Steve Hay stvhay

EspressoBIN Ultra Notes

Get Connected

Initial Power On

Connect the MicroUSB port on the device to a computer. This is the serial UART.
Connect the power adapter and press the power button to turn on.
Connect to the serial UART via screen, minicom or other utility. One can see the UART device by monitoring /dev/tty* on the host computer and specify the device using the minicom -D option.
Log into the system with default password root/admin.

graph LR;
	scorp2kk_uboot(https://github.com/scorp2kk/u-boot/tree/master);
	radxa_kernel(https://github.com/radxa/kernel/tree/linux-5.10-gen-rkr3.4);
	radxa_repo_bsp(https://github.com/radxa-repo/bsp);
	radxa_uboot(https://github.com/radxa/u-boot/tree/stable-5.10-rock5);
	radxa_repo_rbuild(https://github.com/radxa-repo/rbuild);
	radxa_apt(apt repository);

	rockchip --> jeffyCN_xorg-xserver

Configuring and Using GPG+Freight

gpg_email="[email protected]"
reponame="aptrepo"
apt_manager="apt"
apt_distro="sid"
apt_archs="arm64"
apt_origin="stvhay"
apt_label="stvhay"

Preliminaries

adduser hays
usermod -aG sudo hays
visudo # NOPASSWD

References

Espressobin Ultra Build Sequence

sequenceDiagram
    participant gtibuild
    participant U_boot
    participant A3700_utils
    participant atf_marvell
    participant mv_ddr_marvell
    participant linux

Linode Network Setup with systemd-networkd

When you get into having multiple subnets routed to your machine, you have to turn off Linode's automatic network configuration service and do it yourself. This example is using systemd-networkd.

/etc/systemd/network/99-default.link

# This machine is most likely a virtualized guest, where the old persistent
# network interface mechanism (75-persistent-net-generator.rules) did not work.
# This file disables /lib/systemd/network/99-default.link to avoid

Dual Stack Networking for Docker, LXC, and LXD

Premise: Maybe one day IPv6 will be here, and we can stop forwarding so many gosh darn ports and just use a firewall. So with that hope, the goal of this document is to create container networks satisfying the following:

Continue using private subnets for IPv4, and gain control of which subnets to use so that you can route packets on your internal network appropriately.
Given a public IPv6 /48 or /56 subnet, dole out /64 subnets to each container network, which is the standard IPv6 subnet size.

I'm going to use one of my Linode subnets for this doc: 2600:3c03:e002:a200::/56

Docker and LXD are going to be given both each 16 /64 subnets to play with:

	#!/bin/bash
	typeset -A overlay kernel bsp

	defconfig="linux-defconfig-full.config"
	overlay=(
	[name]="overlay"
	[repo]="https://github.com/radxa/overlays.git"
	[commit]="main"
	)
	kernel=(

	#!/bin/bash

	pkg_dir="$(pwd)/debian-packages"
	mtune="cortex-a76.cortex-a55"
	march="armv8.2-a+dotprod+rcpc+ssbs+sb" -mtune=cortex-a76.cortex-a55
	# -march=armv8.2-a+dotprod+rcpc+ssbs+sb -mtune=cortex-a76.cortex-a55

	#function get_source_packages () {
	# pkg_file="$1"
	# cat $pkg_file \| xargs apt-cache show \| grep Source: \| awk '{print $2}' \| sort \| uniq > $pkg_file.sourcepackages

	#!/bin/bash

	CROSS_COMPILE="/usr/bin/"
	CFLAGS="-mcpu=cortex-a76.cortex-a55"

	timestamp="$(date -u +%Y%m%dT%H%MZ)"
	bsp="$(pwd)/bsp/bsp"
	rbuild="$(pwd)/rbuild/rbuild"
	dst="out/build-$timestamp"