[LEDE-DEV] Help needed: IP175D + RT3662 issues on a "new" device

John Crispin john at phrozen.org
Sat Dec 24 09:31:46 PST 2016



On 24/12/2016 18:22, John Crispin wrote:
> 
> 
> On 24/12/2016 17:49, Martin Blumenstingl wrote:
>> On Sat, Dec 24, 2016 at 5:17 PM, antonio rossi <braccoz at hotmail.com> wrote:
>>> Hi everybody,
>>>
>>> i'll try to make this as short as possible despite the large amount of
>>> information needed to describe the issue:
>>>
>>> i am working on adding support to LEDE/OpenWRT for DIR-815 A1, it's a
>>> RT3662+RT3092 device with an infamous IP175D switch IC.
>>> i managed to get everything working properly, save for the fact that
>>> network (both ethernet&wifi) works only after the stock bootloader
>>> does some funky stuff to set up the internal cpu port and/or switch
>>> to receive images via TFTP.
>> if only ethernet was affected then this might have been some pinctrl issue.
>> but since wifi is also affected this might be more of a clock (gate?) issue.
>>
>> maybe John has more information (or even a suspect) here?
> 
> 
> 
> 
> this sound like either
> 
> 1) the cpu port is not configrued properly,
> 2) the switch is not configured properly
> 3) pinmux is failing
> 
> the odd thing is that wifi fails without bootloader init. how does wifi
> fail exactly ?
> 
> 	John

also post your dts file please

	John

> 
> 
>>
>>> this basically means that when the bootloader receives something via
>>> TFTP then any image booted (either initramfs or from flash) works properly,
>>> while a squashfs image booted from the internal flash (so the bootloader
>>> does not set up cpu port/switch) doesn't.
>>> the boot log of working and non-working states are basically identical. the
>>> switch is recognized and everything. i've tried every possible combination of
>>> parameters in the DTS file to no avail.
>>> I also investigated GPIO pins, but the only pin that has somewhat of an effect
>>> is GPIO24 which is connected to the switch reset pin, and it has no effect on
>>> the issue itself (it just resets the switch that then needs to be reconfigured
>>> with the vlans and such)
>>> What catches my interest are those few lines from the bootloader that says
>>>
>>>
>>> 1.set ge1 as normal mode
>>>
>>>
>>> 1.set ge1 as mii mode
>>>
>>>
>>> enable turbo mii
>>>
>>>
>>> Using Eth0 (10/100-M) device
>>>
>>>
>>> i am guessing ge1 refers to the gigabit ethernet cpu port and the turbo mii,
>>> well, i can't set that mode, just normal mii. it's weird because changing
>>> phy-mode in the DTS actually affects the cpu port and switch but there is
>>> still something missing to make it work without bootloader/stock rom
>>> intervention.
>>>
>>> here are the bootlog pastebins, both of working and non working state
>>> (in this order).
>>> any help is appreciated.
>>>
>>> http://pastebin.com/rNwewZ8J
>>> http://pastebin.com/Gw9DNfFZ
>>>
>>> here is the DTS
>>> http://pastebin.com/iLSbw9gJ
>>>
>>> here is the swconfig output (identical in both working and non-working state)
>>> http://pastebin.com/iBK47i9m
>>>
>>> best regards,
>>> Bracco
>>> _______________________________________________
>>> Lede-dev mailing list
>>> Lede-dev at lists.infradead.org
>>> http://lists.infradead.org/mailman/listinfo/lede-dev



More information about the Lede-dev mailing list