[PATCH v3 0/8] arm64: rockchip: Initial GeekBox enablement

[Giuseppe CAVALLARO]

On 3/9/2016 10:42 AM, Tomeu Vizoso wrote:
> On 9 March 2016 at 10:00, Giuseppe CAVALLARO <peppe.cavallaro at st.com> wrote:
>> On 3/9/2016 9:56 AM, Giuseppe CAVALLARO wrote:
>>>
>>> Hello
>>>
>>> On 3/9/2016 9:35 AM, Tomeu Vizoso wrote:
>>>>
>>>>
>>>> in order to get the onboard network on the Radxa Rock2 to work at all
>>>> on today's linux-next, I had to revert both commits:
>>>>
>>>> * "stmmac: Fix 'eth0: No PHY found' regression" for the drivers to
>>>> probe and the link to come up, and
>>>
>>>
>>> I've just sent two patches (on top of net.git) to fix this problem
>>> pls let me know if these help on your side too.
>>>
>>>>
>>>> * "stmmac: first frame prep at the end of the xmit routine" for the
>>>> network interface to work at all.
>>>
>>>
>>> I cannot reproduce it right now but I am looking at this again and let
>>> you know.
>>> I will check your output file too
>>
>>
>> hmm, I think that the problem is around the "normal descriptor"
>> management that is configured on the boxes where the driver is
>> failing now. Using enhanced descriptors the network is ok.
>> I will keep you up-to-date.
>
> Hi,
>
> I have done some tests and these are the results:
>
> * net.git (133800d1f028): probe failed

ok, failure due to the "stmmac: Fix 'eth0: No PHY found' regression"

>
> * net.git (133800d1f028) + revert: everything fine

ok that is expected but the fixed-link will be NOK

>
> * net.git (133800d1f028) + revert + "stmmac: fix MDIO settings": everything fine

thx for having tested it. So it works on real transceiver (tested by
you) and with the fixed-link (tested by me).

>
> * today's linux-next: probe failed
>
> * today's linux-next + revert of 88f8b1bb41c6 stmmac: Fix 'eth0: No
> PHY found' regression: probe succeeded but no network at all
>
> * today's linux-next + revert of 88f8b1bb41c6 (stmmac: Fix 'eth0: No
> PHY found' regression) + revert of 0e80bdc9a72d (stmmac: first frame
> prep at the end of xmit routine): probe succeeded, dhcp succeeds and
> nfsroot works for a few seconds before timing out

ok, I was looking at this problem now that seems to related
the "stmmac: first frame prep at the end of xmit routine"
that, at first glance, is breaking the gmac 3.50 with normal descriptor.

> Btw, could you CC LKML on future patches that need testing?

sure

peppe

>
> Thanks,
>
> Tomeu
>
>> peppe
>>
>>
>>>
>>> Regards
>>> peppe
>>>
>>>
>>>> But a few seconds into the nfsroot boot I start getting:
>>>>
>>>> [    9.136701] systemd[1]: Started Load Kernel Modules.
>>>> [   18.116321] nfs: server 10.42.0.1 not responding, still trying
>>>> [   18.516224] ------------[ cut here ]------------
>>>> [   18.521024] WARNING: CPU: 1 PID: 0 at net/sched/sch_generic.c:303
>>>> dev_watchdog+0x284/0x288
>>>> [   18.529456] NETDEV WATCHDOG: eth0 (rk_gmac-dwmac): transmit queue 0
>>>> timed out
>>>>
>>>> I'm attaching the full boot log.
>>>>
>>>> Regards,
>>>>
>>>> Tomeu
>>>>
>>>
>>>
>>
>
>