Support for TP-Link TD-W8960N v5 (& TD-W8968 v3)

Hi everyone!

I'm experimenting OpenWrt on the TP-Link TD-W8960N v5 (while updating the WIKI).

I've serial access.
I was able to flash the OpenWrt image built for the Plusnet FAST 2704N V1 from CFE and back-flash the original TP-Link image as well.

Using the F@st2740N image everything works very well (I tested Wifi, switch, vlans).

Because the led and buttons are not correct, I'm asking you to help me identify the correct GPIO associated and possibly make a patch like the one provided for TD-W8968 v3.

What else is missing to build a dedicated target in OpenWrt downloads?

Thanks for you time and help!

Hi, I am looking to hire Openwrt developer to help me with getting to a stable working state a router model: TD-W8960N v5.
I have 600pcs so I am ready to offer a reasonable amount to get some help.
So far I have flashed it with:
openwrt-brcm63xx-generic-TD-W8968-squashfs-cfe
On a first look everything is working as it should apart from the wifi, its very slow 2mbps download / 9mbps upload.
I have noticed that Generic MAC80211 802.11bg (radio0) shows only b and g standart, for some reason n is missing.
I guess that is a driver issue. And I guess the slow wifi is related to that too. Even as g radio should get something close to 20mbps.
I am just a user, I don't have the knowledge to sort this out.
I really hope that someone will be interested to collaborate.
Thank you!
Ivo
Whatsapp/Viber: +35799944162

You're unlikely to get this working. The BCM43217 radio is not supported by brcmfmac, but only by b43 which has no support for aggregation (i.e., it's an 11abgn radio, but cannot use the important features of an 11abgn radio).

You are likely limited by the speed of the CPU in performing software-based cryptography.

Well, that's a pity. So this is a 100% dead sentence or there is a chance although I can see very small if any?

I can't imagine anyone volunteering to work on HT for b43 (even less with VHT/ HE being in more demand today), at least not for a 'reasonable amount'. It would require months of work, and at least two independent teams to work on it (Broadcom is hostile towards opensource b43 development and there's zero hardware documentation or cooperation from them, this means clean room reverse engineering (as it has been done during early bcm43xx/ b43 development) would be necessary - one team documenting the hardware and the proprietary driver/ firmware, another -distinct- team working on implementing the opensource driver, no cross-pollination possible for legal reasons). It's probably significantly cheaper to dispose those ~decade-old 600 devices and to buy new (non-broadcom) ones, than to fund this endeavour.