0:11

So friends and welcome back to your weekly

0:13

Linux talk show. My name is Chris. My

0:15

name is Wes and my name is Brent.

0:18

Well, hey gentlemen. Coming up on the show

0:20

today, we're going to really focus on the

0:22

Linux kernel again. There's been a lot of

0:24

hoopla about the state of rust and going

0:26

on there. So we're going to recap the

0:28

latest and then dive into what Linus and

0:30

Greg have said recently. We'll also chat with

0:32

Hannah from scale, who's going to give us

0:34

the four and one on what you need

0:37

to know on what you need to know to go to go

0:39

to go to go to go to. some picks and

0:41

a lot more. So before we

0:43

get to any of that, I

0:45

got to do the right thing,

0:47

and I got to say time-appropriate

0:50

greetings to that virtual lug. Hello,

0:52

mumble room! Hello, room! Hello, Chris.

0:54

Hello, Brian. Hello, Brian. Hello, that's

0:57

a serious showing. It's like it.

0:59

It's a little echo-y because we

1:01

had to get a huge room to

1:03

fit out. A lot of people.

1:05

Thank you everybody for joining us

1:08

over there. That's where you go

1:10

to get tail scale for free 100

1:12

devices, 3 users, no credit card required.

1:14

This is a modern networking solution

1:16

for connecting your devices securely to

1:18

each other, your applications, servers, systems,

1:21

whatever it might be, to each other

1:23

directly, on a mesh network, protected

1:25

by WAGO. It really is what

1:27

we've all wanted to see from

1:29

the moment we heard wire guard

1:31

was coming to the Linux kernel.

1:33

Secure remote access to your systems

1:35

that just works so intuitively intuitively

1:37

intuitively. and it's easy to deploy. It

1:39

actually is a zero config, no fuss, vp,

1:42

and I've been running it for years now. You

1:44

set it up and it just goes. You don't

1:46

really have to think about it. And with

1:48

the 100 device plan, you can use it

1:50

on just about everything you got and support

1:52

the show. So you go to tail

1:54

scale.com/ unplugged, you get it for

1:56

free on 100 devices, support the show, and

1:58

then you might do what... I did and I

2:01

inevitably rolled it out in the company. You

2:03

know, it's great for your back-end infrastructure

2:05

when you have multiple different

2:07

data centers and you want

2:09

to bring everything together. Thousands

2:11

of companies do this like

2:13

instacart, hugging face, dueling, more. They all

2:16

use tail scales so go try it out for

2:18

yourself or for your business. The free

2:20

plan is at tail scale.com slash,

2:22

unplugged. Well

2:24

as you listen to today's episode, we're

2:27

going to get into the Linux kernel

2:29

and we have a question you'd like

2:31

to answer just something put in the

2:33

back of your mind and when it

2:35

comes to you as we're talking about this

2:38

stuff, Bustin, if you could be

2:40

Linus Torvalds for a day, what would you

2:42

change or get done? Oh, that's a fun

2:44

one. You know, like the pet thing, you've

2:46

always hated in the kernel, the one tweak

2:49

you'd think or is something different?

2:51

You've been hearing us talk about

2:53

planet Knicks and that's coming up

2:55

really fast and that runs right

2:58

along scale. And so we wanted

3:00

to get you up to speed

3:02

on what you need to know

3:04

if you're going to attend scale

3:06

22X this year. And Hannah from

3:08

scale, she is the chair of

3:11

publicity, joins us to talk about

3:13

that. Hannah, welcome to the show. It's

3:15

great to talk to you. Yeah, thanks so

3:17

much. I'm so happy to be here to

3:19

talk about scale. We are really close. Scale

3:21

22X is coming up on March 6th. That's

3:23

less than two weeks now. Yeah, and so

3:26

we wanted to have you on to help

3:28

everybody kind of know what they need to

3:30

do in order to attend and what they

3:32

should do once they get there like on

3:34

day one. Yeah, of course, so I

3:36

would be remiss if I didn't

3:38

give you the quick spiel about

3:41

what scale is first before how

3:43

you can attend. Scale stands for

3:45

the Southern California Linux Expo. We're

3:47

in our 22nd year and we're

3:49

hosted in Pasadena, California. And so

3:51

we're so excited to be back.

3:53

And if you've never heard of

3:55

scale or have not been, it

3:58

is all things open source. We're

4:00

an entirely volunteer run conference. Scale is

4:02

run on love and passion for open

4:04

source. And so it is a great

4:06

time to get together and talk and

4:08

meet other people who care about all

4:10

things open source. And the other thing

4:13

that I absolutely love about scale is

4:15

we have been around for a long

4:17

time. We have great roots in Southern

4:19

California and beyond. And we also use

4:21

scale as a launching pad for other

4:23

open source communities who might not have

4:25

the infrastructure. or want to plan a

4:27

conference. So you might see some of

4:30

your other favorite open source events co-located

4:32

at scale. So Devops Day LA, Kauai

4:34

Summit, Planet Knicks, Linux training, so we

4:36

have a little bit of everything if

4:38

you are in the open source world,

4:40

which I imagine everyone listening here is.

4:42

So now that we've got kind of

4:45

the details out of the way, if

4:47

you want to come to scale, we

4:49

would love to have you in a

4:51

little under two weeks, you can go

4:53

to our website. website so Cal Linux

4:55

expo.org and get registered. We really pride

4:57

ourselves on the conference being approachable and

5:00

I think we're probably one of the

5:02

only conferences that run a four-day event

5:04

for under $100 for a ticket. So

5:06

we think it's really good value but

5:08

if you want a little extra motivation

5:10

to get your ticket before the event

5:12

you can use the promo code Linux

5:15

for 50% off that ticket. Great, okay,

5:17

promo code Linux, and so, also I

5:19

just want to say I do love

5:21

these side events that have been happening

5:23

over the last few years. It's such

5:25

a brilliant idea because, you know, you

5:27

folks really are the experts now at

5:29

the infrastructure and the event, and that

5:32

is such a massive undertaking for smaller

5:34

projects or groups or communities to undertake.

5:36

It's a really incredible thing you're doing.

5:38

Yeah, we're quite proud of it. Event

5:40

planning is no joke and we have

5:42

hundreds of volunteers that make this conference

5:44

run year-round and it would not be

5:47

possible without volunteers. So another shameless plug

5:49

is if you want to volunteer and

5:51

help with the conference you can always

5:53

there's an email on the website where

5:55

you can volunteer either the week of

5:57

or if you want to work on

5:59

our network or marketing shameless plug for

6:02

my team we're always taking volunteers to

6:04

do it but Yeah, we love having

6:06

other events there that would not have

6:08

the resources otherwise. So there's a place

6:10

for everyone who cares about open source

6:12

here and it's in Pasadena in two

6:14

weeks. That's great. I feel like too

6:17

getting involved like that would really build

6:19

some skills that you know could be

6:21

marketable in the workplace there. I just

6:23

kind of want to make it clear

6:25

when people show up they do need

6:27

to proceed to it like there's a.

6:29

proper registration process and all of that

6:32

that they need to follow when they

6:34

first get there or at least get

6:36

their badges and whatnot right? Yep so

6:38

we're we're a pretty standard conference you

6:40

can register for your tickets ahead of

6:42

time at so Cal Linux expo.org with

6:44

that promo code Linux and you're all

6:46

registered you will go to the building

6:49

that says welcome there's a big Linux

6:51

penguin that'll give you an idea of

6:53

what buildings we're in. The Pasadena Convention

6:55

Center is a little confusing. It is

6:57

two buildings with a theater in between.

6:59

So if you're facing the convention center,

7:01

the building on the right is where

7:04

you check in your first day. And

7:06

we've got self-serve systems where you enter

7:08

your name and your registration number and

7:10

it'll print your badge and then you

7:12

get a... a swag bag like most

7:14

conferences full of open source swag. And

7:16

then you get to go in and

7:19

see all the topics for the four

7:21

days. So make sure you go to

7:23

the right side building on your first

7:25

day, get your badge and once you

7:27

have your badge, you can go to

7:29

all the events throughout the four days.

7:31

Yep. And then it's easy after that.

7:34

That's the hardest part is just figuring

7:36

out which building to go in and

7:38

get your badge. It's not too bad.

7:40

I would say the hardest part is

7:42

picking which talks to go to. There's

7:44

so many good ones that happen all

7:46

at once. And the other good thing

7:48

about our conference is don't get FOMO

7:51

because if you have two concurrent talks

7:53

we stream and record all of them

7:55

so you can go back and listen.

7:57

are the ones you missed as well.

7:59

Also no small feet which we really

8:01

appreciate too. Yeah the tech on that

8:03

is insane. It takes dozens and dozens

8:06

of people all here to get those

8:08

recordings up but we're quite proud of

8:10

them. You can always check out our

8:12

past talks from events on our YouTube

8:14

channel as well. So if you're still

8:16

on the fence you don't know what

8:18

kind of content. Check out the YouTube

8:21

channel from the years past. We've got

8:23

a lot of great content there and

8:25

it gives you an idea of what

8:27

we're putting on for the week as

8:29

well. for sure i will say it

8:31

never quite captures the social element in

8:33

the hallway track and you know going

8:36

out and having lunch and all that

8:38

stuff but that's just the bonus that

8:40

you get to come discover on your

8:42

own when you visit so sounds like

8:44

they need to get registered they can

8:46

use promo code Linux to save some

8:48

money anything else we need to let

8:51

folks know Yeah, just like you said,

8:53

like the hallway track is really the

8:55

highest value ad. So if it's your

8:57

first time coming to scale, introduce yourself.

8:59

I always hang out at registration. So

9:01

you're always welcome to say hi to

9:03

me. So you already have one friend.

9:05

But get to know other folks. We

9:08

love the cross population of communities and

9:10

it comes to really amazing things outside

9:12

the conference. So I would really encourage

9:14

you not to be shy and make

9:16

some new friends there as well. Absolutely.

9:18

Yeah, thank you so much for having

9:20

me and I hope to see everyone

9:23

listening at scale in March. Wonderful. Well

9:25

I know the world is full of

9:27

news these days but jeez the colonel

9:29

has been seeing some newsworthy notes from

9:31

day to day and you guys have

9:33

been doing a deep dive. My favorite

9:35

thing. Should we go through what's been

9:38

happening and dust them off a little

9:40

bit? Yeah, we started capturing this for

9:42

the members in the members bootleg and

9:44

then it really has developed further. So

9:46

let's go back like Brent says and

9:48

kind of and kind of just briefly

9:50

just cover what's happened here. Obviously we've

9:53

talked about rust and Linux kernel a

9:55

bit. We last September we did an

9:57

episode about it and that was sort

9:59

of the. recap on the state of

10:01

things. Right. An ongoing effort to, you

10:03

know, add the ability to write new

10:05

code to add drivers to the Linux

10:08

kernel using Rust, which is a secondary

10:10

language in the kernel and it's a

10:12

big change and that's why it's been

10:15

a slow and long effort and why

10:17

we're continuing to talk about it. Even

10:19

the idea of adding another language is

10:21

a big deal, right? And of course,

10:24

you know, you have to say, Rust

10:26

is... It's not new, but it's still

10:28

moving fairly fast and does things, it's

10:30

a very different language than C. So

10:33

you've got that to contend with. So

10:35

with our current event, we actually need

10:37

to go back to late January of

10:40

2025, and the Rust-DMA patch proposal sparks

10:42

a bit of a conflict on the

10:44

Linux kernel mainly list. A patch is

10:46

proposed to enable Rust-written device drivers to

10:49

call the Linux kernel's core direct memory

10:51

access, or what is referred to as

10:53

DMA. Obvious drivers need this, right. The

10:55

goal here to expand Russ's usability within

10:58

the kernel. Christopher Hillwig, though, had some,

11:00

I would say, concerns and probably raised

11:02

the largest rejection. There's a quote from

11:04

him that says, no rust code in

11:07

kernel slash DMA, please. Yeah, there's a

11:09

lot to digest here. For the first

11:11

part, you know, fairly simple patch, not

11:14

a huge amount. Three files changed, 273

11:16

insertions, and the text, the request reads.

11:18

Add a simple DMA coherent allocator rust

11:20

abstraction. And then abstraction has a specific

11:23

meaning here. There's sort of the automatic

11:25

bindings that get generated to be able

11:27

to like talk to the C data

11:29

structures from the rust side of things.

11:32

But then there's the like the higher

11:34

layer part of really using rust and

11:36

rust type system in what they're calling

11:39

abstractions, which is where you do the

11:41

work to wrap the C side at

11:43

a semantic level to sort of encode

11:45

how to safely use where possible where

11:48

possible. the seaside from rust. So then

11:50

you have this abstraction layer that can

11:52

kind of sit between. And it's actually,

11:54

for the most part, for non-exceptional cases.

11:57

you're not even allowed to go call

11:59

the C directly in drivers. You're not

12:01

supposed to, you want to use this

12:03

abstraction. So that's where this code is

12:06

trying to sort of bridge the gap

12:08

to enable downstream rust things to be

12:10

able to use the bus, the DMA

12:13

bus. And that's the idea is that

12:15

downstream rust things could read the DMA

12:17

bus. And the very events that transpired

12:19

this week come back to this very

12:22

patch and this very discussion. So this

12:24

event that happens in January. is pretty

12:26

noteworthy as time goes on, the discussion

12:28

kind of escalates. You'll see this, there's

12:31

several people in there that are kind

12:33

of anti-rest, there's folks in there trying

12:35

to explain stuff, there's other folks that

12:38

seem to be just kind of in

12:40

a watch and see mode. Maybe it's

12:42

worth also touching on, as you said,

12:44

Helwig said, no rust code in kernel

12:47

slash DMA, and what that's referring to

12:49

is sort of the various trees inside

12:51

the kernel source tree itself. So that's,

12:53

we'll see coming back to this already,

12:56

if you just look at the diff,

12:58

you know, how you view whether it's

13:00

a part of the subsystem or not

13:02

versus like where it lives in the

13:05

source tree could be a whole separate

13:07

question, but this was all under rust

13:09

slash at the top layer. So it's

13:12

all kind of in the rust subtree

13:14

in its own. It is, of course,

13:16

wrapping code that lives in the DMA

13:18

side, but. Yes. And so it's in

13:21

its own contained area that is important

13:23

to understand later. So this kind of

13:25

bruise for a bit, you know, there's

13:27

Fosdom, there's a talk about Rustford Linux

13:30

there, held by Miguel Odeja? Is that?

13:32

Ojeda? Ojeda? And he's the lean maintainer

13:34

of Rustford Linux, he presented at Fosstom.

13:37

He highlighted there some of the progress

13:39

and some of the things that landed

13:41

in Linux 613. He's able to get

13:43

more responses from various maintainers, both, you

13:46

know, neutral for and against. I think

13:48

it seemed to have. He's able to

13:50

get more responses from. people interested in

13:52

the project, but it was an interesting

13:55

survey. As you can imagine though after

13:57

that talk the debate kind of heats

13:59

up again until later on, Linus and

14:01

Hector Martin kind of start to get

14:04

into it. And Hector Martin takes the

14:06

social media to call out the process

14:08

as being, you know, just really hard

14:11

on the contributors who are trying to

14:13

get rust in the Linux kernel. That

14:15

very act sparks a debate on the

14:17

Linux mailing list about social brigating and

14:20

trying to influence the kernel process through

14:22

social media posts. Linus steps in says

14:24

this is essentially social says it pushes

14:26

him away from supporting the patch, says

14:29

to Hector maybe the problem is you.

14:31

And that got quite a bit of

14:33

attention just a couple of weeks ago.

14:36

To kind of try to take us

14:38

from the dust up over that to

14:40

actual like tangible policy, Miguel introduced what

14:42

he's calling on February 11th, the Rust

14:45

Colonel policy, which is meant to clarify

14:47

how Rush should integrate with the Colonel.

14:49

hoping to reduce tensions. Yeah, he wrote,

14:51

given the discussions in the last days,

14:54

I decided to publish this page with

14:56

what our understanding is. Hope it helps

14:58

to clarify things. Did it clarify things,

15:00

Wes? Well, that probably depends on who

15:03

you ask. There was obviously some pushback.

15:05

But maybe it's worth looking at kind

15:07

of like what it was. Yeah. You

15:10

know, it had some points about. Like

15:12

who's pushing this? It's not the Rust

15:14

project or the foundation or even necessarily.

15:16

There are companies interested, but it's like

15:19

its own, you know, their kernel contributors

15:21

who want to add rust. Then I

15:23

mentioned some key folks, maintainers, other people

15:25

that are involved. Here's some quotes. Some

15:28

subsystems may decide they do not want

15:30

to have rust code for the time

15:32

being, typically for bandwidth reasons. This is

15:35

fine and expected. Now, in the Colonel

15:37

Maintainer Summit 2022, we asked for flexibility

15:39

when the time comes that a major

15:41

user of rust in the kernel requires

15:44

key APIs for which the maintainer may

15:46

not be able to maintain rust abstractions

15:48

for it. This flexibility is the needed

15:50

counterpart to the ability of maintainers to

15:53

decide whether they want to allow rust

15:55

or not. It also says some stuff

15:57

about what happens if things break, so

15:59

on their FAQ, who's responsible if a

16:02

sea change breaks a build with rust

16:04

enabled? That's has been a big contentious

16:06

question, right? Well this site's position their

16:09

understanding is the usual kernel policy applies.

16:11

So by default changes should not be

16:13

introduced if they are known to break

16:15

the build, including rust. However, exceptionally for

16:18

rust, a subsystem may allow to temporarily

16:20

break rust code. The intention is to

16:22

facilitate friendly adoption of rust in a

16:24

subsystem without introducing a burden to existing

16:27

maintainers who may be working on urgent

16:29

fixes for the seaside. That feels like

16:31

a big compromise from the rust side.

16:34

where they're essentially saying we're willing to

16:36

let rust code break for a while

16:38

and we'll try to sort it out

16:40

if it just means the seaside will

16:43

roll with us here. That's a big

16:45

give. And you've seen various versions too,

16:47

right? Like that's where there's, you know,

16:49

some subsystem maintainers are interested in figuring

16:52

out some of the rust stuff for

16:54

themselves. Other ones appoint like a co-maintainer

16:56

or like a, you're maintaining the rust

16:58

side of this subsystem, well I'm going

17:01

to do the seaside. as some people

17:03

felt was kind of an agreement before,

17:05

was to accommodate all of that. And

17:08

the idea would be if the rust

17:10

code does break, you actually fix it

17:12

before it gets to Linus' tree, and

17:14

it actually ships. It's not like you

17:17

ship broken rust code. There's an idea

17:19

that gets caught, but that is still

17:21

a pretty big compromise. Now. There are

17:23

concerns with some of this in practice

17:26

particularly around like okay you say it's

17:28

fine to break it Yeah, am I

17:30

gonna get am I gonna get yelled

17:32

at still and who knows it's even

17:35

broken? You have to make sure that

17:37

you've set up and I don't think

17:39

they're not doing this but just you

17:42

have to make sure you've set up

17:44

like who do you then if we've

17:46

all agreed you're not going to fix

17:48

it but it being broken is a

17:51

problem then like who is actually in

17:53

practice fixing it? Yeah. also said, I

17:55

don't think having a web page in

17:57

any form is useful. If you want

18:00

it to be valid, it has to

18:02

be in the kernel tree and widely

18:04

agreed on. It also states factually incorrect

18:07

information, e.g. some subsystems may decide they

18:09

do not want to have rust code

18:11

for the time being, we read that,

18:13

while Linus in private said that he

18:16

absolutely is going to merge rust code

18:18

over a maintainer subjection. Yeah. And this

18:20

has been sort of the, one of

18:22

the complaints has been mixed messaging here.

18:25

Like, oh, it's a maintainer's choice, but

18:27

also I am going to be publishing

18:29

rest code. Helwing's been kind of emerging

18:31

as maybe the most vocal critic of

18:34

rust getting in and I think as

18:36

probably raised the most. points in the

18:38

most recent rounds of going back and

18:41

forth. And he's kind of calling out

18:43

Linus here. He says, quote, yeah, like

18:45

you said, well, Linus in private said

18:47

that he absolutely is going to merge

18:50

Russ Code over a maintainer's objection. He

18:52

says this is a very dishonest way

18:54

of communication. Yeah, and that was pointed

18:56

at Miguel over including that quote about

18:59

you don't have to take it. Right.

19:01

And that's where there's a bunch of

19:03

nuance and will will eventually see spoilers

19:06

that Linus sort of... does eventually chime

19:08

in, but I think there has been

19:10

a question in many folks' minds of

19:12

like, where, you know, has there been

19:15

enough communication? We've seen Greg H. chime

19:17

in a few times in various small

19:19

ways, especially to clarify stuff about, oh,

19:21

what was broken in this previous case

19:24

that people were pointing to at one

19:26

time. So, you know, while this is

19:28

going on, right? Burned out. Disappointed with

19:30

leadership in the kernel frustrated with their

19:33

post approach to rest integration cease is

19:35

maintaining upstream Linux kernel code for the

19:37

arm stuff You know the Solly project

19:40

has to rework that now We've seen

19:42

others that burned out like that that

19:44

is so you understand the context and

19:46

the background while all of this is

19:49

sort of playing out. While also at

19:51

the same time publicly Greg and Linus

19:53

kind of would from time to time

19:55

say fairly pro rust stuff even though

19:58

this superheated debates happening in the background

20:00

during all of this you'd see comments

20:02

from both of them saying oh yeah

20:05

no rust is coming. Yeah and I

20:07

think there has been maybe some lack

20:09

of clarity in general or at least

20:11

there's been some confusion around the like

20:14

okay so it's clear that rust has

20:16

been kind of an experiment in the

20:18

kernel but... To some people's mind experiment

20:20

means it's not been like we're not

20:23

doing it. Whereas I think maybe in

20:25

some of the higher up maintainers minds

20:27

experiment means well we don't know how

20:29

we're doing it and we're figuring out

20:32

the right way to do it and

20:34

it may be painful and bright things

20:36

but like in general we're doing it.

20:39

We may decide we ultimately don't want

20:41

to do it and stop doing it

20:43

but for the moment we're doing it.

20:45

We should not just introduce rust. It's

20:48

going to make it hard to maintain.

20:50

This is going to make it more

20:52

of a bear. We should just improve

20:54

the C stuff. If we want, we

20:57

should go back and retrofit that. And

20:59

I think what they're ultimately pushing back

21:01

on is, well, where do you draw

21:04

the line? Do you completely rewrite the

21:06

Linux kernel one day in rust? Is

21:08

that where this leads to? Every time

21:10

we find a bug now, are we

21:13

just going to write a rust version?

21:15

Like that's, they don't want that, that's

21:17

ridiculous. We should just fix the problems

21:19

and see is what they're, I'm sure

21:22

you've seen that opinion in the mailing

21:24

list. Yeah, well, I think it's also

21:26

like, you know, it means you have

21:28

to have double the infrastructure, you're now,

21:31

at some point, maybe become dependent on

21:33

the right version of cargo. And you

21:35

can just imagine an important sea project,

21:38

even if you can. even if you're

21:40

willing to take like a co-maintainer path,

21:42

it's a lot of trust, it's a

21:44

lot of change, and you are going

21:47

to have code that's either making, you

21:49

know, has... expectations of you or is

21:51

directly interfacing with code that you're using

21:53

that you don't maybe have a full

21:56

understanding of and imagine a situation where

21:58

you're trying to push a feature and

22:00

It does break something in rust and

22:03

now you're kind of trying to be

22:05

polite So you're trying to wait for

22:07

the rust stuff to get fixed, but

22:09

now maybe you've missed the merge window

22:12

like you can understand there There is

22:14

a little bit of concern there that

22:16

does seem legitimate and that's where this

22:18

is you know a complicated question because

22:21

it is ultimately one, like most things

22:23

in the real world of tradeoffs, right?

22:25

Like there are real costs to having

22:27

multiple languages regardless of the language, and

22:30

there's real costs for the rust in

22:32

particulars, and it's a big change. But

22:34

what the rust for Linux folks, and

22:37

as we'll see some others, are advocating,

22:39

is that especially maybe if we focus

22:41

this on net new code and, you

22:43

know, new drivers in particular, there's a

22:46

lot of upside and a lot of

22:48

wins that can help the kernel overall.

22:53

One password.com/unplugged that's the number one password.com/unplugged

22:55

all lower case and you're gonna want

22:57

to go there because this is something

22:59

that if I had when I still

23:02

worked in IT I think it would

23:04

have sustained me for many many more

23:06

years the reality is your end users

23:09

don't and I mean without exception work

23:11

on only company owned devices applications and

23:13

services maybe get lucky and they mostly

23:16

do but I don't find that to

23:18

be the reality today And so the

23:20

next question becomes, how do you actually

23:23

keep your company's data safe when it's

23:25

sitting on all of these unmanaged apps

23:27

and devices? Well, that's where one password

23:29

has an answer. It's extended access management.

23:32

One password, extended access management helps you

23:34

secure every sign-on for every app on

23:36

every device, because it solves the problems

23:39

that your traditional IAMs and MDM's just

23:41

can't touch. And it's the first security

23:43

solution that brings unman managed devices and...

23:46

applications and identities under your control. It

23:48

ensures that every credential is strong and

23:50

protected. Every device is known and healthy.

23:52

Every app is visible. This is some

23:55

powerful stuff, and it's available for companies

23:57

that use Octa or Microsoft Entra, and

23:59

it's in beta for Google Workspace customers,

24:02

too. One password changed the game for

24:04

password management, and now they're taking everything

24:06

they've learned there and expanding it to

24:09

the login level and the application level.

24:11

You know what a difference it makes

24:13

when people have proper password management. Now

24:16

let's have proper login management and authorization.

24:18

One password also has regular third-party audits

24:20

and the industry's largest bug bounty. They

24:22

exceed the standards set by others. They

24:25

really do. So go secure every app,

24:27

every device, and every identity. Even the

24:29

unmanaged ones. You just go to one

24:32

password.com slash, unplugged. Also, Linus did chime

24:34

in just a few days ago. Greg

24:36

came in with a rather compelling case,

24:39

actually, for new drivers, at least, to

24:41

be written in rough, like West was

24:43

saying. He says, quote, yes, mixed language

24:46

code bases are rough and hard to

24:48

maintain, but we're kernel developers, damn it.

24:50

We have been maintaining and strengthening Linux

24:52

for longer than anyone ever thought was

24:55

going to be possible. We have turned

24:57

our development model into a well-oiled engineering

24:59

marvel creating something that no one else

25:02

has ever been able to accomplish. Adding

25:04

another language really shouldn't be a problem.

25:06

We've handled much worse things in the

25:09

past and we shouldn't give up now

25:11

on wanting to ensure that our project

25:13

succeeds for the next 20 plus years.

25:16

We've got to keep pushing forward when

25:18

confronted with a new good ideas, and

25:20

embrace the people offering to join us,

25:22

actually doing the work to help make

25:25

sure that we all succeed together. Yeah,

25:27

I like that. I really think Greg's

25:29

done some good, you know, just sort

25:32

of on the edges and a little

25:34

stronger here, but just with a very

25:36

nice tone. It's so positive. It makes

25:39

you want to go help make Linux

25:41

better yourself. And I think it's right

25:43

to... call out that, you know, whether

25:46

you agree with the mission or not,

25:48

you have to, you've got to agree

25:50

that the Russ Relinics folks are trying

25:52

hard. You know, I noticed, I think

25:55

it was a week or two ago,

25:57

and they're working on the FOBUS, and

25:59

Greg posted on Masto about it, and

26:02

he made sure to call out, you

26:04

know, from the get-go, this has rust

26:06

API bindings, it's ready to go for

26:09

the rust crew, we're introducing a seaside

26:11

and a seaside and a rust side,

26:13

and a rust side, and it, and

26:16

it, and it, and it's good, and

26:18

it's good, and it's good, and it's

26:20

good, and it's good, try out, try

26:22

out, try out, try out, try out,

26:25

try out, try out, try out, try

26:27

out, try out, influenced not some small

26:29

amount by you know he's been super

26:32

involved with the you know maintaining the

26:34

stable side of things and dealing now

26:36

with the kernel Cvee process so yes

26:39

he started that post by saying yeah

26:41

as someone's who's seen almost every kernel

26:43

bug fix and security issue for the

26:46

past 15 plus years well hopefully all

26:48

of them and who sees every kernel

26:50

Cvee issued I think I can speak

26:52

on this topic the majority of bugs

26:55

and here he's trying to be clear

26:57

quantity not quality or The majority of

26:59

bugs we have are due to the

27:02

stupid little corner cases in C that

27:04

are totally gone in rust. Things like

27:06

simple overrides of memory, error path cleanups,

27:09

forgetting to check error values, and use

27:11

after free mistakes. That's why I'm wanting

27:13

to see rust get into the kernel.

27:16

These types of issues just go away,

27:18

allowing developers and maintainers more time to

27:20

focus on the real bugs that happen.

27:22

I. e. logic issues, race conditions, etc.

27:25

etc. Yeah, the memory problems developers grate

27:27

themselves. And then, so a moment that

27:29

really impressed me was Linus. And you

27:32

know, I've been critical of him recently

27:34

on some of his decisions, but I

27:36

really felt like Linus showed he still

27:39

really got his hands around this entire

27:41

thing. He popped in on the mailing

27:43

list a couple of days ago. And

27:46

remember, this went all the way back

27:48

to January, the beginning of the nuance

27:50

of all of this. And none of

27:52

it missed Linus' catch and I was

27:55

impressed by that. And he was responding

27:57

to Hillwick I believe. writes, you're not

27:59

forced to take any rust code at

28:02

all, or care about rust code in

28:04

the DMA code. You can just ignore

28:06

it, but the quote, ignore the rust

28:09

side, end quote, automatically also means that

28:11

you don't have any say on the

28:13

rust side. You can't have it both

28:16

ways. You can't say quote, I want

28:18

to have nothing to do with rust,

28:20

and then in the very next sentence

28:22

say quote, and that means that the

28:25

rust code that I want to ignore

28:27

cannot use the C interfaces I maintain.

28:29

can be involved in it, and by

28:32

being involved with it, they will have

28:34

some say in what the rust bindings

28:36

will look like. They basically become the

28:39

maintainers of the rust interfaces. And what

28:41

Linus did here, in my kind of

28:43

short, horrible read there, is he recognized

28:45

that what Hillwig was originally complaining about

28:48

wasn't really applicable here. The Rust folks

28:50

weren't asking to change anything on the

28:52

seaside. They weren't touching anything on the

28:55

seaside for the DMA bus that we

28:57

originally talked about. As we touched on,

28:59

it was in a separate subtree. It's

29:02

in its own separate subtree. And all

29:04

they would be doing is reading the

29:06

DMA bus like anything else in the

29:09

Linux kernel. And that is a line

29:11

too far for Linus. You can't be

29:13

not involved and also say what these

29:15

programs can and cannot use in the

29:18

Linux kernel. As he puts it. But

29:20

that wall of protection goes both ways.

29:22

If you don't want to deal with

29:25

the rust code, you get no say

29:27

on the rust code. Put another way,

29:29

the nobody is forced to deal with

29:32

rust does not imply everybody is allowed

29:34

to veto any rust code. I think

29:36

I just want to reiterate, you know,

29:39

people have been saying, Linus needs to

29:41

say more, Linus needs to chime in,

29:43

Linus, if you look at this mailing

29:45

list threat, he's chiming in pretty frequently

29:48

throughout. It does seem like, I mean,

29:50

it seems maybe it failed, but he

29:52

did clearly also have some private communiques

29:55

with Miguel. But I was impressed that

29:57

the original nuance all the way back

29:59

in. January did not miss his grasp

30:01

and that was a fundamental nuance

30:04

to kind of saying to Hillwig hey

30:06

I respect your arguments you're often

30:08

right you call in a line even says

30:10

you call me on my BS I'm calling

30:12

you on your BS here that's the other

30:14

thing if you just hear the parts we

30:16

read the most intense parts of the you

30:18

know the debate and Linus putting down his

30:21

policy right but he also says things like

30:23

you know you're saying like I respect what

30:25

you're saying I respect you technically I respect

30:27

you technically and I like working with

30:29

you and I like it when you call me

30:31

on my BS. I say some stupid

30:33

things at times. There needs to

30:35

be people who can stand up to

30:38

me and tell me I'm full of S.

30:40

But now I'm calling you out on yours.

30:42

It's pretty good. It's a

30:44

very different lineness than we've

30:46

seen previously. I love it.

30:48

I think one of the MVPs in this

30:51

threat is McGill, don't you think?

30:53

Yeah. I will say there's been, you

30:55

know, we focused on a lot of

30:57

the more intense... parts as I was

30:59

just saying. But there was a lot

31:01

of good interactions in there as well. You

31:03

know, I saw James bottomly asking a lot of

31:05

like legit and good nature questions, trying,

31:08

a lot of developers trying to learn

31:10

more. Kent Overstreet chiming in how this

31:12

could be beneficial for B cacheFS or

31:14

how, you know, how to maybe reframe

31:16

things from time to time. Yeah, and

31:18

you know, there's just, there's a lot

31:20

of different levels of different levels

31:22

of knowledge, right. Some folks have used...

31:24

Rust but like for Usurland stuff and

31:26

never thought about the kernel side of

31:28

it some people have never touched rust

31:30

at all so you get levels of like do

31:33

you is your understanding of rust accurate

31:35

are you are you asking about a

31:37

nuance of like where the type system

31:39

isn't complete are you asking sort of

31:41

like you don't understand you know how the compiler

31:43

side of it works there's a lot of

31:45

areas and yeah Miguel has just sort

31:47

of been seemingly tirelessly and for

31:49

the most part quite you know politely

31:52

answering trying to clarify This

31:54

has not been an easy process

31:56

for the community and it has

31:58

not been without its costs. really

32:00

good contributors, you know, and

32:02

I know that it's it's cliche to

32:04

say the process is messy, but

32:06

sometimes, I mean, you actually see

32:09

the mess unfolded. It is, it's

32:11

hard to watch. Neil, did you have

32:13

any thoughts from sort of the

32:15

Asahi side, you know, watching that

32:17

project as closely as you do now

32:20

being tied in over there for door?

32:22

Did you have any thoughts on this?

32:24

Not because he hasn't stepped

32:26

in and doesn't have good

32:28

thoughts about this, but like

32:30

the downward spiral with this

32:33

whole, the whole conversation that's

32:35

had, could have been completely

32:38

avoided if he had said this

32:40

earlier. I get why he didn't.

32:42

I understand the desire to

32:44

have the community sort itself

32:46

out. I just think that this

32:49

is one of those things that

32:51

it is going to require. more tough

32:53

love type stuff as we keep going

32:55

because like there's this contention with groups

32:57

of people saying oh this is not

32:59

this can't ever be enterprise ready and

33:02

so it's a toy there are people

33:04

saying I don't want to deal with

33:06

this because multi language is terrible ignoring

33:08

the fact that the Linux kernel actually

33:10

been a multi language project since almost

33:13

the beginning it's been see an assembler

33:15

and then see pearl python in assembler

33:17

and make files and shell and and

33:19

like There is code in there that generates

33:22

code that generates code that is then

33:24

used to compile code. Like it is

33:26

a morass of complexity all

33:28

the way through it. Not acknowledging

33:30

that and then kind of complaining

33:32

about rust bringing in a new

33:34

thing of multi-languageness and complexity I

33:36

think kind of is missing the

33:38

forest for the trees there. There's

33:40

a lot more complexity in the

33:42

Linux kernel today than there ever

33:44

has been. Basically I think a lot

33:47

of people are going to need reality

33:49

checks. fairly frequently as this keeps going

33:51

on. Now I'm not going to say that

33:53

I'm like the most amazing fan of rust.

33:55

I think that there are things about rust

33:58

that I'm not the biggest fan of. But

34:00

it is pretty clear at this

34:02

point that there are significant

34:04

drivers towards, pun intended,

34:07

towards rust in the

34:09

Linux kernel and people who

34:11

are interested and enthusiastic about

34:14

it. And I think the

34:16

job of a good community

34:18

leader is to encourage and

34:20

support the people that are

34:22

doing these sorts of things

34:24

to enable the next generation

34:26

to be successful. Like, the

34:28

Linux kernel has a very,

34:30

very big problem that it's, that

34:32

it can't, that it hasn't really

34:35

started looking at addressing, which is,

34:37

the churn rate of maintainers in

34:39

the Linux kernel is unbelievably low.

34:41

Most of the people that have

34:44

maintaining subsystems, the Linux kernel, have

34:46

been doing it since almost the

34:48

beginning of that subsystem. And you

34:50

look at the progress of bringing in

34:52

new people to stay in for the

34:55

long haul, and it's not terribly high.

34:57

And that can be concerning as everyone's

34:59

getting older and what's going to

35:01

happen when a critical subsystem loses

35:03

a maintainer for reasons. Like we

35:05

lost wireless subsystem Larry Finger like

35:07

last year. And this is not

35:09

going to be a new thing.

35:12

This is going to keep happening.

35:14

And so like we have to

35:16

start, I mean, we, I, the

35:18

people in the Linux internal

35:20

community, the subsystem leaders

35:22

and whatnot, have to

35:24

start acknowledging that. There has

35:26

to be some give and take. There has to

35:29

be some give to be where the people are

35:31

or where the people want to be.

35:33

And some people want C+, some people

35:35

want Russ, some people want to be

35:37

on forges rather than using mailing lists.

35:40

This is all going to be things

35:42

that have to be dealt with because

35:44

you need the next generation or

35:46

the project dies. So this is

35:48

something where I think Linus is

35:50

going to have to take some

35:52

tougher leadership on because otherwise I... I

35:54

don't know what's going to happen.

35:57

I thought maybe we could end, I saw

35:59

a nice post. during part of this

36:01

discussion from Keys Cook, who's for

36:03

a long time leading the kernel

36:05

self-protection project and doing a lot of good

36:07

security, vulnerability-focused work on the

36:09

C code in the kernel. And Chris,

36:11

you mentioned kind of the contention like, well,

36:13

where does this end? And are we trying

36:16

to rewrite the whole kernel? And what is

36:18

the goal of this experiment, really? So,

36:20

Keys wrote, speaking to the what is the goal

36:22

question, I see the goal is eliminating

36:24

memory safety issues and new drivers and

36:26

subsystems. The pattern we've seen in

36:28

Linux. With security flaws is that

36:30

the majority appear in new code.

36:33

Focusing on getting new code written in

36:35

Rust puts a stop to these kinds

36:37

of flaws. And it has an exponential

36:40

impact, as Android and Usenix have found,

36:42

and then he included a bunch of

36:44

citations. In other words, I don't

36:47

see any reason to focus on

36:49

replacing existing code. Doing so would actually

36:51

carry a lot of risk. But writing

36:53

new stuff in Rust is very effective.

36:55

Old code is more stable, and

36:57

it... already has fewer bugs, and

36:59

yet we're still going to continue

37:02

the work of hardening C,

37:04

because we still need to shake

37:06

those bugs out. But new code can

37:08

be written in rust and not have

37:10

any of these classes of bugs

37:13

at all from day

37:15

one. Jupiter broadcasting.com slash

37:17

River. You know I'm all

37:19

about decentralized systems like Linux,

37:22

podcasting, and Bitcoin.

37:24

Self-hosted, user-controlled, no vendors

37:26

pulling the strings. And Bitcoin

37:28

really is the Linux of money. When we

37:31

first started talking about booths

37:33

on the show, Bitcoin was around

37:35

30-something thousand dollars, and now it's

37:37

around a hundred thousand something dollars?

37:39

Now why did that happen? It's because

37:41

of a couple of things, and

37:43

scarcity and network effect come to

37:45

mind. There's only 21 million Bitcoin

37:47

that will ever exist, and there

37:50

is unlimited printed fiat chasing

37:52

it. This is why the Bitcoin ETFs

37:54

were the most successful ETF launches in

37:56

history. There was just a little story

37:58

that went by in the last year. But they

38:00

were, the most successful ETFs in

38:03

history. So, if you're like me,

38:05

you don't really want an ETF

38:07

because that's essentially an IOU.

38:09

You want to own the real thing.

38:11

The challenge is finding a legit,

38:13

secure, trusted space that isn't

38:15

a scam or isn't something that is going

38:18

to go away in a few years.

38:20

You want a place you can buy,

38:22

DCA, or sell your Bitcoin. And my

38:24

answer to that is River. Go to

38:26

Jupiter broadcasting.com/river that'll take you to

38:29

our partner page. River is where

38:31

I buy my Bitcoin, my family's

38:34

Bitcoin. What I like about them is

38:36

they have proof of reserves and anyone

38:38

can audit their supply at any time.

38:40

They also are a Bitcoin only focused

38:42

company. So that means top to

38:45

your services for Bitcoiners, slick mobile

38:47

and web apps, lightning support, the

38:49

works. And one of my favorite

38:52

features is actually their cash savings

38:54

account with a 3.8% interest paid

38:56

in Bitcoin. Now your Galaxy

38:58

Brain move here is you stash some

39:00

cash in there, you earn the 3.8% and

39:02

then you smash by Bitcoin when the price

39:05

dips a little bit. And trust me,

39:07

it adds up fast. River is the

39:09

most trusted spot in the United

39:11

States for individuals and businesses

39:13

to buy, sell, send, or

39:16

receive Bitcoin. That's why I use

39:18

them. and that's why I am pumped to

39:20

recommend them to you. So if you just want

39:22

some Sats to boost the show or you're looking

39:24

for a long-term store value, River

39:26

makes it easy to get started with

39:29

Bitcoin in three simple steps. Get

39:31

started by going to

39:33

Jupiter broadcasting.com slash River. It's

39:35

real easy that'll take you

39:37

to our partner page for

39:39

businesses or individuals. Jupiter broadcasting.com

39:42

slash River. Well,

39:45

it is certainly conference season coming

39:47

up and we have one who's

39:49

planning way ahead Texas Linux Fest

39:51

is of course back again this

39:53

year But with a new season

39:55

October 3rd to 4th this year.

39:57

That's in Austin, Texas and there's

39:59

a call for paper open. You

40:01

have until August 3rd, so no

40:03

excuses for getting a talk in,

40:05

but we would love to see

40:07

what you can throw at Texas

40:09

Linux Fest. Make it good. That

40:11

way, we feel like we have

40:13

to go. Yeah, yeah, it's a

40:15

great little fest. They're looking for

40:18

sponsors to 2025. Texas

40:20

Linux fest.org for more details

40:22

and I think Austin in October

40:24

is going to be wonderful.

40:26

October 3rd through the 4th. at the

40:29

JJ Pickle research campus? Is

40:31

that real? Is that real? Is that

40:33

real? Is that not right? I

40:35

don't even know. I named that

40:37

one maybe. JJ Pickle? Yeah, it

40:39

is a different location from last

40:41

year. That's a good old Texas

40:43

boy name right there. JJ Pickle.

40:45

Get some culture in too while

40:47

you're at it. Is it close to

40:49

barbecue? That's what we need to figure

40:52

out. I mean, Carl is, I believe,

40:54

involved, so. You would hope, right?

40:56

Right. Right. Right. The

40:58

Boost. Yes, it is.

41:00

And Vamex is our

41:02

baller booster for episode

41:04

603. The only 603

41:06

there will be. And

41:08

he came in with

41:10

100,000 sets. Not bad

41:13

at all. Since I am

41:15

of course advocating for a

41:17

K8 challenge. Yes, I know.

41:19

And after that gets shot

41:21

down, maybe some kind of

41:24

exciting network challenge. Oh! They

41:26

know so well. Yeah, they

41:28

do. Also, I randomly saw

41:30

hybrid sarcasm in the Gathio

41:32

contribution release logs as I

41:34

was getting it set up.

41:37

Really? That's cool. Contease tears

41:39

to hybrid. That's really neat.

41:41

Okay, so a networking challenge. You skipped

41:43

right over the first idea. What was

41:45

the first idea? Oh, well, because as

41:47

he knows, the K8's challenge. That was

41:49

an obvious skip, right? It's obvious. I

41:51

mean, unless you want to do, I

41:53

do have plenty of raspberry pies. We

41:55

could do like a pie, a pie

41:57

thing, but I don't know if we have

41:59

another. of pies for all of us.

42:01

The idea of networking, quote unquote, as

42:03

something we should do has come up

42:06

about a bajillion times though. So we

42:08

do need to hone in on networking

42:10

and figure out what we can do

42:12

there. Because it's come up a lot.

42:14

Thank you very, very much. They, we

42:16

really appreciate that. You are the Max.

42:18

Thanks for the boost. And if

42:21

anybody has any networking challenge ideas

42:23

that we could realistically pull off.

42:25

Because I remember, one of the

42:27

things we do like. is if the

42:30

audience can participate too. Please do

42:32

some of those in. We just have to

42:34

not take down the live stream while we're

42:36

doing the show. Right, right, right. The

42:38

Dudamide comes in with 42,000 cents. The

42:41

answer to the ultimate question. I

42:43

didn't have time to get on with

42:45

the BSD challenge, but it was fun

42:47

listening to your adventures. I'm wondering, um,

42:49

how would it be if you used

42:51

something like PDF sense or Open Sense

42:54

or Trunez or Trunez Core as a

42:56

base for the challenge? Technically,

42:58

they all use free BSD, so would

43:00

that still count? No. But I

43:02

say no, but I think it would

43:04

count as one of the, you know,

43:07

because one of the points is

43:09

like a re-spend of free

43:11

BSD, because otherwise we'd have to

43:13

say ghost BSD counts. And

43:15

ghost BSD didn't count, right?

43:18

Well, I was just going to be

43:20

nice and say, you know, make it

43:22

your own challenge. Oh, yeah, if you

43:24

learn something about how the BSD went.

43:26

Oh, yeah, everybody gets a trophy. What

43:28

do you think, Brent? You get the

43:31

tie breaker here. Oh, if you put

43:33

me in that hard position.

43:35

I think we gained a

43:37

lot of background and knowledge

43:39

and context running free BSD

43:41

and that meant that when

43:43

we went into running a

43:45

non-free BSD, you know, one

43:48

of the derivatives, we had all

43:50

of that context and it gave

43:52

us a bigger appreciation. So you're,

43:55

now, you're siding with me, I

43:57

think, is what I'm hearing there.

43:59

Good. And also, worth checking it

44:02

on open sense. They've been doing a

44:04

lot of good stuff over there. Well,

44:06

I think it also says something

44:08

that all these projects are based

44:10

on previous T. So there's certainly

44:13

value to diving into any of

44:15

them. Yeah, they don't like the GPO.

44:17

Well, adversaries sent us in 32,768 sets.

44:19

The traders love the ball. Adversaries, of

44:21

course, obviously. I try to get it

44:23

right to the way you get it

44:25

wrong. And that I get it, I

44:27

get it wrong. So my trick is

44:30

just do both. What? Yeah, adversaries,

44:32

adversaries, you just, you know, use both. Okay.

44:34

Or, you know, we get a little more

44:36

colloquial and 17. Yeah, it's big 17!

44:38

Lucky 17! Lucky 17 came in, it says,

44:41

I loved all the BSD coverage. What

44:43

a wild ride that must have been. Oh,

44:45

thank you. I'm glad to hear it was

44:47

positive. I know this sounds silly, but

44:49

we did have the discussion, like,

44:52

should we be talking about BSD

44:54

in our Linux podcast in our

44:56

Linux podcast. So I'm glad that

44:58

people liked it. Thank you for that.

45:00

It's always good to hear from

45:02

you Big 17. I heard Chris

45:05

had to go and reinstall if

45:07

config on his next box. Yeah,

45:09

I still got one machine with

45:12

GoSD. It lingers. It lingers. Turd

45:14

Ferguson's here with 24,500 stats. Turd

45:16

Ferguson! How's about a video game

45:19

challenge? Three of you get a

45:21

co-op game working on Linux extra

45:23

points if the audience can join.

45:26

Oh, that sounds super fun. Turd

45:28

says, what about a no build

45:30

Gentoo challenge? Same rules as the

45:33

BSD challenge, but no building software.

45:35

How does that work? I mean, they have a

45:37

binary cash, right? Now? Could you get Knicks

45:39

working on Gentoo? That'd be hilarious.

45:41

Is this strap a small little alpine on

45:44

there? If we could somehow do the

45:46

Gentoo challenge without doing the Gentoo

45:48

challenge? I think Turd's on to

45:50

something here. You know, you get

45:52

the pre-built stage stage two or

45:54

whatever. Okay, and then here's a

45:56

philosophical philosophical question for you Brent.

45:59

If your favorite Linux distribution

46:01

was a pizza topping

46:03

what would it be and why? Oh man.

46:05

Well I think I have to say

46:07

pineapple because it's so divisive

46:10

there's a Linux for

46:12

everyone. Pineapple? What? I'm

46:14

challenging the question just a

46:16

bit because instead of distro specific

46:19

I'm just gonna say Knicks and

46:21

I don't know if this is

46:23

technically a topping more of a

46:26

sauce but uh... I'm going to say,

46:28

Knicks is like pesto because it

46:30

goes with everything. Oh, all right.

46:32

Not the cheese or the dough?

46:35

No, pesto. Okay. All right. I

46:37

was going to say it's the

46:39

sauce, but all right. All right.

46:41

All right. I was going to

46:44

say it's the sauce, but all

46:46

right. All right. All right. All

46:48

right. All right. I was going

46:50

to say it's the sauce, but

46:52

all right. All right. Okay. All

46:54

right. All right. All right. Right.

46:56

That's all I'm saying. That's

46:58

all I'm saying. Thank you, Turd.

47:01

Appreciate the boost. Our

47:03

buddy, our pal Jean Bean

47:05

Boosin, with four thousand, four

47:07

hundred and forty-four cents. Affleck!

47:09

Well, first question. Did y'all

47:12

play any with Beehive? Yeah,

47:14

I did a bit. Not

47:16

as much as I would have

47:18

liked. Beehive is the one

47:20

thing I've played around with

47:22

before where I've like loaded...

47:24

FreeBSD just to try beehive.

47:26

And I think I've also at some

47:28

point tried some sort of version that's

47:30

on MacOS. I mean, I haven't done

47:32

anything. You mean X-5? Yes, X-Ive. But

47:35

it's neat to see. I mean, you

47:37

know, we'd, KVM has worked so well

47:39

for us on Linux and FreeBSD

47:41

deserves a great virtualizer and

47:43

now it has one. At least

47:45

they're starting to integrate Vertio

47:48

support. uh... pair of virtualization

47:50

that sort of stuff uh...

47:52

show up in in in

47:54

beehive in the future i

47:56

mean like apple integrated it

47:58

with their hypervisor framework for Macoess.

48:00

So it kind of supersedes

48:02

Exhive in that respect, but I'd

48:04

like to see the Verdeo stuff

48:06

be usable in more hypervisors. Agreed.

48:09

You know, Wes, I'm looking here.

48:11

Do you see this here? It

48:13

seems like maybe the message got

48:15

degraded. You're trying to connect and

48:17

pull it again. I think the

48:19

facts got cut off for the

48:21

report. But I think he had

48:24

something about skipping the Gentoo challenge

48:26

in there. Did you see that?

48:28

Yeah, Gene's trolling us here, rightly

48:30

so. Oh, you say you're in

48:32

need of another challenge, huh?

48:34

How about that Gentoo challenge?

48:36

You've been putting off forever?

48:38

I'm not from Gen. Never

48:40

heard of it. No, that

48:42

was Brent's challenge. Yeah, because

48:44

I had to do it. I

48:46

take offense to that. I don't

48:48

think we've been putting this off.

48:51

Neil. Well Nirlpi

48:53

sent in 5,000 Sats. You're so

48:55

boost! No message on this one,

48:57

just a kind little send. Well Monti

48:59

came in with a row of adorable

49:01

ducks, 2,222 Sats. I got my

49:03

Alby Hub set up. He says,

49:06

congratulations. That's a nice little journey.

49:08

Tried several times to get Nick's

49:10

Bitcoin up and running, but just

49:12

couldn't get it to build. Alby

49:14

Hub though, I had it up

49:16

in 10 minutes. Any suggestions on

49:18

the simplest way to boost from

49:20

Alby to Boost from Alby Hubby

49:22

Hub. For this I installed the ALB browser

49:24

extension and opened up Linux Unplugged on

49:26

the Podcast Index. That probably is the

49:28

simplest way actually, really, if you have ALB

49:30

Hub. So here's the order. If you're curious

49:33

about boosting. Fountain makes it the easiest because

49:35

it's all hosted by them. Okay, that's the

49:37

easiest. Breeze is the next easiest because it's

49:39

all hosted inside the app. You get a lightning note

49:42

in the app. Put some Sats in there, you can

49:44

go find the podcast, you can boost, you don't

49:46

really have to

49:48

change podcast apps.

49:50

B-B-B-B-R-R-R-R-R-R-R-R-R-R-R-R-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E-E And then the next tier, which

49:53

you're hearing people talk about, is Alby Hub.

49:55

And Alby Hub is where you self-host the entire

49:57

thing. I mean, it is, after all, an open

49:59

source peer to... your network and anybody can

50:01

participate. And Alby Hub is a

50:03

stack of basically a lightning demon

50:06

and a bunch of great software

50:08

and a nice UI on top

50:10

of that to manage all of this. But

50:12

you don't really need it to boost,

50:14

but if you want to go to like, you

50:16

know, Chad Level Tier, Tech Guy, then

50:19

you go to the Alby Hub setup.

50:21

But if you just want to send

50:23

some stats to the show, something

50:25

like Fountain or Breeze, BR-E-E-E-E-Z, make

50:27

it really easy. Also really impressed

50:29

you tried the next Bitcoin

50:32

route too and still still

50:34

ended up on albiyab and

50:36

love it. It's great to

50:38

hear. Distrusts too, Boosin, with

50:41

10,000 cents. Boy, they are

50:43

doing a lot with Mayo these

50:45

days. Short and sweet. Planet Knicks,

50:47

y'all. I agree. Planet Knicks,

50:49

indeed. Yeah. Well, Congeroo paradox sends

50:52

one, two, three, four, five, Satoshis.

50:54

They're sending in a late update

50:56

on their free BSD challenge. I

50:58

tried out open BSD on a

51:00

raspberry pie for as a server,

51:03

but Busy Life with two young children

51:05

didn't let me do much else

51:07

apart from the manual install and

51:09

installing Neovim. Hey, you got the

51:11

important work done. I will say that I

51:14

very much liked the vibe of open

51:16

BSD. The installer is very guided and

51:18

clear and once rebooted, you have a

51:20

message that states that. You have mail.

51:22

I do like that. I love the

51:24

old Unix mail system and you got

51:27

messages in there. Said Mail has

51:29

very interesting information for newcomers. I

51:31

think I'll keep playing with this

51:33

in my free time. And once

51:36

I get a feel for BSD,

51:38

on the next BSD for me.

51:40

P.S. I'll let you guys decide if

51:42

this suffices or if I'll have to

51:45

go sober with Windows 11. I'm feeling

51:47

like it does because we didn't get.

51:49

Very many open or if all

51:51

open BSD reports, and I think it's

51:54

extremely valuable to get an open BSD

51:56

report and he got it on a

51:58

raspberry pie for So I'm saying And even though

52:00

it's not free BSD, I think he wins. I

52:03

think he wins right there. So no Windows

52:05

11 for you. You've at least

52:07

avoided that. Enough points to avoid

52:09

that. Otterbrain comes in with 5,000

52:11

stats. That's a charge-ar boost. You're supposed.

52:13

Might do them. Oh, good. I've been

52:15

hoping to get more to a monitor

52:18

setups. By the way, I want more

52:20

do-a-monter setups out there. Autobrain

52:22

says, my dual monitor setup

52:24

for work is a 2K

52:27

fast refresh display on an

52:29

X-pin display drawing tablet. Whoa,

52:31

interesting. Whoa. Really? A drawing

52:33

tablet? And it works great

52:35

on Knicks with Ghanom. A

52:37

drawing tablet on the desktop

52:40

for Linux. Autobrain,

52:42

I'd love to know what you're doing

52:44

there. That's interesting.

52:46

That's really interesting.

52:48

See Chris, I think I have

52:50

a problem with these requests for

52:53

multi-monter setups because it just makes

52:55

me really jealous Yeah, how come

52:57

it doesn't come with a budget?

53:00

Yeah You know, that would be

53:02

a great. Oh man, you know

53:04

how much if JB like had

53:07

like, you know, like some of

53:09

these YouTubeers that just get these

53:11

stupid ad deals and they just

53:14

go spend money like crazy Ours

53:16

would be we do technique makeovers

53:18

One day. One day when Linux is

53:20

the most popular thing ever and everybody

53:22

out there cares about it real hard.

53:24

I'm sure we'll get to do that. That'll

53:27

be an opportunity. Thank you everybody who

53:29

boosted the show. That's all the boost

53:31

above the 2000 set cutoff for time.

53:33

But I want to give a special

53:35

shout out to our SAT streamers. 35

53:37

of you streamed 70, 762 SATs to

53:40

the show. Just as you sat back

53:42

and listen, thank you very much. When

53:44

you combine that with our boosters. That

53:46

means we stacked a grand total of

53:48

309, 141 sets. Thank you very

53:50

much. That system is nice

53:53

and easy. Once you get

53:55

set up with fountain or

53:57

breeze or something like that.

53:59

like that. You can boost and also

54:01

check out the splits. You never know

54:03

where sometimes sending some stats to a

54:05

project, editor Drew gets set. It's just

54:08

the way the whole system works is

54:10

so freaking cool. And you just need

54:12

a podcasting app like Fountain or something

54:14

like that to try it. You can find them

54:16

at podcast apps.com. And then you

54:18

can also listen to live stream in

54:20

your podcast app and all that kind

54:23

of stuff too. And you're using a

54:25

decentralized index instead of like apples or

54:27

Spotify. We

54:29

have a couple of picks and this one's

54:31

kind of a time appropriate pick if

54:33

you will that is because Amazon is

54:35

about to remove your ability to download

54:38

any e-book that you've bought from them

54:40

in the past. Yeah I think this

54:42

takes effect on the 26th February 26th

54:44

2025. Yeah it's just a few days

54:46

after the show comes out so we

54:49

wanted to get this in here and

54:51

the app pick that we have for

54:53

you is Amazon Kindle bulk downloader. Yeah

54:55

right the backstory here since time forgot

54:57

you could... Go and download like an actual file and

54:59

then transfer it over USB or even email things to

55:01

your Kindle. You know, then you could use open source

55:04

tooling to strip DRM and then have like a proper

55:06

backup file or something you could use with a different

55:08

e-reader if you wanted or you know, like any of

55:10

the things that you might think that you want to

55:12

be able to do with a thing that you purchased.

55:14

Well, now Amazon's doubling down on the whole. You didn't

55:16

buy this. It's just a license. It's just a

55:18

license. It's just a license. It's just a license.

55:20

It's just a license. It's just a license. It's

55:23

just a license. It's just a license. It's just

55:25

a license. But, so you can still do that

55:27

if you, I think you have to have a

55:29

compatible device, because maybe they've already stopped this for

55:32

newer devices or something. I have not

55:34

tested that. Okay. You can go on to their

55:36

site and download them one by one. If you

55:38

have a few books, that's fine, but if you

55:40

have, you know, maybe you have a big library.

55:42

So, I don't have a crazy library, but

55:44

I had, you know, a hundred or so,

55:46

didn't want to do that. So I was

55:48

able to find a little open source app.

55:50

I think they intended to be open source

55:52

so buy everywhere on that one there's not

55:54

yet a license file okay well I'm thinking

55:56

I'll open an issue on that license TBD

55:58

but you know you're archiving Kindle books

56:01

in a rush anyway, so maybe that's not

56:03

your biggest concern. You can get the code

56:05

and run it on yourself on your computer

56:07

all you like. It's a typescript app set

56:09

up with buns, so I just did it

56:11

in like a little a bunch of container

56:13

and that worked totally fine. And then you

56:15

just give it some end vars with your

56:17

Amazon login info, and it's not very big,

56:19

so you can go check out the source.

56:21

And then it'll do, it spins up a

56:24

chrome-up a chrome instance to control-up-up-up-tier to control

56:26

via puppeteer-up-up-up-up-up-up-up-to-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a-a- If you do it locally too,

56:28

I didn't have a problem, but they do

56:30

have support for actually not doing it headless

56:32

and popping the browser up if you need

56:34

to do like a manual login if the

56:37

auto login doesn't work. Hmm, that's nice.

56:39

Yeah, it worked, it worked quite well,

56:41

just downloaded real fast, had progress bars

56:43

and everything. Get your books people, get

56:45

your books, get your books, and go

56:47

find somewhere else to download. We have

56:49

a bonus pick. This came in from

56:51

Bayer, listener, listener, Bayer, and I love

56:54

it. I double checked. I cannot believe

56:56

we have not picked this before because

56:58

it is 100% up our alley. It is

57:00

called N-Ping. It is a ping tool and

57:02

it is developed in rust.

57:04

It supports concurrent ping for

57:06

multiple addresses and then it

57:09

on the command line gives

57:11

you a visual chart display

57:13

with real-time data updates and

57:15

other features. Look at this boys. Look

57:18

how beautiful this is. Yeah, okay.

57:20

I like it. I am experiencing

57:22

a very strange problem with my one

57:24

workstation up in my office at the

57:26

studio, particularly in the mornings,

57:28

it's dropping packets. I can't even ping

57:31

my router, so it's on my land.

57:33

And then it kind of seems to like

57:35

settle out by the afternoon and it never

57:37

happens. But like it sucks because I get

57:39

here early in the morning, like I was in

57:41

at the studio at 4.30 a. to do the

57:44

show and that's when it's absolutely the worst.

57:46

And so it's like come on I'm here

57:48

early to get this done and it doesn't

57:50

and it wasn't until about oh I don't

57:52

know nine eight a.m. somewhere in that you

57:54

know hour it starts to smooth out. So

57:56

being able to throw this on there I

57:58

could actually visualize the issues, which was really cool. I

58:01

was going to say, you know, this seems like something you maybe traditionally

58:03

would use a tool like smoke ping for. But having this in a

58:05

zelage session in the background could be a nice way to do it.

58:07

I just need it for a couple of hours, you know, run it

58:09

there. So N-ping, we'll put a link to that, and it is MIT

58:11

licensed. I like it. So by default, you get like a graph per

58:13

thing that you're pinging. You can ping multiple things, has a slick, has

58:16

a slick little graph going, has a slick little graph going, has a

58:18

slick little graph going, a little graph going, and it, a, a, a,

58:20

a, a, a, a, a, a, a, a, a, a, a, a, a,

58:22

a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a,

58:24

a, a, a, a, recent records that that target resolved

58:26

to and the latency to each

58:28

of them. So there's a lot of

58:31

handy info right there. You know, I think

58:33

next week might be our last live episode

58:35

before we take off to scale in Planet

58:38

Knicks. Whoa. So and we may or

58:40

may not do a live episode from there.

58:42

We probably will have a pre-record all

58:44

of this. We'll try we'll try to get

58:46

this on the Jupiter broadcasting

58:48

calendar calendar and of course we'll

58:51

try to market pending in podcasting

58:53

and two. Two. Join us for 12

58:55

p.m. Pacific 3 p.m. Eastern. See you

58:57

next week. Same bad time, same bad

58:59

station. Now don't forget, links to what

59:01

we talked about today are at linuxunplug.com/603. You

59:03

can use that promo code Linux to take

59:06

50% off your scale tickets. And of course,

59:08

we're really excited about Planet Knicks. We want

59:10

to see you there. And the team at

59:12

Flocks is making it possible for us to

59:15

attend Planet Knicks. And we'll have so much

59:17

great coverage for you coming soon. And last

59:19

but not least, I'd really encourage you to

59:22

consider setting up our mumbleroom. We have it

59:24

going now for the launch on Tuesdays and

59:26

for lup on Sundays. It's a great way

59:28

to just listen and get high quality opus

59:31

right from the mixer. But if you just

59:33

listen, that's all that really matters. Thanks so

59:35

much for joining us on this week's episode

59:37

of the Unplug Program. We'll see you right

59:39

back here next Tuesday. As in Sunday!