
>> call dev_close to set net dev to offline state if net dev is online.
>> Doing the actual phy loolback test require phy is power up, So phy_resume
>> and phy_suspend are used.
>
> O.K, so you at least need some comments, because this is not obvious.
>
>>From your description, it sounds like you can call phy_resume() on a
> device which is not suspended.
Do you mean after calling dev_close, the device is still not suspended?
If that is the case, is there any way I can ensure the device is suspended?
In general, suspend is expected to
> store away state which will be lost when powering down a
> device. Resume writes that state back into the device after it is
> powered up. So resuming a device which was never suspended could write
> bad state into it.
Do you mean phydev->suspended has bad state?
>
> Also, what about if WOL has been set before closing the device?
phy_suspend will return errro.
int phy_suspend(struct phy_device *phydev)
{
struct phy_driver *phydrv = to_phy_driver(phydev->mdio.dev.driver);
struct ethtool_wolinfo wol = { .cmd = ETHTOOL_GWOL };
int ret = 0;
/* If the device has WOL enabled, we cannot suspend the PHY */
phy_ethtool_get_wol(phydev, &wol);
if (wol.wolopts)
return -EBUSY;
if (phydev->drv && phydrv->suspend)
ret = phydrv->suspend(phydev);
if (ret)
return ret;
phydev->suspended = true;
return ret;
}
Best Regard
Yunsheng Lin
