Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1670226 > unrolled thread

Re: [PATCH NET] net/hns:bugfix of ethtool -t phy self_test

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH NET] net/hns:bugfix of ethtool -t phy self_test l00371289 <linyunsheng@huawei.com> - 2017-06-20 05:20 +0200
    Re: [PATCH NET] net/hns:bugfix of ethtool -t phy self_test Andrew Lunn <andrew@lunn.ch> - 2017-06-20 15:30 +0200
      Re: [PATCH NET] net/hns:bugfix of ethtool -t phy self_test l00371289 <linyunsheng@huawei.com> - 2017-06-21 04:10 +0200

#1670226 — Re: [PATCH NET] net/hns:bugfix of ethtool -t phy self_test

Hi, Andrew

On 2017/6/20 5:54, Andrew Lunn wrote:
> On Mon, Jun 19, 2017 at 02:00:43PM -0700, Florian Fainelli wrote:
>> On 06/16/2017 02:24 AM, Lin Yun Sheng wrote:
>>> This patch fixes the phy loopback self_test failed issue. when
>>> Marvell Phy Module is loaded, it will powerdown fiber when doing
>>> phy loopback self test, which cause phy loopback self_test fail.
>>>
>>> Signed-off-by: Lin Yun Sheng <linyunsheng@huawei.com>
>>> ---
>>>  drivers/net/ethernet/hisilicon/hns/hns_ethtool.c | 16 ++++++++++++++--
>>>  1 file changed, 14 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c b/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c
>>> index b8fab14..e95795b 100644
>>> --- a/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c
>>> +++ b/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c
>>> @@ -288,9 +288,15 @@ static int hns_nic_config_phy_loopback(struct phy_device *phy_dev, u8 en)
>>
>> The question really is, why is not this properly integrated into the PHY
>> driver and PHYLIB such that the only thing the Ethernet MAC driver has
>> to call is a function of the PHY driver putting it in self-test?
> 
> This whole driver pokes various PHY registers, rather than use
> phylib. And it does so without taking the PHY lock. 
I will consider using phylib as much as possible, thanks.

It also assumes it
> is a Marvell PHY and i don't see anywhere it actually verifies this.
When it said Marvell Phy , I meant Marvell Phy with fibre support.
I will send anther patch to only setting bit in Fiber Control when
it is a Marvell Phy with fibre support.

Thanks for reply.
Best Regards
Yunsheng Lin
> 
> This is all broken.
> 
>      Andrew
> 
> .
> 

[toc] | [next] | [standalone]

#1670756

> >> The question really is, why is not this properly integrated into the PHY
> >> driver and PHYLIB such that the only thing the Ethernet MAC driver has
> >> to call is a function of the PHY driver putting it in self-test?
> > 
> > This whole driver pokes various PHY registers, rather than use
> > phylib. And it does so without taking the PHY lock. 
> I will consider using phylib as much as possible, thanks.
> 
> It also assumes it
> > is a Marvell PHY and i don't see anywhere it actually verifies this.
> When it said Marvell Phy , I meant Marvell Phy with fibre support.
> I will send anther patch to only setting bit in Fiber Control when
> it is a Marvell Phy with fibre support.

There is a lot more broken than just that.

You really should remove all code which is accessing the PHY, and add
support to phylib and the drivers for what you need.

    Andrew

[toc] | [prev] | [next] | [standalone]

#1671286

Hi, Andrew

On 2017/6/20 21:28, Andrew Lunn wrote:
>>>> The question really is, why is not this properly integrated into the PHY
>>>> driver and PHYLIB such that the only thing the Ethernet MAC driver has
>>>> to call is a function of the PHY driver putting it in self-test?
>>>
>>> This whole driver pokes various PHY registers, rather than use
>>> phylib. And it does so without taking the PHY lock. 
>> I will consider using phylib as much as possible, thanks.
>>
>> It also assumes it
>>> is a Marvell PHY and i don't see anywhere it actually verifies this.
>> When it said Marvell Phy , I meant Marvell Phy with fibre support.
>> I will send anther patch to only setting bit in Fiber Control when
>> it is a Marvell Phy with fibre support.
> 
> There is a lot more broken than just that.
> 
> You really should remove all code which is accessing the PHY, and add
> support to phylib and the drivers for what you need.
> 
>     Andrew
After adding genphy_loopback support, I will try it. Thanks for pointing out.

Best Regards
Yunsheng Lin

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web