Hey all,
First some background, I have an XGS1010-12, which is hardware-wise (lack of a reset-button) completly identical to the XGS1210-12. Software, is a different matter; the XGS1010-12 is an unmanaged switch, and in software, makes use of pretty much the entire RTL930x SDK defaults. This also means the stock partition layout, is one with 2 firmware banks (each about 7MiB). Except for mtd0 (u-boot binary) and mtd5 (kernel + initramfs) all mtd's are all completly empty. No data is stored in them, so yes, the secondary firmware slot is empty
The XGS1210-12 however, is webmanaged, and has a single firmware bank configured, that's about 15MiB.
Now of course, I can just install the XGS1210-12 vendor/openwrt firmware, and would have a big partition, and we just treat the XGS1010-12 as a XGS1210-12, no problem right.
But my real question, is of course more then this bit of background To make better use of this switch (at the cost of loss of the second firmware bank support) is to use an 'openwrt optimized partitioning scheme' e.g. do a partition layout that that is sane, in this example, it would be simply spanning both firmware banks, but more things could be envisioned, that's a bit besides the point for now.
In software this could be easy still, we have 2 targets -vendor
or -stock
and openwrt, mostly affecting partition layout I can imagine, but sometimes this would also require 'migration' for users. E.g. boot initramfs, make backup! (you should do this always right) and then move your calibration/factory data from a to b (though this usually would only work on targets that also do U-Boot changes ...).
Anyway, a lot of background/thoughts from my end; what is the generally OpenWRT stance here? I get that we want to offer 'the most easy path from stock' but we don't want to hamper the devices possiblities/future because the vendor did something ... stupid?