hpr-knowledge-base/hpr_transcripts/hpr3705.txt

Episode: 3705
Title: HPR3705: The Year of the FreeBSD Desktop
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr3705/hpr3705.mp3
Transcribed: 2025-10-25 04:24:33

---

This is Hacker Public Radio Episode 3705 for Friday the 14th of October 2022.
Today's show is entitled, The Year of the Freebie SD Desktop.
It is hosted by BinRc and is about 69 minutes long.
It carries an explicit flag.
The summary is I talk about configuring Freebie SD as a desktop OS and give tips for those
coming from Linux.
On this episode of Hacker Public Radio, I'm going to be talking about The Year of the Freebie
SD Desktop.
I've received a few emails people asking me, hey I'm curious about Freebie SD.
How do I get a desktop installed?
How do I use this thing?
Can you help me more?
All I know is Linux.
People asking questions, asking for advice.
I thought it'd be a good idea to record an episode and I chose the title The Year of
the Freebie SD Desktop.
After being inspired by the Linux community, I still would consider myself a part of
the Linux community.
But every time there is a minor system update, you see various commenters saying, oh this
update is so good, you know, the Skynomchell update is so good, it improves the system so
much that this year, the current year will be the year of the Linux desktop.
We will see widespread adoption of Linux throughout the whole world.
I took the title there, or at least the inspiration from those types of commenters.
And community members, a little bit tongue in cheek, I think, the title of the year of
the Freebie SD Desktop and also the concept of the year of the Linux desktop.
I've always said that the year of the Linux desktop is the year that you begin using
Linux as your desktop operating system, your primary or exclusive desktop operating system.
Similarly, the year of the Freebie SD Desktop is the year that you start using Freebie SD
as your primary desktop operating system or your exclusive desktop operating system.
Now I'm not super big on exclusion.
I think another episode that may or may not be coming out, I talk about gaming on Linux
a little bit.
So obviously, there's not any exclusion when you use open source software.
You can use different operating systems for different purposes, and you know, don't
feel bad if you still have a Linux desktop and you want to be a Freebie SD desktop user or
vice versa.
It's all open source, it's all welcome.
So the first step to becoming a member of the Freebie SD Desktop year commission, I don't
know what type of group of people we could have that are involved in the year of the Freebie
SD Desktop.
So the first step is to get an installer, I have a link to Freebie SD downloads.
That's freebie SD dot org slash where.
So the first thing you need to do is pick an installer, pick the correct architecture
for your system.
Don't know a whole lot about systems architectures.
You probably want AMD 64 if you have a computer with like an Intel processor or an AMD processor
that was manufactured within the last 20 years, it's probably going to be AMD 64 otherwise.
If you have something special, I think you'll know what you need.
The rest of this, the rest of this sort of podcast is going to be about installing Freebie
SD on an AMD 64 system and configuring it as a desktop.
So when you click on the architectures button, you have a lot of different download options.
So let's go through them.
There's an installer called bootonly.iso.
It's a net install image, it's for burning to a CD.
There's disc1.iso, that's a supplementary image, CD image for bootonly.iso.
You have DVD1.iso, it's a complete DVD image, it comes with extra packages.
You have memstick.iso, it's a complete image for burning to a USB stick and you have
mini memstick.iso, that's a net install image for burning to a USB stick.
I typically just use the memstick images because I find it easier to install off of a USB.
Then to burn a CD and then re-burn a CD and I don't have a lot of computers with the
CD drive anyway.
So it doesn't really make a whole lot of sense for me to be burning CDs.
One of the memstick images is probably what you want and if it's your first time installing
free BSD, I don't recommend going with the mini memstick option.
Sometimes that makes installing wireless drivers difficult.
So once you have your memstick image downloaded, a couple of commands you can run.
X, unZ, so they distribute these.
I download the zipped ones obviously, it saves a band with X, unZ, and then the path to
your installation image.
And then I use DD, the disk destroyer program to write that installation image to a USB
drive and then I eject that USB drive.
So now that you have a free BSD installer ready to go, let's get on with the installation.
So the first step is pre-installation.
The standard steps for installing Linux apply, so you have to go into your BIOS, disable
secure boot, enable USB booting.
Depending on how you knew your hardware is, enable legacy booting.
The same type of stuff you would do for a Linux installation and then enable the default
boot device or select the boot device at startup time.
This is all hardware specific.
I almost assume that being a Linux user is a prerequisite for this episode, so you probably
know how to do all of these things.
So now actually installing the thing for BSD has a menu driven installer and it walks you
through various steps.
This program is called BSD install.
So the first thing BSD install asks you to do is set a key map.
If you have a strange keyboard, you know all about setting key maps.
If you don't know what a key map is, just leave it as default.
Set a host name and then the third step is to select sets.
There are lots of sets.
This is just, I guess modules of the base operating system.
And if it's your first time installing free BSD, you probably want to install all of them.
I typically only install the lib32 set and add the rest later.
I like to have a leaner system.
The fourth step is to partition your hard drives.
So BSD install makes it really easy to partition your drives.
The file system you want or the option you want to choose in the partition menu will
be called auto and then in parentheses ZFS.
You should probably go with ZFS.
The default UFS configuration is not journal that's not very reliable because it's not
journaled and ZFS is very robust.
So in the automatic ZFS configuration menu for a single hard drive installation, set
your pool type to one disk.
You want to stripe one disk and then select your hard drive.
If you want full disk encryption, pick encrypt disks.
And then for setting swap size, you want to bump up your swap size.
I think by default, it's only two gigabytes.
The general rule of thumb for swap size is however much RAM you have times 1.5.
So if you have four gigabytes of RAM, you want like six gigs of swap.
If you have eight gigs of RAM, you want 12 gigs of swap.
And then if you also are encrypting your disks, you should probably also encrypt your swap.
encrypting your swap space means for example, password you may are made on and turn a web
browser or stored in RAM and they're swapped out when they're not used and you don't want
those things floating around on disk.
So when you're done configuring your file system, you can proceed with the installation.
You'll see a confirmation message asking if you want to destroy the disks.
This is the last step to go back after you commit.
You're going to have ZFS on your hard drives.
And whatever you have on there before is now gone, reduced to ash, never to be seen again.
And if you did choose to encrypt your disks, you'll have a password prompt.
This is the disk encryption password, not any user password.
You should probably set these to be different things.
The fifth step, wait for sets to install, cool little progress menu shows.
The sixth step is setting a password for the root user.
The seventh step is network configuration.
So if your wireless card is already supported, all of the hard parts are done for you.
If your wireless card is not supported, you might have to plug in an ethernet cable and
either download those drivers or you'll have to compile the drivers, for example, I think
Broadcom, a lot of the Broadcom cards you have to compile the drivers into the kernel.
Hopefully though, your Wi-Fi card is just supported.
So select your card at the network config, EM, EM0, EM1, those are ethernet interfaces.
Wi-Fi cards are named after their drivers, so for example,
so I have a, my Wi-Fi card is named WLAN0.
Hey, I think this might actually be incorrect.
I'll open BSD the name after your drivers, but
no, I'm pretty sure on Fubius either also named after your drivers because I have an atheros.
USB Wi-Fi card and it's named like ath0 instead of WLAN0.
If you choose Wi-Fi, the installer is going to stand for networks and give you a menu to select one.
If the network's encrypted, you're given a password prompt.
Annic, totally sometimes it'll ask you to change the FCC region.
If you are having difficulty seeing any Wi-Fi interfaces show up, don't set the region,
just leave it as default and it typically just works.
It can be fiddly. The last time I installed Fubius, it wasn't.
I reinstalled Fubius for this episode, it was fiddly.
But if you can't get it, try different FCC regions.
For example, no region versus US region.
The 8th step is time and date setup. That's really simple. You just set the time and date.
Then you set up services so you're presented with a menu that allows you to enable or disable
services on systems startup.
For most people, you probably want all of them except local unbound,
local unbound is like a local caching DNS resolver.
If you have never used one of those things before, it can cause DNS to behave in a way you've never
experienced and you might become frustrated. Typically, I don't enable that one, at least for a
desktop that's just going to be like a DHCP client. The 10th step is to configure security.
You can enable or disable security features. The more of these you enable, the more confusing
things can become, especially for someone who might be unfamiliar with FreeBSD.
If nothing else, disable send mail and enable the clearing of the temp partition, you can do the
other things if you really want to or if you really feel like it. But that's typically what I do
just for a desktop installation. The 11th step is to add users, so you just add your user.
You might want to add your user to the wheel group if you plan on using sudo.
I also set my shell to tcsh in this menu. You can always change your shell later.
Another thing about the wheel group. A lot of the hard parts about using FreeBSD as a desktop
operating system, a lot of the manual configuration you might have to do to be able to access a lot of these.
I guess lower system things that you know you would expect a desktop operating system to do,
you probably want your user in the wheel group. The final configuration you might want to
install the handbook or modify any configurations, installing the handbook takes a little bit of time.
And then when you're done, apply the configuration and exit. And then before you reboot,
you're given the last option to do any manual configuration. Most of the time you don't need to do
this. After the installation, now it's time to install a graphical user interface. I quoted
myself, what? No gooey? Yeah, there's no gooey by default on FreeBSD. So with a fresh installation,
we want to update the system. Those commands are FreeBSD-updateFetch, FreeBSD-update install,
and reboot. So running those things will update the base system.
The next thing I usually do is install them because I don't like using VI.
I don't want to have to use Ed when I don't have to and I don't like EE.
So we need a better text editor before we start editing all these files. On FreeBSD,
the package utility is what we manage packages with, at least binary packages anyway.
It functions almost identically to any Linux package manager you've used the syntaxes,
PKG, verb, object. That's how apt works, DNF works, for example. On FreeBSD, the only editors
installed by default are VI, Ed, and EE. If you've ever used nano, EE is kind of like nano.
But I like to install VIM. There are a couple of VIM flavors. I like the VIM-tiny flavor,
just because it doesn't pull in as many things. Like I said, I like running a lean system.
You probably also want to install Sudo or do as. So package install Sudo and then run the VI
Sudo command to edit the Sudo or style and then find the line that says wheel.
And remove the leading hash because we want the wheel group to be able to do actions as root.
The next thing I do is a couple of bootloader tweaks. We can tweak the bootloader to make the
system boot faster. So how you do this is you edit a file slash boot slash loader.com.
There's a lot of default stuff in there. Don't touch that too much if you don't know what you're doing.
And then I typically go to the very bottom of that file, add a comment that says custom stuff
below and then comment what each option does. So I add a line that says auto boot underscore
delay equals to so it'll boot faster. I think I should say before I continue. I'm not doing
a lot of really implementation or I guess hardware specific stuff. This is a pretty generic
free BSD installation. But the reason I'm editing these files is sort of to show you where they're
at and show you how to, for example, edit the bootloader, edit the init system, so on and so forth.
I'm not doing a lot of the really custom stuff that might make things behave strangely and in a
non default way. So to learn more about what tweaks you can make to the bootloader, read the loader.com
man page. The free BSD also comes with default configuration examples. So if you read slash boot,
slash default slash loader.com, that is a default configuration file for the bootloader.
Lots of good examples in there very well commented. The other thing we can do is tweak the init
system. Oh boy, my notes got out of order. That's fun. We can tweak the init system by editing slash
Etsy slash RC.com. So the configuration I had in the init system is, I append a line that says
background underscore DH client equals yes. This makes the DH client start in the background.
And sort of gets you to a desktop faster maybe than if you were waiting for a DHCP lease from your
DHCP server. If you read the man page for RC.com, it'll tell you all about the different things you
can do with RC.com. If you read slash Etsy slash default slash RC.com, that is an example file filled
with comments and lots of examples on what you can do with the init system. At this point, once I
have everything sort of a sane base fresh installation, I like to take a ZFS snapshot.
This makes it really easy to roll back to a fresh known working system configuration.
So how do I do that? You run the command ZFS snapshot dash R,
Z root at fresh install. So I'm snapshotting the Z root, which is the root of the ZFS pool.
And then I'm naming the snapshot fresh install. And then to list your snapshots, you can do ZFS
list dash T snapshot. So when the system becomes unreparable, we can roll back instead of having
to reinstall with a simple command. So how you roll back is ZFS data set ZFS rollback optionally
dash R, Z root at fresh install. To roll back every data set because ZFS rollbacks aren't recursive,
you can use a little bit of XArg's magic. So how do I do that? ZFS list dash T snapshot, pipe,
grip, fresh install, pipe, cut, dash D, in quotes of space, because we want the delimiter to be a
space, dash F1, pipe XArg's dash capital I percent ZFS rollback percent. A little bit of
shell magic to get recursive rollbacks on ZFS. I think it's a good idea to take ZFS
snapshots before you make any potentially dangerous configuration. And possibly after you've
done something that's kind of difficult to configure, it saves a lot of headache in the long run
because ZFS is accessible from the recovery shell. The last word of advice is to roll back with caution.
If you run a recursive rollback, for example, and you roll back your home pool, I believe that's
what it's called. Yeah, if you roll back your Z root slash user slash home, if you roll back your
home pool, all of your user data will be lost since that snapshot. So proceed with caution.
And then a little bit of homework, write a series of cron jobs that automatically take snapshots
and clear up the old ones. Sort of as a last line of defense version control.
Now for the part everyone's waiting for graphically user interfaces. So the first thing we need to do
is install graphics drivers. This varies depending on your GPU, but generally you'll just run
package install DRM-Kmod. After you install a package, you'll get a message. That's the cool
thing about free BSD is packages will actually give you a message telling you how to use them
after you install them. So the message from the DRM-Kmod package says, how do you install this
driver? So it says for AMD GPU, you enable it one way for Intel GPU, you enable it another way
for radio and KMS, you enable it a different way. So to enable one of these, you need to add a line
to your init system, right, slash Etsy slash RC.com. The earlier you place things, different lines
and starting services in this file, the sooner they start. So for example, since I have intel
graphics, you add the following line to enable intel graphics in the file slash Etsy slash RC.com.
So you append a line that says KLD underscore list equals in quotes i915 KMS. When you boot the
system, that loads the intel graphics drivers. To load kernel modules on the fly, you can use the
the command KLD load. The interesting thing about graphics drivers on free BSD is it also increases
the resolution of a VT. VT is like the virtual terminal. It's kind of like a frame buffer console,
but not quite. You just run the command KLD load i915 KMS. I've noticed if you unload
graphics drivers on the fly, it crashes the system, but that might just be anecdotal.
The last thing you need to do is add your non-root user to the video group. You have to be in the
video group to be able to use these drivers. So how do we do that? PW, run the command. PW,
group mod, video, dash m, and then your user name. Again, notes out of order, but I'm going to
continue anyway and upload this in this way because the commentary on notes out of order is kind
of fun. So audio, hopefully audio just works. Sometimes audio doesn't, but supported audio
interfaces are enumerated in the SND man page, and details about enabling and disabling these
drivers in slashboot slash loader.com are also explained in the man page. How you manage volume
on free VSD use the mixer command. So to set the mic volume to 50% and the speaker volume to 90%,
95%, you run mixer, mic, 50 colon 50. That'll set the mic's left and right audio channels to 50%.
mixer, vol 95 colon 95, that's the output volume to left and right output volumes for the speakers to
95%. You can also use plus or minus, I believe. So for example, mixer, mic, plus five will increase
5%. I think it works that way. The mixer, TUI command can also be used. This one's not installed
by default. You'll have to install mixer, TUI. This program functions almost identically to
all some mixer on Linux. It makes, it gives you a terminal based interface, like a graphical
terminal interface to modify volume on various audio interfaces. And then depending on your hardware,
the volume keys on your keyboard might not work. I typically add a key binding that calls a
shell script or even just a command to increase or decrease volume. This is like the typical
solution for people who run tiling window managers on Linux, people who really run any window
manager without a desktop on Linux. So this should be a familiar workflow, TUI. I think a lot of
the types of people who are interested in running free VSD. The next thing I do is install XORG.
How you do that? Package install XORG. That's the command you run. The TWM window managers included
with XORG by default. And this is useful because we can test our XORG configuration, test our
mouse support. Before we install a larger desktop environment that may or may not have problems
of its own. So early troubleshooting really prevents a lot of headache, you know, test early
test often. How you test your XORG after installs start X. If it works, you should CTWM start up.
Hopefully, again, hopefully XORG configurations just work out of the box. If they don't work,
troubleshooting XORG and getting a functional XORG or X11 configuration.
That is really operating system agnostic. That's almost identical on every
Unix-like operating system except for the fact where the configuration files are stored.
Now on to desktop environments. So you can refer to the handbooks instructions on
desktops. I have a link for that.
I've tested three of these on free VSD. So KDE works XFCE works pretty reliably.
Genome is sometimes reliable. If you're running a big desktop environment, you might have to modify
the Paul kit rules, P-O-L-K-I-T policy kit rules to do things like reboot the system from the GUI.
A lot of these larger desktops rely on free desktop.org components. I don't really like those
types of things. Like I said, lean system. That's what I like. So I don't like having debust running.
I use the cyclist tools, but for the sake of completeness, I will install each of these desktop
environments for the masses. I installed each of these desktops independently and sequentially
on the same system using ZFS snapshots to roll back to a bare-bone system without any desktop
environment installed. This is another good use for ZFS snapshots. For example, you want to test
something that might break the system or might create a lot of garbage files that you don't
necessarily want on the system. It's really useful to be able to snapshot things and roll back.
So how you install Genome? You're on the following commands package. Install Genome. After that's
done, you have to add ProcFS to your Etsy FS tab. I did this with a printf,
but if you want to edit that file in VIM, your FS tab in VIM, that just looks like Proc, tab,
Proc, tab, ProcFS, tab, RW, tab, zero, tab, zero, new line. That'll enable the Proc file system.
And then to mount that on the fly, you would do mount dash A.
To enable these things, all of the services required to get Genome started.
CISRC debuts underscore enable equals in quotes capital YES. So yes, in capital letters.
CISRC GDM underscore enable equals in quotes yes and all caps.
CISRC Genome underscore enable equals in quotes yes and all caps and then reboot the system.
When your system comes back up, you should see the GDM display manager you should be able to
log into Genome. I was actually surprised about how well Genome was functioning on FreeBSD 13 point.
Am I on two or one? I can't remember which version I'm running. FreeBSD 13.1.
In the past, I've had a lot of troubles with Genome, but it seems to work fairly reliably now.
At least Genome on FreeBSD used to be difficult for me in the past. I then rolled back and installed
KDE, so how you install KDE on FreeBSD. Package install KDE5 SDDM. Again, you will add the ProcFS
here at CFS tab. Same way as above. Then you will enable a couple services. CISRC debuts underscore enable
equals yes. CISRC SDDM underscore enable equals yes and then reboot and you should get SDDM.
So the interesting thing about the CISRC command is it actually just modifies the file slash
Etsy slash RC.com. It modifies your init system. These CISRC commands are identical to just
opening up your Etsy RC.com with a text editor. For example, adding a line that says debuts underscore
enable equals yes. As far as I know, these commands mostly function like just insert a line here,
modify a line here. That's how you install KDE and then I have a screenshot of KDE running.
I'm not sure if I said it, but I also have a screenshot of Genome on FreeBSD above.
I then roll back and I install xfce. So how you install xfce package install xfce
xfce4 dash goodies. The goodies package pulls in a lot of the
xfce applets and things that make it more usable than just a window manager.
xfce doesn't provide its own login manager unlike Genome or KDE. So we're going to install
LightDM because it's fairly small, fairly fast. The graphical toolkit matches xfce.
It looks good together. So how you install LightDM package install LightDM-GTK-greeder
and then you enable LightDM with a CISRC command. CISRC LightDM underscore enable equals yes.
Another way you could accomplish this would be like
echo LightDM underscore enable equals yes and then two angle brackets, right. The append
output redirect and then append it to your EtsyRC.com. I reboot the system. You should
see LightDM be able to log in to xfce. Then I roll back and I finally install what I wanted,
which is suckless. So the suckless tools, they're tools that suckless. I have a link for suckless.org.
They're fairly small tools. The philosophy is really hey, let's
we need, although we have to use x11, we can still make desktop environments and window managers
that are modular and not really a pain to use. That stay out of your way and let you get work done.
I typically run suckless software on most computers. I think except for Fedora,
I just run Genome on Fedora because I can't be bothered to install suckless when
I can get a full screen terminal running Genome identically to how I can get a full screen terminal
running suckless software. So I just wrote a make file that bonifies the compile options so that
these tools build on FreeBSD and automatically patch them to add a theme. The way you modify
suckless software is by patching the source code or the header files or the
make options. So you have to modify each of those components of software individually for whatever
system you're on. So I just wrote some suckless duct tape that I installed suckless on FreeBSD
and writes my plan 9 theme. The make file is fairly large but pretty much what it does is just
pulls down all of the get repositories for the suckless software I use.
I have a make option to patch them, install dependencies, add a patch for themes, build and install them.
It makes it easier. At least for me when I reinstall these things I like having a consistent
a consistent way of installing think of really quickly. Although I don't
and reinstall operating system very often I thought I'd just do it this time around.
And then in addition to using the suckless tools I just use XDM. It's small and fast,
it's configurable if you care. I typically just leave it how it is except for disabling the console.
How you install XDM, pseudo package install XDM, and then pseudo service XDM enable.
In FreeBSD there are a couple ways to modify your
init system. And then I have a screenshot of suckless software with my fancy plan 9 theme
and web browser open. So final note on desktops before I continue to some more general things.
Sometimes the desktops behave unexpectedly like users can't manage power settings or reboot
the system. If the user you're logging into a desktop environment is with logging into a desktop
environment with is in the wheel group. Most of these issues are resolved for the non-wheel group
users. You have to write policy kit rules, poll kit rules. I used to do this in the past. Now I
just add every user to the wheel group. But to be fair I'm the only one using my systems as a
desktop operating system. I don't share my computer very often with people. Especially considering
that I'm using like DWM and really a terminal only interface. It's kind of hard to share that
with an average person. Another thing to watch out for a lot of the big desktops are compiled without
support for modifying network connections. So for example there's not a Wi-Fi menu in the
GNOME settings. That may or may not be something you care about. But it's just as easy to configure
those things on an operating system level. I think almost easier because you know for sure it's
going to work. I'll get into network configuration in a minute. Similar to Arch and Gen 2,
there's a little bit of legwork left for the end user. I know a lot of people might not like
this little bit of legwork. You know modifying configuration options, tweaking things to get
things to work. How do you like them to work? But I always say you know this is a good thing. You
never know what you might learn if you don't give yourself the opportunity. So there's a lot of
weird operating system or even like desktop specific things that I've learned just because I've
put them on freeBSD. You know I've never modified a poll kit rule in my life until I started using
freeBSD. You might not even know that poll kit rules exist if you don't use freeBSD. I think exposing
yourself to strange and unfamiliar systems. It builds a lot of skills that you might not ever
have been aware of before. So you have to give yourself an opportunity and you have to stay
diligent. But most of the time everything just works. At least in my anecdotal experience.
So another thing I do, I make shell tweaks. So I like colors in my system shell.
I don't use colors on OpenBSD. That was kind of a strange thing. I also like aliases. So because
the default shell on freeBSD is CSH, we can modify the configuration file to automatically do
the fancy for us. So in my CSHRC, I leave the default stuff in the prompt section.
I have a couple sampled prompts. I just set my prompt to be bold and then leave it at that.
I think I have a colorful one in there too. One of those might be red instead of bold.
You know, comment or uncomment if you want to copy my configuration.
And then in the bind key section, I add the control R keybinding to do reverse history search.
Kind of like how you would use the control R keybinding on bash to search through history.
I then add some aliases. So ls alias la to ls dash a capital f alias lf to ls dash capital f capital a alias ll to ls dash a capital
I lost that. So my l alias is for the following command ls dash l capital a capital f i alias
regular ls to ls dash capital g capital f. So the capital g gives you colors. The capital f gives you
like a slash, for example, if it's a directory, I think a star, for example, if it's executable.
Yeah, so it gives you a star if it's an executable.
Some interesting things. And then I alias lc to ls dash g capital g capital f.
The reason I alias lc is because on plan nine, lc is equivalent to ls dash dash column or
something like that. I think it's, yeah, so lc is the command on plan nine to give you a multi-column
ls output. And the muscle memory, it's a lot easier just to leave the muscle memory and adjust the
aliases. So some other packages you can install. The things that I like to install.
Firefox, Gimp, FAT, MPV, FFMPEG, ImageMagic, MUT, and NewsBoot. If you've installed a large desktop
environment, most of the applications are automatically pulled in. If they're not pulled in,
you can use xargs to pull in hundreds if not thousands of gigabytes, possibly up to petabytes
of programs. So how you do that pseudo package search, whatever your desktop environment is,
you pipe it through cut. You're going to delimit it with a space, take the first field,
and then pipe it through xargs, pseudo package install dash y.
FreeBSD package is kind of interesting. You can, I don't think it works with regular
expressions, but if you do a package list or a package search followed by a cut and then pipe
it through xargs, it'll just run. Instead of running each of those package commands,
it can currently, it'll just put them all together and run one big package command. It's really
cool. If you want to go GNU, you can do sudo package install core utils, emacs, bash, gcc,
gmake, whatever other going to do utilities you want. They're there, they're available.
I don't use them, but if you like bash, maybe more than CSH,
why not, right? That's how open source is all about making your system do what you wanted to do.
Using the computer, how you want to use it. So if you want GNU on top of freeBSD, why not?
I think the only caveat I'll say here is don't change a lot of the root users default stuff to
GNU stuff. Things start to break when you change the root user shell.
But that is also the same on Linux. For example, you don't want to change your root user shell to
one of those wacky, wacky shells people use just so they can look cool on forums like fish or
CSH. You don't want to do that. Now that you have a desktop configured, it's another good time to
take a ZFS snapshot. So you always have a working desktop configuration to go back to. For example,
if you break it in the future, you can always do a ZFS rollback, get back to your working desktop
configuration, and then proceed from there. The last section I have, actually this is not the last
section, is a quick start section. So this is a lot of, this is sort of where I take the assumptions
of people knowing how to use Linux and build on top of them. So instead of SystemD,
FreeBSD uses RC scripts for starting and stopping services. Everything is pretty much just out
shell scripts. To modify the processes that start at boot, you can edit slash Etsy slash RC.conf.
I like being able to edit my init system with the text editor. It's something that I really,
really miss when I have to use a Linux system. If you are familiar with like a system CTL
command on Linux, you can start and stop services this way. Using a different command, the service
command. So service shshd enable service shd start service shd restart stop disable one start status.
A lot of the similar system CTL commands you might be familiar with on Linux, you will just use
the service command on FreeBSD. I believe the service command on Linux also works the same.
The only thing here is that each service has its own init file. So sometimes a service might take
different arguments than one you might expect. So for example, if you modify your sshd service to
maybe use a different sshd config, you could do like service sshd start dash alt config.
Because it's all shell scripts, it's fairly easy to do.
And the other thing I should say, the syntax is backwards. So it's service, the service you want to
start or stop, the service that you're supplying an argument to, and then an action with system D,
it's system CTL verb object on FreeBSD its service object verb. This syntax messes me up sometimes,
but it's not a huge deal. You realize as soon as you do it that you've done it, you just have
to flip a couple words around. It's not super difficult. Networking. So network interfaces are
configured classically using the if config command. If you want a network interface to persist
across reboots, you add this information to slash Etsy slash rc.conf. A lot of things are
managed in the init system, including for example, you want to start a loopback device or a
bridge device, you'll just add that interface to your init configuration. And then when you
boot the system, that bridge devices automatically started that extra loopback devices automatically
started. Wi-Fi is managed with a program called WPA underscore supplicant. Read the man page for
more information. If I actually go to my file. So WPA supplicant is
fairly easy to use. So you just edit the file Etsy slash WPA supplicant.conf. And then you just add
a network and you can set priorities, add keys, so on and so forth. It's a plain text way of
configuring your Wi-Fi networks. And the only thing that ever trips me up with WPA a
supplicant is that when you assign a network a priority, higher values have a higher priority,
as opposed to lower values having a priority. Sort of intuitively, I think, oh, priority zero,
that's going to be like nice levels where a lower number has more priority. No, it's the opposite.
As for firewall, I have a link to the free BSD handbook. Use the PF firewall. The PF firewall
is very easy to use. Again, plain text configuration. A very simple PF configuration can be like one line
or two lines. Let me find a sample PF configuration. I will just read it to show you how short it is
compared to something like NIP tables configuration.
Yeah, so my PF configuration on my web server, set skip on L0, so set skip on the loop back
device. TCP underscore services equals in quotes. And then in brackets, SSH comma HTTP comma HTTPS
block in all pass in proto TCP to any port dollar TCP under services keep state pass out all.
That is all the firewall configuration you need to block all ports except for SSH HTTP and HTTPS.
That's 2280 and 443. I really like PF because it's simple and you can also, it's really flexible.
There's a lot of things you can do, for example, read a list of IP addresses or from a file and then block
those. There's lots of cool things you can do with PF and it's quite, I like it quite a lot even
though I do have a little bit of an affinity for firewall D.
The next thing I talked about, the next quick starter, the general upgrade process. So on FreeBSD,
the packages you install are separate from the base system. The base system is managed separately
from the packages. If you keep this in mind, you're going to be okay. So to update the entire system
from top to bottom, base and packages altogether. The commands look something like package update
and and package upgrades. So when I say and and that means 2 ampersands. You then run the command
FreeBSD-updateupgrade-r and then the release, in my case 13.1-release. When 13.2 comes out,
you use 13.2-release or 14.0-release. FreeBSD-update install and then you reboot.
And then you reboot the system and log in as root and then you run FreeBSD-update install again.
You want to update and upgrade your packages again that way,
if there are any major standard library or or A, B, I interface changes, they will be upgraded
and just work again. And then you run FreeBSD-update install. One more time, this will remove all of
the unnecessary A, B, I, A, B, I stuff from the package update. And then when you reboot,
your whole system will be new, all of the packages will be updated if there's any library changes,
for example. And everything should just work. This is a little bit more
comprehensive than, for example, doing a major version upgrade on something like Fedora.
But it's largely because on Fedora, the base system and the ancillary user installed packages are
both managed to the same package management system. Whereas on FreeBSD, the base system is distinct
from the user installed packages. This is for stability purposes. As for shells, FreeBSD
uses TCSH as the default shell. It includes SH for born like compatibility. So when you do
shell scripting, I always say just do POSIC shell because every single unit like operating system
as a POSIC compatible shell or even a system with BASH, if you run BASH, DASH, DASH, POSICS,
I believe, I'm going to check that. Yep, if you run BASH, DASH, DASH, POSICS, it will run in
POSICS compatibility mode. You can install BASH if you want on FreeBSD. I typically just use the
default shell wherever I'm at. But you know, right, right, POSICS compatible shell scripts.
And then your shell scripts will work on Linux. It'll work on FreeBSD. It'll work on
something strange like Illumos. Not Plan 9 though. Not Plan 9.
Package management. That's the next quick start section. On FreeBSD, there are two main ways
of managing software. There are binary packages and ports. Don't mix binary packages and ports.
If you don't know what you're doing, it can cause problems. So a brief explanation of ports,
ports are like Gen2. You download a port tree. It's a series of recipes to make the software.
You end up spending a lot of time watching compiler output, possibly manually troubleshooting
dependencies, resolving dependencies. But there are a few programs that help with this.
So synth is a program that builds ports, port master is a program that builds ports,
and PUDRIARE, PUDRIARE. I never know how to say it. That's useful for creating like a local binary
package repository. That's what PUDRIARE is for. It builds all of the ports and then you can
serve them as a binary package repository. It's very useful. And then to be more verbose about
using the binary package management system. So for example, package install FUBAR,
package search FUBAR, package remove FUBAR, package install FUBAR again.
Package auto remove, package update, package upgrade. Really similar stuff to like a DNF or an
app. The file system, I guess I should probably have titled this hierarchy. But the hierarchy of
FUBAR is different than a Linux system. Read the man page on hire. That's man, H-I-E-R.
It tells you why things are the way that they are. So the biggest difference on FUBAR
is that it's logically organized. On Linux, everything you install ends up in slash bin,
which is actually just a simlink to slash user slash bin. And slash S bin is just a simlink
to slash user slash S bin. On FreeBSD, the file system is a lot more organized. There's a
stronger concept between base operating system and random stuff the user installed from the package
management tools. This is useful because of a package breaks, for example, the system still
boots. So on FreeBSD, slash bin requires contains everything required to boot the system.
Slash bin contains everything required for fundamental administration. Slash user slash bin
contains all of the other non-essential binaries that are part of the base system. And then
everything in user local is what is installed by the package management system. So user local bin,
that would be, for example, where VIM is installed, perhaps where Audacity or Firefox is installed,
right? Another thing, user installed programs are configured in slash user slash local slash Etsy.
This is a little bit confusing at first, but you get the hang of it when you start to think,
user local is where all of the other stuff is. Everything in the root directory is default
system. Everything in user local is what I've installed. So, for example, a document root for
like Apache on FreeBSD. It's not far dubbed up HT docs. It would be user local far dubbed up HT docs.
It gets long, sometimes your prompt gets long, but you get the hang of it after a while,
and it starts to make a lot of sense and really becomes enjoyable to use the system that way.
Another issue with this logical separation is it can cause confusion when compiling software
from source, specifically not from the port system, but like something you download off GitHub,
or for example, all of the suckless tools I compile, I have to manually modify the linker options,
and sometimes the make file options, so that the linker knows where to look for libraries at.
If you know how linker options work, if you know how make files work, a lot of times you can just
change, for example, slash user include to slash user local include, and it fixes a lot of things.
Things usually just compile. As for actual file system support, apparently I didn't know this,
but apparently extended to extend the three and extended four. Now I have read right support
using the EXT2FS driver. The last time I looked into extended four support on FreeBSD
was probably about half a decade ago. It might have existed then, but I wasn't aware of it.
I probably wouldn't try to boot from extended four, honestly, but these things exist for
compatibility purposes. The other big file system, UFS, it's not journaled by default,
proceed with caution. If you don't know what file system to pick, just pick ZFS, I promise,
a lot of the complicated ZFS stuff, including snapshots, you don't even have to be aware of snapshots
to use ZFS. You can use it as a very dumb file system if you want. I then have a section called
ZFS nonstarter. I'm not an expert on ZFS, but I sort of show you some of the cool things about
ZFS in this shell session. So I write ZFS is cool because we can create partitions on a whim, we can
create partitions, and ZFS partition is called data set. I think of it as a logical partition,
except unlike Linux, LVM, it's a lot more flexible. So we can create data sets, create data sets with
a quota, destroy data sets, create and use encrypted data sets, so on and so forth. So ZFS list,
list all your data sets, ZFS list dash T snapshot, list all your snapshots,
how you create a data set, ZFS create Z root slash the path to your data set. So I create
Z root slash crypt. I'm thinking I want to make an encrypted partition. So I set the quota,
ZFS set quota equals 5G on Z root slash crypt. I listed it tells you it's only using 96 kilobytes,
it has 5 kilobytes available. That is the quota. I then realized I didn't turn encryption on.
So I ZFS destroy Z root slash crypt. I then recreate it. ZFS create dash O encryption equals on
dash O key location equals prompt dash O key format equals pass phrase Z root slash crypt. It
prompts me for a password. I enter it. I then ZFS list Z root slash crypt. It tells me hey,
this thing exists now. I make a file called super secret and slash Z root slash crypt,
list that file. I get encryption information. I unmount, unload the key, reload the key,
so on and so forth. Additionally, you can modify mount options with ZFS. I didn't do this in that
example, but I probably wouldn't mount my crypt partition at Z root slash crypt. I'd probably
mount it as a sub directory of my home partition. ZFS is really flexible and a lot of the things
you can do don't require like modifying your FS tab. It's all done through ZFS. It's really,
it's a really cool file system and I need to use it more. You know, I know quite a bit about
LVM Linux logical volume management, but that sort of has its own drawbacks compared to ZFS.
I really like ZFS. You know, ZFS even does a raid. I only run it. I've only ever ran it on one
disc at a time, but you know, the ability to create partitions on the fly, the ability to set quotas
on the fly, the ability to change mount points on the fly. It's all really powerful stuff and I'm
kind of surprised there's not something like this and Linux already. And I almost wish there
was ZFS on BSD, on open BSD because I would just use open BSD forever and ever, but we gave what we
can take. So my conclusion here, I really think free BSD is a viable desktop operating system,
especially for the types of people who already use Linux in a terminal-centric way. You know,
Unix is Unix. I think for anyone who's ever installed like Arch or Gen2, or gone from a Debbie
and minimal installation to a graphical installation, a graphical environment, you know, something
as big as GNOME or something as small as like open box. I think free BSD is not, it's not really
all that difficulty, you know, so sort of demystifying free BSD. It's just a Unix. It's just like
a Unix because it is a Unix and a lot of the things you do are familiar because, guess what,
X11 is the same everywhere you go. Operating systems or Linux distributions that use WPA
Supplicant, same as everything else. You know, the really the big major differences are from
a Linux, no GNU, no system D, and maybe a better way of organizing the system where instead of
dumping everything into Slash Bin, you instead dump everything into, you know, an Ancillary
Directory link. This is the garbage I've installed from the software repositories containment zone,
everything else that is actually, you know, I should take care of that's over here. So I have
some other stuff. I don't run Firefox inside of a jail, but I have two links on how to run Firefox
inside of a jail. Jails are like Cherutes, jails are like Docker containers. The only difference is
jail is slightly more secure than a Cherute. And unlike Docker, when you build a jail, it doesn't
have the aspect of downloading random stuff off of GitHub and downloading random operating systems
off of GitHub and Docker containers off of GitHub. And blindly trusting that the guy who set it
up actually knows what he's doing. And then I have a list of free BSD distros that come with a
desktop out of the box. Ghost BSD is free BSD with the Matei desktop Hello System. It's free BSD
with an Apple like GUI. I believe Hello System is still in development, but I'm kind of curious
to see. I think the Apple desktop is called like Aqua or something. I'm kind of curious to see
what an Aqua like desktop would be on free BSD, especially considering that large portions of
Apple software is pretty much just BSD. Another one, Midnight BSD is free BSD with XFCE and a
different package management system. There is also Nomad BSD. It's a live GUI free BSD with
open box. I'm probably wrong about this, but the way I think of Nomad BSD, I conceptualize it
similar to Puppy Linux. But you know, you can obviously install it on a hard drive and use it
as if it was anything other than Puppy Linux, right? But that's all I have on the year of the
free BSD desktop. If you have any questions, you can leave a comment or send me an email or record
a follow-up show. If you have any complaints, you can do the same. If you have any grievances to
air, you can also do the same. If you... I'm trying to think of what else you might want to do.
If you want to learn more about free BSD, I will definitely say you should check out the handbook.
The handbook is good reading, even if you have no intentions of ever installing or running the
system, just because it gives you a really big overview on how to do things with this system,
how things are structured. If you're curious, but not committed. For the same reasons, I tell people
to read the 9 front FQA, even if they have no intentions of running the system, because it is
informative and it gives you a good overview of the system without actually committing to that system.
This episode, I've waited till the very end to say it. I'm recording this on free BSD. My USB
microphone just works. Unlike the 9 front episodes, I have a plan 9 episodes. You know,
where the microphone doesn't work. It just works on free BSD, you know. I'm recording in audacity,
so on and so forth. And I'm still looking. Oh, I think one last caveat, I should say this is
kind of an aside. So Linux has DRM enabled in the kernel. This means you can watch DRM controlled
content in your web browser. Free BSD does not have DRM controlled content in your web browser,
or in the kernel for that matter. So for example, you can watch Netflix on Linux. You can't watch
Netflix on free BSD. Even though the Netflix servers run on free BSD, but you know, the whole
debate on DRM and the kernel, my legitimate response is, why would you ever want to watch Netflix?
Right. There's so many things we can do besides just watching. We have so much doing to do in life.
Why not do it? And I think with that sort of multimedia caveat, the DRM caveat, I think I'm
going to end the show. So if you have any questions, feel free to leave a comment or send me an email.
If you get stuck somewhere along the way, trying to install free BSD,
for sure, for sure, send me an email or leave a comment and I will try my hardest to help you.
Right. That's what open sources all about. The people before me taught me for free. Now it is
my responsibility to teach the people after me. All right. We have to pass down this knowledge,
pass down these skills, help people where we can and help guide them to the answers and the
solutions. And I think that's my favorite aspect about open source software and free software in
general is sort of the, you know, passing of the torch before you become too cynical about software.
You have a period of time where you tell people all about all the things you've learned and help
people along the way. I think also if you do install free BSD, you should record an episode
and giving it a review. Your perspective as a person who has never used free BSD before,
what you like about it, what you don't like about it, the interesting things you found out about it,
maybe your favorite part about it, your least favorite part about it, why you're sticking with
free BSD, why you're going back to Linux, why you're considering being a hobbyist computer in any
capacity whatsoever when an operating system like macOS or Windows just works. All right. So if you
install free BSD, you know, record an episode giving a legitimate review on it. You know, we're so
curious about a lot of the new Linux users, but I'm more so curious about the people who have been
using Linux and Unix for a long time, trying a different type of Linux or Unix that's just
different enough, it might be a little bit uncomfortable. But as I always say, there's a lot of fun
in computing. You know, there's still a lot of fun in Linux and a lot of fun in BSD. So again,
thanks for listening to Hacker Public Radio. And stay tuned. Eventually, I will record the year of
the Open BSD desktop episode. This has been Hacker Public Radio, and that is all I have to say.
This has been RC, and I am ending the recording.
You have been listening to Hacker Public Radio. Hacker Public Radio does work. Today's show was
contributed by a HBR listener like yourself. If you ever thought of recording broadcast,
you click on our contribute link to find out how easy it leads. Hosting for HBR has been
kindly provided by an onsthost.com, the Internet Archive and R-Sync.net. On the Sadois status,
today's show is released on our Creative Commons Attribution 4.0 International License.