On Fri, Jan 26, 2018 at 4:57 PM, Jean Flamelle eaterjolly@gmail.com wrote:

Considering how complicated and deep the issue of libre firmware can go, I wish they provided more documentation.

sigh yeah, one of the first things i do when tracking a new board is, create a page and document the compile and board-bringup process.

the RK3288 was absolute hell for completely different reasons. the internet is *completely overwhelmed* by ill-informed "wannabe-a-hacker" wordpress sites claiming to have the world's most easiest, most best, "it's"-ism'd documented method of installing {insert free os here}. most of them turn out to be chroot startup methods or keep the UEFI-partition-infested variant of u-boot that google insists on spamming the world with.

booting any given processor is actually really really simple and straightforward, but you have to accept that it is at least a three-stage bootstrap process. once you recognise - and accept - that pattern, it's really quite easy to spot, and you don't *need* so much in the way of "documentation". finding the source code for the components (u-boot, linux kernel), *that* tends to be the main challenge, and you have to find the people who can tell you where to find those.

in the case of the RK3288, that turned out to be #linux-rockchip on freenode. it's going to vary on a case-by-case basis.

l.