[LEDE-DEV] [OpenWrt-Devel] [PATCH 0/4] Gemini forward-port to kernel v4.14
Roman Yeryomin
roman at advem.lv
Mon Apr 16 15:21:49 PDT 2018
On 2018-04-17 00:36, Linus Walleij wrote:
> On Mon, Apr 16, 2018 at 2:00 AM, Roman Yeryomin <roman at advem.lv> wrote:
>
>>> Looking at your tree...
>>> I don't see any (affecting) differences in ethernet driver itself.
>>> Probably it's something else.. MDIO?
>>>
>>
>> After looking into ethernet I've found several issues.
>> 1. skew delay settings were not ported from old driver to dts for
>> nas4220b
>> board
>> 2. kernel config in you patches (accidentally?) disabled bridge
>> support
>> 3. driver crashes if you try to disable unused gmac in dts because of
>> access
>> to uninitialized port(0|1) member of struct gemini_ethernet
>
> Ouch. Sorry for my sucky backport. :(
>
>> So after fixing all above ethernet on nas4220b is working ok.
>
> You are my hero :)
>
>> Can you confirm that after enabling bridge support back (just remove
>> CONFIG_BRIDGE from gemini kernel config and rebuild) ethernet comes up
>> on
>> D-link boards? That is with default network config.
>
> I think it's best if I test your entire set-up on the D-Link routers.
> Do you have a git branch I can grab?
Will try to prepare that tomorrow.
Target kernel config must be cleaned up, it's redundant in some places
and overrides some things which it should not.
I've managed to get usb working, it was kernel config error again (and
Hans' patch he sent earlier is actually working). But still usb has
problems, e.g. I couldn't get both ports working, only one of them.
SATA seems to be working out of the box, which is good!
Regards,
Roman
More information about the Lede-dev
mailing list