[Cerowrt-devel] tp-link 4300 evaluation

[Lance Hepler]

On Mon, May 27, 2013 at 6:33 AM, David Lang <david at lang.hm> wrote:

> On Mon, 27 May 2013, Dave Taht wrote:
>
>
>>>  That's tragic. I just picked up a Netgear WNDR4300 (openbox on sale at
>>> the
>>>
>>>> local Fry's) to see if I could hack up a CeroWrt clone on it. It seems
>>>> to
>>>> be mostly the same hardware as the WNDR3700v4 and the TP-Link
>>>> WDR43[01]0,
>>>> with things just wired up slightly differently.
>>>>
>>>>
>>> As I understnad it, the difference between the WNDR3700v4 and WNDR4300 is
>>> that the 4300 has a slightly better wireless chip.
>>>
>>> Unfortunantly from what I've seen so far, they did something wierd with
>>> the storage and as a result the stock openwrt can't access it. I've seen
>>> reports of people getting it to run from an initramfs, but this means
>>> that
>>> no settings can be preserved across reboot.
>>>
>>> If you've seen anything different, I'd be very interested to hear about
>>> it
>>> (I picked up a 3700v4 and a couple 4300's for testing)
>>>
>>>
>> according to a birdie, "it looks like it's an ONFI with quirks, or nobody
>> has realised that it's ONFI at all.". Perhaps that's enough clue to get
>> someone started? but I fear jtag debugging will be needed. Flash chips
>> tend
>> to have interesting race conditions....
>>
>
> Given that we have the GPL source for the kernel available from Netgear,
> I'm a little puzzled that we are having to reverse engineer this instead of
> working from the source.
>
> Even if the first version was little more than a cut-n-paste of the
> netgear butchered driver until people have time to analyse it fully.


Found the original NAND driver. If you download the original fw source it's
in wndr4300-GPL.git/git_home/linux.git/drivers/mtd/nand/ath_nand.c. It also
borrows a few things from
wndr4300-GPL.git/git_home/linux.git/arch/mips/include/asm/mach-atheros/atheros.h.

I'm already seeing lots of similarities to the ar934x_nfc.c flash
controller driver in upstream openwrt. That one seems to be missing some of
the ECC features of the original ... and all of the ONFI bits...

I'll poke at it some this week, see if I can bring something up.

Lance
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.bufferbloat.net/pipermail/cerowrt-devel/attachments/20130527/749aa9a9/attachment-0002.html>