[PATCH v2] ath79: add support for onion omega

[Lech Perczak]

W dniu 2021-08-15 o 10:30, Jan-Niklas Burfeind pisze:
>
> On 8/14/21 7:08 PM, Lech Perczak wrote:
>> [...]
>>>> The six bytes at 001fd00 spell OMEGA.
>>> Is this the only data in 0x10000-0x20000? What's in 0x10000-0x11000?
>>>
>>>>>> +                read-only;
>>>>>> +                compatible = "nvmem-cells";
>>>>>> +                #address-cells = <1>;
>>>>>> +                #size-cells = <1>;
>>>>>> +
>>>>>> +                macaddr_uboot_1fc00: macaddr at 1fc00 {
>>>>>> +                    reg = <0x1fc00 0x6>;
>>>>>> +                };
>>>>>> +            };
>>>>>> +
>>>>>> +            partition at 20000 {
>>>>>> +                compatible = "tplink,firmware";
>>>>>> +                label = "firmware";
>>>>>> +                reg = <0x020000 0xfd0000>;
>>>>>> +            };
>>>>>> +
>>>>>> +            art: partition at ff0000 {
>>>>>> +                label = "art";
>>>>>> +                reg = <0xff0000 0x010000>;
>>>>>> +                read-only;
>>>>>> +            };
>> The whole ordeal looks very much like typical pre-safeloader TP-link
>> flash layout, so I expect no writable U-boot environment there, at least
>> for stock U-boot.
>> Of course, it would be best to check if it's possible to write it using
>> serial console, on actual device.
>>
> I just tried to copy part of the string "OMEGA" from its current position.
>
>> md 0x9F01FD00
> 9F01FD00: 4F4D4547[...]
>
> cp 0x9F01FD00 0x9F01FE00 3
>> md 0x9F01FE00
> 9F01FE00: FFFFFFFF[...]
>
> Reading the partition works fine; writing not so much; at least not
> within uboot.
I rather meant checking if it is possible to 'saveenv' from within stock 
U-boot, rebooting and seeing if changes to the environment change stay 
there.
Given, that the typical SPI-NOR sector size on this target is 64k, it's 
very unlikely that there is writable environment there, as the MAC part 
doesn't follow its format, and U-boot would need to use 4k page erase 
for that,
as does pepe2k's u-boot_mod, but OpenWrt currently cannot reliably 
support writing such environment.
>
> [...]
>
> _______________________________________________
> openwrt-devel mailing list
> openwrt-devel at lists.openwrt.org
> https://lists.openwrt.org/mailman/listinfo/openwrt-devel