hi, Andrew On 2017/6/23 11:20, Andrew Lunn wrote: > On Thu, Jun 22, 2017 at 04:53:33PM +0800, Lin Yun Sheng wrote: >> Use function set_loopback in phy_driver to setup phy loopback >> when doing ethtool self test. >> >> Signed-off-by: Lin Yun Sheng <linyunsh...@huawei.com> >> --- >> drivers/net/ethernet/hisilicon/hns/hns_ethtool.c | 64 >> ++++-------------------- >> 1 file changed, 10 insertions(+), 54 deletions(-) >> >> diff --git a/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c >> b/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c >> index e95795b..660b51e 100644 >> --- a/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c >> +++ b/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c >> @@ -259,66 +259,22 @@ static int hns_nic_set_link_ksettings(struct >> net_device *net_dev, >> >> static int hns_nic_config_phy_loopback(struct phy_device *phy_dev, u8 en) >> { >> -#define COPPER_CONTROL_REG 0 >> -#define PHY_POWER_DOWN BIT(11) >> -#define PHY_LOOP_BACK BIT(14) >> - u16 val = 0; >> >> - if (phy_dev->is_c45) /* c45 branch adding for XGE PHY */ >> + if (phy_dev->is_c45 || !phy_dev->drv) >> + return -ENOTSUPP; >> + >> + if (!phy_dev->drv->resume || !phy_dev->drv->suspend || >> + !phy_dev->drv->set_loopback) >> return -ENOTSUPP; > > No, i said to have a function is phy.c which does this check. The phy > core needs to keep track of if loopback is already enabled, and return > an error if it is attempted to turn it on twice, or disable it when it > is not enabled. what should the Mac driver do when it receive the error? I assume that it will return it to userspace. because the rtnl_lock ensure only one ethtool is executing, if the function return error, it means some bug in the mac driver, am I right? will add the function in phy.c next version.
Best Regards Yunsheng Lin