EcoNet: MT7612E not working

[Caleb James DeLisle]

On 26/10/2025 15:20, Daniel Golle wrote:
> On Sun, Oct 26, 2025 at 03:07:55PM +0100, Caleb James DeLisle wrote:
>> This is a bit of a long-shot but I'm at a loss here. On both of my EcoNet
>> devices with MT7612 chips, it fails to start. MT7613 and MT7603 both work.
>> I'm suspicious of my PCI controller because it's something I worked on, but
>> I don't see any evidence of PCI issues.
>>
>>
>> Is there anything really unique about MT7612, particularly how it uses PCI ?
>>
>>
>> MT7613E / Slot 2 -> Works (TP-Link Archer VR1200v (v2))
>>
>> MT7603 / Slot 1 or slot 2 -> Works (ZyXEL PMG5617GA and SmartFiber XP8421-B
>> respectively)
>>
>> MT7612 / Slot 1 or slot 2 -> Fails (SmartFiber XP8421-B and ZyXEL PMG5617GA
>> respectively)
>>
>>
>> Here is my bootlog:
>>
>> [   13.252551] mt76x2e 0000:02:00.0: enabling device (0000 -> 0002)
>> [   13.258716] pcibios_plat_dev_init: bus 2, slot 0
>> [   13.263689] mt76x2e 0000:02:00.0: ASIC revision: 76120044
>> [   13.269738] mt76x2_eeprom_init
>> [   13.272871] mt76_get_of_eeprom 80004328
>> [   13.276719] mt76_get_of_data_from_nvmem eeprom
>> [   13.385361] mt76x2_eeprom_load 1
>> [   13.388635] mt76x2_eeprom_load found=1
>> [   13.523388] urngd: v1.0.2 started.
>> [   14.042816] mt76x2e 0000:02:00.0: Invalid MAC address, using random
>> address b6:db:5e:21:55:82
> Are you sure the EEPROM data passed to the driver is correct?
> I wonder because it should also contain a valid MAC address...

I'm sure I'm giving it an EEPROM which is in the flash, and starts like 
this:

62 76 01 00 00 00 00 00  00 00

So I understand that means the MAC is zeroed. "mt76x2_eeprom_load 
found=1" is my log saying it passed the check.

>> [   14.083630] mt76x2e 0000:02:00.0: ROM patch build: 20141115060606a
>> [   14.113187] mt76x2e 0000:02:00.0: Firmware Version: 0.0.00
>> [   14.118716] mt76x2e 0000:02:00.0: Build: 1
>> [   14.122875] mt76x2e 0000:02:00.0: Build Time: 201607111443____
>> [   14.572779] mt76x2e 0000:02:00.0: Firmware failed to start
>> [   14.578409] mt76x2e 0000:02:00.0: probe with driver mt76x2e failed with
>> error -145
>>
>>
>> I've added a few debug logs to confirm eeprom is loaded correctly and passes
>> mt76x2_check_eeprom().
>> Firmware and ROM patch match (working) vendor OS bit-for-bit.
>>
>> Any ideas where else to look?
> If the problem is not the missing or wrong EEPROM data,


Thank you, this is very helpful.


> maybe the PCIe
> reset line is inverted or configured wrongly (-> pinctrl)?
>
> Just guessing, of course...
>

Thanks, I'll go back and re-check all of the reset commands, while I 
still have no confidence in my own PCI impl, I'm using a very lightly 
modified version of 
https://github.com/cjdelisle/kernel-49/blob/master/arch/mips/tc3262/pci_en75xx.c 
which is vendor code, but I'll go around and look for suspicious GPIO or 
pinctrl accesses in the rest of the init for that kernel.


Thanks again,

Caleb