Raspberry Pi OS 12 (Bookworm) released

[TheOldPresbyope]

There is no local DNS resolver running on my ISP-provided router. (The ISP is Comcast). 

The LAN is 10.0.0.x, the router gateway is 10.0.0.1, and the router is running DHCP.

Here we see the hotspot has grabbed 10.42.0.x  for its LAN. I imagine the subnet is user-definable but I haven't explored this.

Code:

kreed@pi400:~ $ ifconfig
eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
       inet 10.0.0.253  netmask 255.255.255.0  broadcast 10.0.0.255
       inet6 xxxx:xxx:xxxx::faa4  prefixlen 128  scopeid 0x0<global>
       inet6 fe80::6655:9a1c:2c4e:76d6  prefixlen 64  scopeid 0x20<link>
       inet6 xxxx:xxx:xxxx:xxxx:xxxx:xxxx:xxxx:e569  prefixlen 64  scopeid 0x0<global>
       ether e4:5f:01:29:07:f7  txqueuelen 1000  (Ethernet)
       RX packets 155  bytes 52781 (51.5 KiB)
       RX errors 0  dropped 0  overruns 0  frame 0
       TX packets 189  bytes 26374 (25.7 KiB)
       TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
       inet 127.0.0.1  netmask 255.0.0.0
       inet6 ::1  prefixlen 128  scopeid 0x10<host>
       loop  txqueuelen 1000  (Local Loopback)
       RX packets 10287  bytes 898671 (877.6 KiB)
       RX errors 0  dropped 0  overruns 0  frame 0
       TX packets 10287  bytes 898671 (877.6 KiB)
       TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

wlan0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
       inet 10.42.0.1  netmask 255.255.255.0  broadcast 10.42.0.255
       ether e4:5f:01:29:07:f8  txqueuelen 1000  (Ethernet)
       RX packets 5869518  bytes 6426459127 (5.9 GiB)
       RX errors 0  dropped 127  overruns 0  frame 0
       TX packets 1724949  bytes 197714515 (188.5 MiB)
       TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

Code:

kreed@pi400:~ $ cat /etc/resolv.conf
# Generated by NetworkManager
search hsd1.md.comcast.net
nameserver 75.75.75.75
nameserver 75.75.76.76
nameserver xxxx:xxx:xxxx::1
# NOTE: the libc resolver may not support more than 3 nameservers.
# The nameservers listed below may not be recognized.
nameserver xxxx:xxx:xxxx::2
nameserver xxxx:xxx:xxxx:xxxx:xxxx:xxxx:xxxx:74b8

Note that I have full IPv6 service from my ISP and I obscure IPv6 global addresses out of (what may be) an over-abundance of caution. Someday I need to get more familiar with IPv6 security. techniques Someday.

Regards,
Kent

Interesting. My network uses the 192.168.1 address space. I'll have to see what network the AP assigns. 

Also prolly need to investigate how to assign a static address space for example 172.42.1.x if needed. I'm thinking it might not since lately it seems that pretty much everything supports mDNS or some sort of auto discovery.

Looks like it always assigns the same 10.42.0 network so no need for any static addressing.
inet 10.42.0.1  netmask 255.255.255.0  broadcast 10.42.0.255

- Connection hot switching works fine as long as you are first ssh'd in to the Ethernet address
- AP mode is fully bridged to the Ethernet side forming a true Hotspot.

It's a super nice simplification of networking. Things like dhcpcd, hostapd, bunch of config files etc don't appear to be  needed anymore :-)

Btw, you can use PSK's generated by wpa_passphrase instead of plaintext passwords for the psk= param. Also to generate UUID's just cat /proc/sys/kernel/random/uuid

Turns out to be easy to assign the static network 172.24.1.1/24 to the AP. Just add another line to the [ipv4] section

Code:

[ipv4]
address1=172.24.1.1/24
method=shared

Now the AP comes up as 172.24.1.1 and, for example, my phone connected as 172.24.1.238.

There's so much to nmcli and some newer features---such as hotspot---have options that aren't documented as well as I'd hoped.

Regards,
Kent

[Tim Curtis]

Turns out to be easy to assign the static network 172.24.1.1/24 to the AP. Just add another line to the [ipv4] section

Code:

[ipv4]
address1=172.24.1.1/24
method=shared

Now the AP comes up as 172.24.1.1 and, for example, my phone connected as 172.24.1.238.

There's so much to nmcli and some newer features---such as hotspot---have options that aren't documented as well as I'd hoped.

Regards,
Kent

Cool.

[Tim Curtis]

Boot time is like 5 secs :-)

   

[TheOldPresbyope]

Congrats on scoring a Pi5 and, yep, they're speedier.

I was overthinking the local display issue. After whittling down the cruft I found on the InterWeb™, here's what seems to be a minimum setup for starting chromium-browser in a local display on Bookworm Lite:

Code:

# First, install the wayfire package if it's not present already
pi@raspberrypi:~ $ sudo apt install wayfire

#create the configuration file to look like the following
pi@raspberrypi:~ $ cat .config/wayfire.ini
[autostart]
chromium=chromium-browser https://moodeaudio.org \
--ozone-platform=wayland \
--enable-features="OverlayScrollbar" \
--no-first-run \
--noerrdialogs \
--disable-infobars \
--kiosk

# Simply start wayfire
pi@raspberrypi:~ $ wayfire

How easy is that?! The wayfire windows manager takes over the screen and starts chromium-browser full screen.

The wayfire.ini file can have oodles of configuration sections and options (see Configuration  and the example wayfire.ini).

My first attempt at screensaving failed but I sure it'll work when I get time to play some more (real life and all that).

It's interesting to capture the output of wayfire to a file and examine the actions as it, for example, figures out that it's HDMI port 1 which is active (HDMI-A-1) and that, after enumeration of capabilities, the display mode is set to 1920x1080 @ 60.000Hz.

I dunno what caused the long delay I reported previously. It hasn't happened again.


Regards,
Kent

[TheOldPresbyope]

Bookworm Hints and kinks

- It's fun to use the Linux virtual tty capability to run something like the localui on tty0, say, track the journal log on tty1, and run interactively on tty2.

-By default, Raspberry Pi/Debian Bookworm doesn't log to /var/log/kernel.log and /var/log/syslog. The journalctl command is your friend.

Regards,
Kent

[Tim Curtis]

Cool. Maybe Firefox is also worth testing since its also included in Bookworm.

[TheOldPresbyope]

Cool. Maybe Firefox is also worth testing since its also included in Bookworm.

I'm not well versed in Firefox's kiosk mode but I gave it a try.

1.. install firefox package. That pulled down some 300MB more stuff. Lots of libs, etc.

2. changed the wayfire.ini file to make the only autostart program line

Code:

firefox=firefox --kiosk https://moodeaudio.org

3. from the command line, invoke wayfire

Result: it takes some 40 seconds for the firefox window to appear with the moodeaudio homepage.

By comparison, it takes only 4 seconds for the chromium-browser window to appear with the moodeaudio homepage.

Also, running firefox consumes 300MB more memory than running chromium-browser.

Both of these browsers have built-in preferences. I didn't try to edit the default settings for either. Perhaps one or more of the firefox preferences can be tweaked to shorten the load-in time and/or the amount of memory consumed.

Regards,
Kent

[Tim Curtis]

Cool. Maybe Firefox is also worth testing since its also included in Bookworm.

I'm not well versed in Firefox's kiosk mode but I gave it a try.

1.. install firefox package. That pulled down some 300MB more stuff. Lots of libs, etc.

2. changed the wayfire.ini file to make the only autostart program line

Code:

firefox=firefox --kiosk https://moodeaudio.org

3. from the command line, invoke wayfire

Result: it takes some 40 seconds for the firefox window to appear with the moodeaudio homepage.

By comparison, it takes only 4 seconds for the chromium-browser window to appear with the moodeaudio homepage.

Also, running firefox consumes 300MB more memory than running chromium-browser.

Both of these browsers have built-in preferences. I didn't try to edit the default settings for either. Perhaps one or more of the firefox preferences can be tweaked to shorten the load-in time and/or the amount of memory consumed.

Regards,
Kent

Case closed. We stick with chromium.

[TheOldPresbyope]

More on Bookworm/Wayland/localUI

Good news: the same uSD card with Bookworm and wayfire on it boots on my Pi5B, Pi4B, and Pi3B+ and runs either Chromium or Firefox using Wayland protocol rather than X11.

There's obviously some cache-building going on when wayfire and a browser are first invoked. Here's just the chromium-browser results but there are similar differences for firefox (and as reported before, always longer times than for chromium-browser)

Pi5B - wayfire/chromium-browser 1st run: 24s to moodeaudio homepage
           wayfire/chromium-browser 2nd run: 4s

Pi4B - wayfire/chromium-browser 1st run: 26s
           wayfire/chromium-browser 2nd run: 9s

Pi3B+ - wayfire/chromium-browser 1st run: 40s
             wayfire/chromium-browser 2nd run: 12s

If I reboot, then this process starts over.

Note that the transition from terminal screen to full screen moodeaudio homepage is rather busy visually: first, a period of dark screen with cursor outline in white; second, a period of white screen with cursor solid in black; then dark screen with rotating red circular line segments; finally, the browser window.

Second good news: the Pi5B and Pi4B were outputting to an HDMI LCD screen; the Pi3B+ was outputting to an official Raspberry Pi 7-in LCD touchscreen. I didn't notice any performance issue with the Pi3B+ relating to Wayland/wayfire. A USB mouse worked with either display; haven't tried touchscreen.

Addendum: Of course the touch function of the touchscreen display should just work...but it doesn't. I should have checked the repo for issues first: Touchscreen not working with latest Bookworm image.

So, ran rpi-update on my 10-10-2023 Bookworm Lite installation. The kernel was bumped from

Code:

pi@raspberrypi:~ $ uname -a
Linux raspberrypi 6.1.0-rpi4-rpi-2712 #1 SMP PREEMPT Debian 1:6.1.54-1+rpt2 (202
3-10-05) aarch64 GNU/Linux

to

Code:

pi@raspberrypi:~ $ uname -a
Linux raspberrypi 6.1.61-v8_16k+ #1696 SMP PREEMPT Thu Nov  2 16:49:24 GMT 2023
aarch64 GNU/Linux

and I can now use the touch function to make selections on the touchscreen display.


Regards,
Kent

[Tim Curtis]

Interesting.
- This "If I reboot, then this process starts over." seems odd or maybe its normal??
- I thought RpiOS Bookworm used X11 on < Pi 4/5 by some sort of detection mechanism.

Ya there will be some kernel and other updates before RpiOS Bookworm and the Pi Imager settles down. Pretty much par for the course for a major OS and hardware bump.