What's your level of confidence reading C code (not necessarily writing it), availability of a serial connection to your device, and tolerance for testing things?
With luck "all" you'll have to do is adapt a DTS to describe the device to the kernel.
I'll look at it in a bit (in the middle of some testing and merging here) to take a guess as to how hard it might be, but here are a couple references. The first is the learning process the early adopters went through. The second describes some of the references I found helpful when learning about DTS files myself.
https://openwrt.org/docs/guide-developer/defining-firmware-partitions (Really about DTS, in general now)
https://openwrt.org/toh/zyxel/zyxel_nbg6716
The hardware is very similar to the TP-Link TL-WDR7500, but with a large NAND flash
The Hynix H27U2G8F2CTR-BC listed on that page is parallel NAND, and I don't know if there is any support for it yet on the ath79 platform. Having been working on serial NAND on that platform ("SPI NAND"), I know that there aren't any boards running on the ath79 NAND target on master quite yet.
Best guess is somewhere between "hard" (a lot of groundbreaking work, but platform drivers for parallel NAND and your chip available) and "very hard" (requiring submissions and acceptance to upstream Linux).