> The purpose of the nvram init script is to reset important variables
> in-case previous firmwares crapped them up. It is something that has
> been done since the first wrt54{g,gs} and is well documented.
OTOH it's not clear *why* it has been done all this time.
You say "reset important variables in-case previous firmwares crapped
them up", but what are those firmwares?
Also, this presumes we know better, which it seems we don't (as
evidenced by this thread).
Finally, if nvram settings are good enough to load the current firmware
and reach the init script that resets the nvram settings, then maybe
those nvram settings aren't that bad after all. It seems a lot safer to
not reset those settings.
I understand that poor nvram settings may cause the box to be "working
enough to run the nvram script" but still be unreachable, in which case
the nvram reset may be a lifesaver. But I think we need to be more
careful to try and distinguish the actual problematic cases and only do
something when we find one of them. For that we need to know what are
those problematic cases and which boards they affect and how we can
reliably recognize them.
Stefan
_______________________________________________
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
