0:12

Friends and welcome back to your weekly

0:14

Linux Dot Show. My name is Chris.

0:16

My name is Wes and my name

0:18

is Brent. Hello gentlemen coming up on

0:20

the show today Linux 6.14 is landing

0:22

and there are big improvements in there

0:24

for gaming file systems and a lot

0:26

more but you might have noticed there's

0:29

a Windows feature that's sneaking into Linux

0:31

kernel. We'll dig into what that's all

0:33

about. We'll round out the show with

0:35

some great boost, some feedback, some picks,

0:37

and more. So before we get into all of

0:39

that. Let's say hello to our virtual

0:41

luck, time-appropriate greetings, Mumble Room. Hello, hello.

0:44

Thank you, this, and hello, Brian. Hi.

0:46

Hello. Got a handful of you on

0:48

the on air and a handful of

0:50

you in the quiet listening. The mumble

0:52

room, of course, is always going when

0:54

we get our stream up and running on

0:57

a Sunday morning. And a big

0:59

good morning to our friends at

1:01

Tailscale. talescale.com slash unplugged. The

1:03

easiest way to connect your devices

1:05

and services to each other. Wherever they

1:07

are, it's modern networking the way you

1:10

really, really want it protected by. Why

1:12

I go? Great for companies, great for

1:14

self-hosters, it's secure remote access to

1:16

your production systems, your mobile systems,

1:19

you can combine complex network infrastructures

1:21

into a flat mesh network. Maybe

1:23

you got multiple VPSs, you got

1:26

some on-premises stuff, all of that

1:28

can come together, and it's really,

1:30

really fast. It's easy to deploy, it's zero

1:32

config, it is a no fuss, fuss, and

1:35

then there are things like... ACLs and additional

1:37

feature sets that you can build on top

1:39

of like integrating in with your company's authentication

1:41

infrastructure. So when you really need to take

1:43

it to the next level, tailscales there

1:46

and it works like nothing else. But the

1:48

real magic is if you go to tailscale.com/unplug,

1:50

they're going to give it to you for

1:52

free for up to 100 devices and three

1:54

users, no credit card required, not a trial.

1:56

That's the starting plan. And like me, you'll

1:59

start using it. and it will change your

2:01

networking game. It is really something else

2:03

to now have everything on my own

2:05

tailnet and nothing on the public internet.

2:08

On my devices sink over my tailnet,

2:10

we access and manage all of the

2:12

Jupiter broadcasting infrastructure over our tailnet, we

2:14

access and manage all of the Jupiter

2:16

broadcasting infrastructure over our tailnet, and then

2:19

display it to the public. It really

2:21

is very powerful. That's why thousands of

2:23

companies like Insticart Hugging Face and Du

2:25

lingo have switched to using tail scale

2:27

scale, and so many in our audience

2:30

love it too. Try for yourself or

2:32

for your business, but support the show

2:34

and get started by going to tailscale.com/unplugged.

2:36

Go there, get it for free on

2:38

100 devices. tailscale.com/unplugged. Race yourself, we have

2:41

25 days until Linux Fest Northwest. If

2:43

you can believe it, now of course

2:45

the schedule is live, we'll have a

2:47

link in the show notes. Linux Unplugged

2:49

is going to be live at 10

2:52

a.m. in HC 108 on Sunday. Whoo!

2:54

Live from Linux Fest and I just

2:56

heard from Noah that he'll be joining

2:58

us from the ASNOA program and also

3:00

Wes has a talk after the live

3:03

show in the same area maybe not

3:05

the same room but the same conference

3:07

same building I think but maybe upstairs

3:09

Linux powered payments running the lightning node

3:11

with Nick's Bitcoin at 1.30 p. Yeah,

3:14

trying to just share some of the

3:16

stuff we've found useful. If you do

3:18

want to engage with the Lightning Network,

3:20

probably the best way to run a

3:22

node. At least if you want full

3:25

control is something like Nick's Bitcoin powered

3:27

by Nick's and Nick's OS. Also, you

3:29

can just come say hi to us,

3:31

because probably a lot of us will

3:33

be there. We're going to have to

3:36

figure that out, because there's no way

3:38

we're done with the show by one.

3:40

Yeah, maybe it publishes a bit late.

3:42

Publishers after my talk. I don't even

3:44

know what we're doing. It's going to

3:47

be crazy, but we've got 25 days

3:49

to figure it out. And we'd love

3:51

to see at Linux Fest Northwest. I'm

3:53

feeling like it's. going to be a

3:55

good weather year. I'm hoping. We'll see.

3:58

Yeah, we've heard that after last year.

4:00

Is it like a seven-year cycle? I

4:02

think it's a two-year cycle. I don't

4:04

know. Well, we'll see. This will be

4:06

the trendsetter. Yeah, if Brent doesn't do

4:09

his cloud seating like last year. Stupid

4:11

lobes. Weather lobes? Hey, it worked well.

4:13

I think we got good seed funds

4:15

from that one? Oh. OK. Well, speaking

4:17

of funds, I just want to take

4:20

a moment up front in the show

4:22

and thank our members and our boosters

4:24

who have been supporting the show. The

4:26

show really is trying to do something

4:28

unique in the space. We're trying to

4:31

make the highest quality product. We know

4:33

how. We aspire to make a magazine's

4:35

worth of content in nearly every episode,

4:37

every single week, for free. And we're

4:39

doing it for a market, the Linux

4:42

listener base, that historically has failed to

4:44

make this kind of thing, sustainable via

4:46

sponsorships alone. And of course. We seem

4:48

to prefer to do things the hard

4:50

way, so we're not leaning into where

4:53

the money is these days. YouTube, click

4:55

bait, drama topics. We're using an open

4:57

platform with an open standard that doesn't

4:59

have an algo that pushes us in

5:01

front of you. And one of the

5:04

things we would like to do in

5:06

the next couple of months is raise

5:08

some funds to buy some gear, some

5:10

headsets that we can use on our

5:12

Linux Fest Northwest coverage and the Bang

5:15

Bus trip to rescue Brent New Bang

5:17

Bus. Which we expect to be doing

5:19

live shows from the van on the

5:21

road. And you know our standards are

5:23

fairly high and I've been researching this

5:26

for weeks I've asked around people's opinions

5:28

that I respect to see what they

5:30

suggest and I've landed on a set

5:32

of Headphone microphone combos that I would

5:34

like to buy for the boys It's

5:37

the Sennheiser, HMD, 26 dual ear broadcast

5:39

headset. Now these things are not cheap.

5:41

They're $500 each and they don't even

5:43

come with their proprietary cable, which is

5:45

also another $100. So it's not that

5:47

I come to this conclusion easily and

5:50

I have ordered one for myself already,

5:52

just so I can. some experience and

5:54

test it. But you know, these are

5:56

really in the sweet spot for a

5:58

quality that I think we could accept

6:01

on air that are under $1,000. And

6:03

we should be able to EQ and

6:05

process them. I'm going to talk to

6:07

Editor Drew on his opinion about that.

6:09

And we hope to make them sound

6:12

decent. And I'd love for each of

6:14

the boys to be able to get

6:16

what too. So it's around $1,300 we're

6:18

going to raise. And I'm comfortable buying

6:20

them with Boosatsatsatsatsatsatsatsats. If you would like

6:23

to help support us getting that gear

6:25

and getting us ready for the VAN

6:27

trip and for Linux Fest Northwest, please

6:29

consider boosting the next few weeks episodes.

6:31

We do come to the audience from

6:34

time to time with these kinds of

6:36

things, not always, but with fewer sponsors

6:38

out there, this is kind of how

6:40

we have to finance these types of

6:42

things. The show goes on if we

6:45

don't reach our goal, but it is

6:47

about trying to make the show as

6:49

good as possible on the road. Not

6:51

just from an audio quality standpoint, on

6:53

the go. So, and again, like when

6:56

we're at a booth at Linux Fest,

6:58

something where we don't have to have

7:00

a bunch of microphone stands and all

7:02

of that. And, you know, if you

7:04

think about it from a gearpiece, it's

7:07

a microphone, it's the stand, it's the

7:09

cable, and then it's another set of

7:11

headphones. And if you can slim all

7:13

that down into one device, it makes

7:15

it just a lot easier from a...

7:18

operational standpoint. Yeah, investing in simplicity, which

7:20

makes it cheaper in a variety of

7:22

ways, less care to bring around, less

7:24

complex, and hopefully makes it easier for

7:26

us to be able to do this

7:29

at more places or more events. Yeah,

7:31

so if you'd like to, that's probably,

7:33

you know, it's probably about 1.3 million

7:35

sat, so if you'd like to boost

7:37

in for the next few weeks and

7:40

have a go to that, just note

7:42

it in your boost, and we'd appreciate

7:44

that, and we'd appreciate that, as well.

7:46

And the show will go on if

7:48

we don't raise the funds, but I

7:51

wanted to put it out there as

7:53

it's a goal of ours, and we'll

7:55

have more. If it works out, and

7:57

thank you everybody who does support the

7:59

show in either way you do by

8:02

either through Treasury, through time, or through

8:04

your talent. We really appreciate it. Well,

8:06

just as we're sitting here recording Linux6.

8:08

14 is cooking and it is a

8:10

major release. It's 2025 has got some

8:13

barn burners and this particular release is

8:15

going to be in the next Fodor

8:17

and a bunch of other distros. So

8:19

it's an important release and it's a

8:21

banger release. Also, should we just note

8:24

at the top that it was the

8:26

rare event where in a way that

8:28

didn't matter at all, Linus Torvald's released

8:30

a day late? Yeah. I'd love to

8:32

have some good excuse for why I

8:35

didn't do the 614 release yesterday on

8:37

my regular Sunday afternoon release schedule. I'd

8:39

like to say that some important last-minute

8:41

thing came up and delayed things. But

8:43

no, it's just pure incompetence. I love

8:46

it, after all this time. Because absolutely

8:48

nothing last minute happened, and I was

8:50

just cleaning up some unrelated things in

8:52

order to be ready for the merge

8:54

window. And in the process, just entirely

8:57

forgot to cut the release. That's all

8:59

right. Yeah. It still, it turns out,

9:01

it turns out, a great kernel. A

9:03

great kernel. It really is. And we're

9:05

going to get into some of our

9:08

favorite improvements and updates. There's a lot

9:10

more to this Colonel, and we could

9:12

do a three-part episode series on it.

9:14

So please feel free to nerd out

9:16

with the links and the show notes

9:19

we have, I think, links to some

9:21

of the best coverage. But let's start

9:23

with the one that's close to our

9:25

heart. Wes is running it on his

9:27

laptop right now. Be CashFS, saw a

9:30

major step towards stability in Linux 614.

9:32

Lots of good stuff in 612 and

9:34

then 613 due to a code of

9:36

conduct ruling, Kent could not contribute. So

9:38

we didn't see any updates in 613.

9:41

So now we're getting a lot of

9:43

good stuff in 614. In particular, there

9:45

is hopefully the last planned big major

9:47

expensive but still automatic and required. on-disc

9:49

format upgrade coming with 614. So it

9:52

just happens automatically in the background, but

9:54

if you have like a giant array,

9:56

that could be some work to do.

9:58

So you might not even notice? Yeah.

10:00

In the past, I have not. running

10:03

is my root of S. Okay, so

10:05

with that comes major scalability enhancements. Yeah,

10:07

and that's the thing, right? A lot

10:09

of this stuff is as, you know,

10:11

the file system gets used in actual

10:14

scenarios and can't go through bug reports,

10:16

you kind of find out, oh, it

10:18

turns out we wanted to slightly tweak

10:20

this structure or a lot of times

10:22

it's adding points where they can keep

10:25

more information or more statistics or metrics

10:27

or like information. or back pointers. A

10:29

lot of times where you kind of

10:31

have more of the structure preserved to

10:33

make it faster when you're trying to

10:36

do some repair operation so that you

10:38

don't have to like go scan as

10:40

much data. To that point, they say

10:42

in here there's a dramatic speed up

10:44

for FSCK. Yeah. And they tested it

10:47

on. Do you see that in there?

10:49

Yeah, right. Ten Pettibite file systems out

10:51

there running B cacheFS right now. Yeah.

10:53

So I don't think my like 500

10:55

gig. Rudolph. My like. Yeah. Good to

10:58

know. If you do ever get that

11:00

stuff. It's a good reminder, right? Like

11:02

if you're not willing to test on

11:04

this format upgrades with giant arrays, then

11:06

B cache of S is not yet

11:09

for you, which is totally fine. It's

11:11

still experimental. But also, this is good

11:13

good signs, I think, of, you know,

11:15

signs of stability going forward that were

11:17

hitting places where once you get to

11:20

a point of not, at least not

11:22

even expecting on this format changes, there

11:24

still could be some if like a

11:26

major issue needs it needs it. That's

11:28

the first step towards like it being

11:31

more trustworthy for like actual serious data

11:33

or is very much so I Should

11:35

have looked more into this I just

11:37

noticed in my notes here the default

11:39

runtime has been set to self healing

11:42

enabled Yeah, there's been a lot of

11:44

self healing work going on right there

11:46

was already ways that you could trigger

11:48

that we don't yet have a full

11:50

scrub but we do have self healing

11:53

and coming in 615 is already merged

11:55

upstream I think in 615 we should

11:57

see Full scrub support. It's already also

11:59

in the upstream userland utils, but I

12:01

think in a future release. So again,

12:04

right, you need you need self-healing capability

12:06

at all, having it happen when it

12:08

detects things on the fly, and then

12:10

from there you can have scrub that

12:12

just sort of walks the whole file

12:15

system to trigger that checking. It's a

12:17

bones release, maybe not one that you'll

12:19

see a ton of changes from a

12:21

user perspective, but that's a sign of

12:23

where we're at with the infrastructure. It

12:26

was interesting, so like the moment the

12:28

614 window opens, he does a poll

12:30

that is just like, here's sinking everything

12:32

up. And then here is, here's all

12:34

these improvements too. So it's really fascinating

12:37

just to read through his description of

12:39

all of it. And it seemed just

12:41

right back at it. And I wouldn't

12:43

be surprised if by the end of

12:45

the year, or before the end of

12:48

the year, the experimental flag is removed

12:50

or... Yeah, I like that as a

12:52

2025 prediction. Could be a little low

12:54

keeper deck. I can't remember if I

12:56

actually made that in the actual predictions

12:59

episode. Do you have the locket in

13:01

sound somewhere? using is ButterFS. And there

13:03

has been some really good improvements landing

13:05

in ButterFS, including three new RAID 1

13:07

read-balancing methods, which improves how reads or

13:10

distribute across devices. So you have rotation,

13:12

which actively rotates and reads to keep

13:14

all devices engaged. That's the preferred default,

13:16

going forward. You have latency, which optimizes

13:18

for unstable or failing device transports. And

13:21

then there's DevID, which gives you manual

13:23

control over which device handles the reads,

13:25

which I could definitely see workflows where

13:27

that would be used. Yeah, absolutely. And

13:29

maybe even just as like an operational

13:32

thing for, you know, while you're having

13:34

issues with the disk, you're going to

13:36

replace or something. And of course, not

13:38

just like changes to raid one, but

13:40

also other improvements to ButterFess have landed,

13:43

which includes support for uncatched buffered IO

13:45

whose pages. will be dropped from the

13:47

page cache once the data is read

13:49

or written. The reason for this feature

13:51

are fast storage devices that can and

13:54

do fill the RAM with too much

13:56

page cache that will not be needed.

13:58

With this feature it is possible for

14:00

the read and write data and to

14:02

drop it from the cash without facing

14:05

the disadvantages and complexity of dealing with

14:07

direct I.O. Man I wish I had

14:09

that problem. My disk I-O was so

14:11

fast that I was overfilling my RAM

14:13

with stuff I didn't need. But isn't

14:16

it great that the file systems in

14:18

the kernel are attempting to keep up

14:20

with the latest and greatest hardware? Or

14:22

at least, you know, a version to

14:24

go of that? I think there's actually

14:27

been some work generally around on cash

14:29

buffer I.O. in the latest couple of

14:31

kernels to just make that better. Okay,

14:33

one last file system thing. Had to

14:35

get the file system things out of

14:38

the way because these probably really impact

14:40

the most of us, some of these

14:42

features. And the one that really sees

14:44

some love is actually your fuse file

14:46

systems. They're going to get a big

14:48

performance boost via IOU ring. So this

14:51

is a fascinating story. A gentleman who

14:53

works at DDN storage. He had a

14:55

talk at the 2023 Linux Storage Summit,

14:57

and also at BPF Summit, and he

14:59

talked about discovering a odd issue when

15:02

really taxing his MVME does. He says,

15:04

I was looking at an MVME driver

15:06

and was wondering why it was able

15:08

to avoid the bottlenecks I was seeing.

15:10

Oh, it was. It was because it

15:13

was using IO ring, but you ring,

15:15

but it was in the wrong direction,

15:17

he says. It was going from user

15:19

space to the kernel, but what Fuse

15:21

really needed was for it to go

15:24

the other way. Can you help me

15:26

make sense of that, Wes? Yeah, I

15:28

think they were investigating issues with Fuse

15:30

and noticed that the NVME driver wasn't

15:32

having the same problems. Okay. And the

15:35

way it was using IOU ring, which

15:37

is a new IO mechanism in the

15:39

kernel that makes it really fast and

15:41

cheap to be able to have the

15:43

kernel, you know, like fill. a section

15:46

of memory with data, say coming off

15:48

of a network driver or something? I

15:50

think our last kernel deep dive episode,

15:52

we did go into that a little

15:54

bit. We did, yeah. And then user

15:57

space can read from that and then

15:59

avoids a whole bunch of contact switching.

16:01

between the two which is always expensive.

16:03

So now we're having some of these

16:05

techniques being able to be applied to

16:08

fuse to speed up basically whatever you're

16:10

doing with fuse which is a whole

16:12

bunch of stuff because fuse is as

16:14

a reminder file systems in user space.

16:16

So you can have a user space

16:19

process managing things but the kernel and

16:21

other programs crucially treat it like it's

16:23

a kernel file system, right? And this

16:25

is where you can have stuff like

16:27

SSFS. And there's a whole bunch of

16:30

very, very cool fuse file systems. But

16:32

because you have to have, right, you're

16:34

talking the kernel API for file systems,

16:36

but then the kernel has to go

16:38

and talk to the user space thing

16:41

managing the fuse file system, that's a

16:43

ripe area where you need a whole

16:45

bunch of data exchange. That can be

16:47

expensive if you're doing context switches, but

16:49

with IOU ring, it can be faster.

16:52

This could have broad ramifications for all

16:54

types of fused file systems. We used

16:56

one, it was essentially a webbed outfused

16:58

file system, for managing the files on

17:00

our next cloud. And this could be

17:03

really nice to see, something we've considered

17:05

generally a little bit slower than a

17:07

real native file system, actually get a

17:09

nice little boost here. And we're seeing

17:11

IOU ring, which we talked about before,

17:14

really kind of come in and change

17:16

the game, which is fascinating to see

17:18

it kind of get rolled out in

17:20

different areas. is probably going to be

17:22

applicable in other areas. Fuse is where

17:25

they're starting, but I bet you this

17:27

idea ends up kind of spreading out

17:29

in other places. Yeah, usually does. Now,

17:31

that is all the file system stuff.

17:33

Great job, everybody. There's more in there,

17:36

of course, including there's a fascinating history

17:38

with Fuse. There has been a couple

17:40

of goes at replacing Fuse to try

17:42

to solve some of these problems. a

17:44

couple of different attempts at like fuse

17:47

two and a different type of fuse

17:49

altogether to specifically try to solve these

17:51

problems. They never really got adoption. And

17:53

then we kind of came along and

17:55

solved it at the kernel level. I

17:58

have good news and bad news. We're

18:00

actually not done with file systems.

18:02

Oh, we're not? No. OK. Because XFS

18:04

gained enhanced real-time device support. Oh, you're

18:07

right. You're right. You're right. XFS got

18:09

some love this cycle. Yeah. And it's,

18:11

you know, I think this is particularly

18:14

worth noting just because last year we

18:16

covered the XFS maintainer swap, or maybe

18:18

it was a little bit longer than that.

18:20

And that maintainer that stepped down was

18:23

a long time really trusted maintainer.

18:25

Really trusted maintainer. And what

18:27

we've seen is just impressive

18:30

development after impressive development.

18:32

So the XFS developer, Derek Wong,

18:34

explained on some of the real-time reverse

18:37

mapping and re-link support that's landing in

18:39

this patch series. He says, Christoff

18:41

and I have been working on getting the

18:43

long delayed port of reverse mapping and

18:45

reflink to the real-time device into

18:48

a manageable shape. With these changes, the

18:50

real-time volume finally reaches the feature

18:52

parity with the data device. This

18:54

is the base for building more functionality

18:56

into XFS. such as the zone storage

18:59

support that Christoff posted about last week.

19:01

I think it continues the story of,

19:03

I mean, just XFS being a great,

19:05

you know, if you don't need the kind

19:08

of things that things like ButterFets and CFS

19:10

and BCASH address, XFS is just

19:12

a great file system. And they're really

19:14

pressing to keep up with in the

19:17

ways that its architecture supports an impressive

19:19

subset of modern file system

19:21

functionality. Okay, now that is

19:23

all of the file system stuff.

19:25

This next category is, holy crap,

19:27

there's a lot of AMD stuff

19:29

landing. But first, we should acknowledge

19:31

the very important addition at the

19:33

kernel level support for the Microsoft

19:36

co-pilot key on modern laptops. Hey, did

19:38

you know that was the thing? There

19:40

you go, Brent. So if you get

19:42

a framework now, and you know, some of

19:44

those new frameworks come with a co-pilot key,

19:46

don't worry, it's going to work. So is

19:48

this in addition to the existing Windows key?

19:51

Yeah, it's on the other side now. you

19:53

know so that you got the on the

19:55

other side of the keyboard on the right

19:57

side i guess i don't know what it does on Linux

20:00

Maybe you can just map it to

20:02

something. Maybe it'll trigger your home assistant

20:04

for you. Sure. I think it should

20:06

be your compose key, but if you

20:08

don't know what that is, then maybe

20:11

it's not useful for you. Hmm. All

20:13

right, well, let's talk about the plethora

20:15

of AMD GPU driver updates. New DRM

20:17

panic supports, you get the fancy blue

20:19

screen now with AMD devices. Yeah, we

20:22

talked about that a whole bunch, but

20:24

as a reminder. to take advantage of

20:26

some of that new functionality, there needed

20:28

to be per graphics driver support. So

20:31

now we're finally seeing that on the

20:33

AMD side. It was already in a

20:35

bunch of like the base ones and

20:37

the Intel side. We're seeing early support

20:39

land for the Rx 90 Zero Series,

20:42

uh, GPUs and beyond, and the big

20:44

one is the AMDP state driver update,

20:46

which makes big improvements for laptop power

20:48

management and efficiency for AMD systems. But

20:51

there's an interesting twist in here. The

20:53

maintainer of the Linux power management subsystem

20:55

is an Intel employee. And you know,

20:57

this release was just a lot of

20:59

AMD code. And so the Intel guy

21:02

is the one that is, you know,

21:04

actually going through this and submitting it

21:06

upstream. And this is what he wrote

21:08

for the patch for 614. He writes.

21:10

The majority of changes are CPU frequency

21:13

updates, which are dominated by AMDP state

21:15

driver changes, like in the previous cycle.

21:17

Moreover, changes related to AMDP state are

21:19

also the majority of the CPU power

21:22

utility updates. Indeed, there are some pieces

21:24

of new hardware support, like the addition

21:26

of the Clearwater forest processors to support

21:28

the Intel Idol, and there are some

21:30

other improvements for other Intel chips in

21:33

there. But the vast majority... of the

21:35

updates for this 613 and 614 kernel

21:37

have been from AMD and this Intel

21:39

maintainers doing his job doing a great

21:41

job is submitting all of this upstream

21:44

and I just find that to be

21:46

a beautiful irony of free software development.

21:48

It also makes me think, I mean,

21:50

A, is there, there's probably a lot

21:53

of nice benefits for these employees of

21:55

diverse companies working together on some this

21:57

stuff, maybe there's insights or you know,

21:59

learnings from each other, but also like

22:01

in a way just open source sort

22:04

of erode some of the, artificial is

22:06

too strong of a word, but I'm

22:08

going to say artificial boundaries imposed by

22:10

like the corporate structure of things, like

22:12

here are a bunch of CPU nerds

22:15

who are nerding out together to make

22:17

all of the CP's work better. Seeing

22:19

what common problems they're solving and also

22:21

both actively contributing to these release cycles,

22:24

bits before the products are even in

22:26

the marketplace, which is so critical when

22:28

you think about the time delay, right?

22:30

So Linux 614 is going to be

22:32

in the most recent Fedor and Abunto.

22:35

But it'll be quite a while before

22:37

614 is in any of the LTS's

22:39

or a route. So they, you know,

22:41

when they get that groundwork in there

22:44

now, people that take advantage of the

22:46

in between distribution releases get to use

22:48

the drivers, but it takes years before

22:50

it's in the enterprise-grade stuff. And so

22:52

it's really... incumbent upon them to try

22:55

to get this stuff in as early

22:57

as possible and that's why and before

22:59

they've even finished all of the driver

23:01

work they're getting the primitives in now

23:03

as early as they can so that

23:06

way the next cycle they can just

23:08

get you know a little bit more

23:10

in there and then hopefully get it

23:12

out to end users as fast as

23:15

possible. We'll see but that's the that's

23:17

the interesting game that they both have

23:19

to play and both AMD and Intel

23:21

just by watching this do seem very

23:23

committed to that process. They're getting it

23:26

done. One password.com/unplugged. That's

23:28

the number one password.com and then unplugged

23:30

all lowercase. You want to go there

23:32

to support the show and to learn

23:34

more about one password. Here's how I

23:36

like to think about this. I have

23:39

a question for you. Do your end

23:41

users always and I mean to always

23:43

use the IT approved applications services and

23:45

devices? I don't know if any users

23:47

do. I think there's just so much

23:49

out there now. There's so much getting

23:52

pushed them at them from a consumer

23:54

level. This is going to change how

23:56

you work level. So how do you

23:58

keep your company's? data saved when it's

24:00

sitting on all these unmanaged apps or

24:02

devices. That's where one password comes in.

24:05

Again, it's one password.com/unplugged. They have the

24:07

answer to this question, extended access management.

24:09

One password extended access management helps you

24:11

secure every sign in for every app

24:13

on every device because it solves the

24:15

problems that traditional Iams and MDMs just

24:18

weren't built to touch. You secure every

24:20

app, every device. and every identity even

24:22

the unmanaged ones like when I was

24:24

a contractor coming in you go to

24:26

one password.com/unplugged check it out this really

24:28

does change the game now one password

24:31

is famous of course for their beautiful

24:33

password management now they've taken up the

24:35

next level from that award-winning product to

24:37

extended access management they're really bringing everything

24:39

they've learned including their best practices you

24:41

know they undergo regular third-party audits they

24:44

have the industry's largest bug boundary And

24:46

it's really great to say that one

24:48

password exceeds those standards that are just

24:50

set by other authorities and other, you

24:52

know, whatever the letter of the law

24:54

might be, they go above and beyond.

24:57

I love to see that. It's really

24:59

the first security solution that brings all

25:01

these unmanaged devices and apps and identities

25:03

under your control. That's the big deal.

25:05

That's the big shift. So go check

25:07

it out and support the show. Go

25:09

make sure every app, every device, every

25:12

identity is secure, even the unmanaged ones.

25:14

Go to one password.com/unplugged. That's one password.com/unplugged.

25:16

As with many previous kernels, we do

25:18

see a little bit of gaming love

25:20

in Linux 614 as well. Looks like

25:22

Linux 614 kernel will officially include the

25:25

completed NT sync driver, which is a

25:27

major improvement for Linux gamers using wine

25:29

and Proton, especially via Steamplay or SteamOS.

25:31

An NT sync feature in our Linux

25:33

kernel? Yeah, that's right. It's kind of

25:35

neat, right? Now we have a... a

25:38

specific synchronization primitive enabled inside the kernel

25:40

specifically to support compatibility with Windows applications.

25:42

Very fascinating. They say here the driver

25:44

enhances the emulation of Windows NTS synchronization

25:46

primitives which are critical for accurately replicating

25:48

Windows behavior and improving performance in many

25:51

Windows games running on Linux. Though NTSink

25:53

was introduced in Linux 13 it was

25:55

initially incomplete and not functional. 614 marks

25:57

its full implementation and usability. As for

25:59

the motivation behind the work, well, it's

26:01

clear in the benchmarks. The gains vary

26:04

widely depending on the application and the

26:06

user's hardware, but for some games, anti-synchronization

26:08

is not a bottleneck, and so no

26:10

change will be noticed. Some other games,

26:12

though, improvements to even 150% are normal.

26:14

Yeah, okay, so there is some nuance

26:17

here as well. Right now, this is

26:19

specifically helpful for, or going to be

26:21

for, upstream wine. And it is the

26:23

newest and shiniest and most compatible with

26:25

Windows implementation, but there's been earlier attempts,

26:27

specifically E-Sing and F-Sing, which have basically

26:30

been like gaming focused mechanisms to enable

26:32

improvements to emulating games specifically. And since

26:34

I think like kernel 514, we've had

26:36

few text two in the kernel. which

26:38

Proton is patched upstream wine to take

26:40

advantage of, but I think upstream wine

26:43

does not. Partially because it was kind

26:45

of gaming focused, and there was some

26:47

thoughts that it was kind of a

26:49

hack and didn't really like the specific

26:51

interface being created for wine protocol. And

26:53

it doesn't solve it for all games,

26:56

like solves it for some of the

26:58

games, which is why some of these

27:00

games won't see an improvement. Yeah. And

27:02

the root of this is, okay, so

27:04

in the Windows world, you've got the

27:06

Windows NT, you've got the... influenced by

27:09

stuff like VMS. And there's a whole

27:11

bunch of objects and you have handles

27:13

to these objects, which if you've ever

27:15

done Windows programming, you've seen handles. freaking

27:17

everywhere, right? And if you're a game

27:19

or just a program, often you need

27:22

to wait for things, right? So like

27:24

in a game it might be, wait

27:26

to be told that like the frame

27:28

I wanted to draw is done painting

27:30

or copying to the GPU memory or

27:32

whatever it is, or I need to

27:35

wait for a keyboard or a mouse

27:37

and event for the user to do

27:39

my next thing, or maybe a network

27:41

package shows up telling you what the

27:43

enemy participant did and how to update

27:45

their game state. And so there's oftentimes

27:48

where you've done your immediate work. Wake

27:50

me up when any of these things

27:52

happens and then based on what it

27:54

is I know what to do next

27:56

to advance the process and Due to

27:58

the way Windows is designed they've got

28:01

some pretty nice APIs for this including

28:03

wait for multiple objects So you can

28:05

kind of just hand it a whole

28:07

smattering of things that you want to

28:09

wait for regardless of it's like a

28:11

file handle or a network thing or

28:14

some status update to a particular process

28:16

or thread and Windows will set that

28:18

up for you and While Linux has

28:20

things like EPOL and other mechanisms too

28:22

specifically wait for things like file handles,

28:24

there's various equivalencies, of course, because right,

28:26

ultimately Windows and Linux make your computer

28:29

work in roughly the same way, but

28:31

the APIs are different enough that like

28:33

originally Wine had to have kind of

28:35

like a helper thread or process that

28:37

could sort of translate between the Windows

28:39

system call in user space and then

28:42

implement it under the hood with the

28:44

right kernel features. And as you can

28:46

imagine... Like that's not going to perform

28:48

super great. It was fine for some

28:50

things. It did work, but like it

28:52

was just going to hit bottlenecks, especially

28:55

with games. And so that's where we've

28:57

had multiple attempts with e-synch and then

28:59

F-synch, which has been the most successful

29:01

and widely used with Proton. But now,

29:03

finally, we're getting NTsynch, which is it

29:05

prioritizes compatibility. So it's like the best

29:08

translation in the kernel for this. style

29:10

of synchronization primitive, but it also looks

29:12

like it has some quite nice performance

29:14

game. So especially for upstream wine, you're

29:16

going to see those big numbers that

29:18

you're seeing reported. If you're doing proton,

29:21

there still has not yet a... in

29:23

theory in the future they'll probably switch

29:25

to NT sync and there'll be some

29:27

slight performance improvements and compatibility improvements. And

29:29

like from a practical standpoint for end

29:31

users they're going to get essentially a

29:34

new device a new virtual device right

29:36

like a slash dev slash NT sync

29:38

device. Yeah and that's the mechanism for

29:40

how things can sort of go request

29:42

that and wait on it from the

29:44

kernel. Yeah and each time each process

29:47

opens up the device they get a

29:49

unique instance. But it's shared across threads,

29:51

which is so they get like their

29:53

own ID, but there's some shared space

29:55

there too Yeah, and that's usually the

29:57

case right so for processes they have

30:00

different memory space and threads get to

30:02

share so that way you can have

30:04

multiple threads that are doing things But

30:06

they can all be they can like

30:08

share waiting or locks or use that

30:10

use that for a shared synchronization primitive

30:13

to like communicate between threads sort of

30:15

like the fuse IOU ring improvements we

30:17

talked about earlier in the show this

30:19

NT sync that is definitely targeted at

30:21

gaming right now seems like there would

30:23

be other complex Windows applications that would

30:26

benefit from this. And so while it

30:28

is initially useful for gaming, I could

30:30

see in a year or two we've

30:32

discovered some other application that didn't work

30:34

before is now working because of this

30:36

too. Yeah, right in the big picture,

30:39

it's just like a big nice step

30:41

up for our Linux's ability to be

30:43

compatible with Windows applications through one. It's

30:45

just cool because... I mean, there's already,

30:47

again, there's already been some of these

30:49

kinds of efforts, but just the idea

30:52

that the colonel is being adapted to

30:54

support this not emulation user space thing

30:56

of wine that somehow seems to work

30:58

surprisingly well. That's just, that's, that's, that

31:00

takes a lot of coordination and planning

31:02

and years of effort to get across

31:05

the finish line. And like, you know,

31:07

folks had to make the case that,

31:09

hey, you know, people are trying to

31:11

run programs not even intended for the

31:13

operating system you design, but can you

31:15

design, but can you make this change,

31:18

please? They were going to sell on

31:20

this proprietary hardware. Yeah, right. But it

31:22

does show you that if you've got

31:24

good code that works and you're willing

31:26

to do the work and you show

31:28

up to the Linux community, kernel community

31:31

with, you know, proof of receipts, they're

31:33

probably. going to take it. They're really

31:35

pretty much down to take most things.

31:37

If you're doing the maintenance, you're doing

31:39

the legwork. You know, like yeah, and

31:41

you have you can point to real,

31:43

you know, use cases or people who

31:46

want it. This is how Samsung got

31:48

a Sam server in the Colonel West.

31:50

Jeff makes a good point to and

31:52

this is definitely true right through the

31:54

history of computing. Games have often been

31:56

one of the best ways to sort

31:59

of like push the limits for how

32:01

to make the limits for how to

32:03

make these things. application. Just in times

32:05

of Brent can switch back to Lightroom.

32:07

Now we were thinking and we'd like

32:09

to hear your thoughts about maybe doing

32:12

a gaming episode next week. Nothing crazy.

32:14

Maybe just one or two games and

32:16

then also a bit of a hardware

32:18

update. So even if you're not a

32:20

gaming person there'll be some hardware talking

32:22

there too. So if you have thoughts

32:25

on gaming on Linux, game you're loving

32:27

right now, particularly those that support co-op.

32:29

Send them in to us. Boost them

32:31

in or go to linuxunplug.com/contact and let

32:33

us know because we've been thinking about

32:35

doing a gaming episode literally for about

32:38

two years So maybe we'll do it

32:40

next week. We've got a couple of

32:42

things. I got a new device in-house

32:44

Reminde me to show it after the

32:46

episode. It's pretty slick I want to

32:48

talk about that next week. So that's

32:51

been on our mind. It's been a

32:53

long time. The gaming episodes tend not

32:55

to be super popular so if you

32:57

know to be super popular so if

32:59

If we really got ahead of it,

33:01

we could do like a gaming meetup.

33:04

If we were playing a game, people

33:06

could join. Well, okay, how about this

33:08

is a goal? We come up with

33:10

a co-op land-style game that we could

33:12

try that we could use at Linux

33:14

Fest. And we demo it. So we

33:17

would love some suggestions for that I

33:19

literally don't know. Maybe producer Jeff could

33:21

help. The other thing to, to, the

33:23

other thing to kind of just sort

33:25

of just think on, for us to

33:27

think on, for Linux, for Linux, for

33:30

Linux Fest, for Linux Fest, is where

33:32

we're, for Linux Fest, is where we're,

33:34

for Linux Fest, for Linux Fest, for

33:36

Linux Fest, is where we're, for Linux

33:38

Fest, for Linux Fest, for Linux Fest,

33:40

is where we're, for Linux Fest, for

33:43

Linux Fest, for Linux Fest, for Linux

33:45

Fest, for Linux Fest, for Linux Fest,

33:47

for Linux Fest, for Linux Fest, for

33:49

Linux Fest, for Linux Fest, for Linux

33:51

Fest, for Linux Fest, Anyways something that's

33:53

been on our mind new hardware and

33:56

all of that so let us know

33:58

your thoughts there and I don't think

34:00

anything's ready yet to try out this

34:02

new NT sync thing. I wish we

34:04

had an opportunity to test it with

34:06

a game, but when one does, I'm

34:09

already going to be on 614, so

34:11

I'll be ready. Did you know your

34:13

ad could be right here right now?

34:15

I could be talking about your company.

34:17

Your Linux Unplug podcast is looking to

34:19

book a sponsor and we're running a

34:22

special for the second half of Q2.

34:24

Now you know we're kind of picky

34:26

who we work with, which is limited

34:28

us, but... If you're listening to this

34:30

show, well, that's a good start. Chances

34:32

are you could be a great fit.

34:35

So if you want to reach the

34:37

world's largest Linux audience on a show

34:39

that never misses a week and has

34:41

been going for nearly 12 years strong,

34:43

reach out to me. Chris at Jupiter

34:45

broadcasting.com I'd love to work out to

34:48

deal with somebody who's in the community

34:50

that has a business or a product

34:52

or a service that they'd like to

34:54

reach. We'll make a pretty good deal.

34:56

And I think it could be neat.

35:01

We've got a special little piece of

35:03

mail here from Grant who asks a

35:05

little for some advice, let's say. It's

35:07

his greetings, Chris, Wes, and Brent. I

35:09

keep hearing you guys talk about this

35:11

Knicks thing, so I'm finally going for

35:14

it. I was wondering if you have

35:16

any tips for Knicks configs on a

35:18

laptop specifically. From what I researched so

35:20

far, most of it has to do

35:22

with power optimization while on battery. Since

35:25

using Linux on a laptop is literally

35:27

Linux unplugged. Maybe another question is, what

35:29

are some community recommendations for Linux on

35:31

a laptop? Love to hear what the

35:33

best Linux community around, Linux Unplugged, has

35:35

to say about this, and thanks in

35:38

advance from Grant. It's a good question,

35:40

Grant. I feel like I'm going to

35:42

be in the laptop market towards the

35:44

end of the year. It has been

35:46

interesting living on a Sahi on the

35:48

Macbook and very limiting, and I eventually

35:51

want to get something, but I want

35:53

a very nice laptop experience. Something real

35:55

smooth. Framework. Framework. Framework. Framework. a contender.

35:57

Of course, I'm going to look at

35:59

System 76. I'm going to be looking

36:01

at ThinkPads, too. But I think Grant's

36:04

asking two questions here. Number one, what

36:06

is a great distro hardware combo? And

36:08

like, what should he look into for

36:10

Knicks to pre-set up as Knicks for

36:12

the best chances on that hardware? Like,

36:15

I know Brent has Looked at config

36:17

specifically for framework laptops. Yeah, well, there's

36:19

the Knicks OS hardware community repo. So

36:21

that's one place you can pull in

36:23

stuff. That's the thing. That's the thing.

36:25

That. You know, because right, a lot

36:28

of it is going to be making

36:30

sure you're fully enabled whatever settings or

36:32

kernel parameters or driver options for your

36:34

specific rig. Because after that, honestly, you

36:36

know, the lunch world's gotten a little

36:38

bit better. I've been pretty pleased with

36:41

the power management options, at least in

36:43

things. So there are you know various

36:45

demons and things like power top and

36:47

tools to monitor but for the most

36:49

part I'm not currently fussing with those

36:52

as much as I used to. Yeah

36:54

I just leave mine in balanced mode

36:56

for the most part and then power

36:58

saver when I need it and boost

37:00

when I when I need it I

37:02

just what I'm looking for is Brent

37:05

he has the framework with I think

37:07

it's the 12th gym mobile right Brent?

37:09

Yes it is. And you have this.

37:11

problem that I witness all the time

37:13

when you're around where you close up

37:15

the laptop and then you come back

37:18

to it a couple hours later and

37:20

it's hot. Yeah, I have to say

37:22

like maybe two out of ten times,

37:24

I guess that's one out of five,

37:26

I will pick my laptop out of

37:28

my bag and it'll be roasting. The

37:31

rest of the time it kind of

37:33

behaves as it should. And I've been

37:35

trying to figure out what is the

37:37

combination of what I'm doing that causes

37:39

that to happen. Now Jeff, I know

37:42

you've struggled with this kind of thing

37:44

before and you probably have a bunch

37:46

of advice for me, but it has,

37:48

I would say, gotten better. So if

37:50

there's constantly tweaks happening where West suggested

37:52

in the NextWest hardware repo to try

37:55

to solve some of these common problems,

37:57

but it seems, I don't know, suspend

37:59

as... just in not so great a

38:01

place as it was, like, let's say,

38:03

on an old thing pad laptop from

38:05

back in the day. So it is

38:08

something that's getting better, and there are

38:10

some tweaks you can do, but I

38:12

gotta say, like, as a warning, part

38:14

of the reality is, sometimes it

38:16

just doesn't go so hot. PJ, I'd

38:18

like to hear your thoughts, but I

38:21

would think if he's on a pretty

38:23

modern kernel, and he's using a 12-gen

38:25

system which has been out for

38:27

a few years, specifically does have

38:30

S3 sleep and that's one of

38:32

the biggest problems that we're seeing

38:34

with modern laptops is Intel

38:36

and AMD have kind of pushed away

38:39

from that and it's up to the

38:41

motherboard really the the

38:43

upstream you know laptop maker to

38:45

put it back in and everything

38:48

defaults to S5 sleep which doesn't

38:50

fully shut down with the frameworks

38:52

I've seen that you need to

38:54

go into the bios and turn

38:57

off USB power. because all of

38:59

the little dongles will just sit

39:01

there and suck power down when

39:04

it's trying to sleep even if

39:06

it is an S3. So you

39:08

might be seeing that, but it's

39:11

a problem. My Lenovo does it.

39:13

My last HD did it. It's

39:15

all S5. These things aren't

39:17

actually going to suspend

39:20

a RAM. They're just basically

39:22

shutting PPD. PPD to tombi.

39:24

And it's really nice because

39:26

you can nest profiles so

39:28

you could have the default

39:30

balance or laptop profile and

39:32

then make your own little

39:34

changes on top. Yeah, okay. So that's,

39:36

you know, maybe the advantage to

39:39

having a distribution tune that

39:41

stuff for you. Something Brent could

39:43

look into, I suppose. On Nick's

39:45

OS, I see there's a power

39:47

management option. Which has stuff like CPU

39:49

frequencies in here power top You know

39:51

all kinds of stuff power profiles demons

39:53

an option as well. So there's various

39:55

things in Knicks OS too. Well, I'd love

39:57

a great experience on my next lab

40:00

Something really, you know, hassle-free, works

40:02

with distros, just great, has some

40:04

decent little performance for maybe some

40:06

basic gaming, and sleeps. Oh, I

40:08

was going to say in general

40:10

for laptops, I think we've all

40:12

been pretty happy, I mean, it's

40:14

not laptop specific, but happy with

40:16

things like Z-Ramp. Yeah. Yeah. I

40:18

think there are also, you know,

40:20

there are compromises to be made.

40:22

You know, I could, if I

40:24

could get it to boot super

40:26

fast, I could possibly live without

40:28

sleep by. Oh, I don't love

40:30

that, but that could always be

40:33

a route I could explore. If

40:35

it meant I could have a

40:37

fantastic laptop that I really love,

40:39

that might be a compromise I

40:41

could make. I fantastic, you mean

40:43

loud fans, right? Yeah, no. I

40:45

do not, sir, I do not.

40:47

And now, it is time for

40:49

the boost. Yes, it is, and

40:51

we start with our Boller booster

40:53

this week, who is the tech

40:55

geek with 45,000 stats. And

40:58

he just sends along, hey, J.B. team,

41:00

keep up the good work. Well, thank

41:03

you for the value. Turd Ferguson Boosin,

41:05

with 41,088 cents. Turd Ferguson! Oz is

41:07

a triple-turd. Gosh. Boosting the two-y challenge.

41:10

Uh-oh. Envy Top, it's not just for

41:12

invidia. Did you know Intel, AMD, and

41:14

even, at least sorta. The M1 and

41:16

M2 GPUs are supported. And V Top,

41:19

it sounds like it's for invidious. Okay,

41:21

another trip here, Brent's traveling. We could

41:23

just, if there was a website, is

41:26

Brent Traveling? It's be, yes. Pack Trippy!

41:28

The two-e love child of ping and

41:30

trace route. Sounds like you could have

41:33

used this just the other day, actually.

41:35

Do we know about trippy? I don't

41:37

know if we do. It's a two-e,

41:39

huh? Let's go look this up. Let's

41:42

go look this up, look this up,

41:44

look this up, look this up, this

41:46

up, this up, trippy, trippy, trippy, trippy,

41:49

a network, a network, a network, a

41:51

network, a network, a network, a network,

41:53

Oh my god, it's a riff on

41:56

the Clippy logo too, which is funny.

41:58

I like this. I like this a

42:00

lot, Brett. I think you could use

42:02

this. I could use this actually. What's

42:05

it written in? The question, oh, look

42:07

at that. Did you know this already?

42:09

I know, I just wanted to know.

42:12

Yeah, 99.6% rest. It's not a hundred,

42:14

but I'll take it. Yeah, it's pretty

42:16

close. So that's trippy. All right, that's

42:19

good. Now, do we have a, you

42:21

said there was three of them? Yes,

42:23

I did. There's one more. Worried you

42:25

might get bored during the two week.

42:28

Try TTY solitaire. Now, with mouse support.

42:30

TTY solitaire. All right. Solid. You know,

42:32

they told you your flight would have

42:35

Wi-Fi. It doesn't. But you already downloaded

42:37

TTY solitaire. Turd Ferguson, coming in with

42:39

some good to-y picks. Turd Ferguson! I'm

42:41

feeling like this challenge is going to

42:44

happen, boys. I'm feeling it. Are there

42:46

any to-y games when you can combine

42:48

these two things? Well, there's the solitaire,

42:51

I suppose. You know on my flight

42:53

to this here continent there was some

42:55

on the entire flight that played solitaire

42:58

and I could not believe it but

43:00

maybe with cards no no no no

43:02

no no on their phone my wife

43:04

could do it with cards she just

43:07

she brings a deck of cards in

43:09

her purse it's a great idea to

43:11

have we were out and about last

43:14

night and we had a half hour

43:16

to kill with the shuttle so she

43:18

just bus out the cards just goes

43:21

to it great battery life Vlkangaroo Paradox

43:23

came in with 30,000 Sats across three

43:25

booths. Linux or something. Great coverage of

43:27

the Russ drama recently. I really appreciated

43:30

the objective breakdown of events and your

43:32

take on it. Also loved The Deep

43:34

Dive. Sorry, Brent, he says. I think

43:37

you know, you know, you know, hold

43:39

on. Brent's got a reputation. Right then.

43:41

Let's get ourselves settled in for a

43:44

wee deep dive. I was trying to

43:46

speed right past that one so you

43:48

wouldn't have a chance to play it.

43:50

Shakesfist at me, you both. You couldn't

43:53

hear a... I think you could sprinkle

43:55

these once in a while and that

43:57

could be enjoyable for people who aren't

44:00

super technical. Good way to learn things

44:02

too. Also plus one for the tooey

44:04

challenge. I'm all for spending as much

44:06

time as possible in the terminal versus

44:09

the browser. I think a tooey challenge

44:11

with a point penalty for mouse usage

44:13

could be a fun spin. Love the

44:16

planet Knicks coverage could not justify flying

44:18

from Europe to go for a couple

44:20

days over a thousand euros total for

44:23

that trip. Can we expect you guys

44:25

to make the trip to Nixon EU

44:27

maybe in the coming years? These usually

44:29

take place October-ish in Germany or around

44:32

there. I imagine the price would be

44:34

an issue, but you know, the beer

44:36

here is worth it. Yes. The truthful

44:39

answer is, I imagine when my kids

44:41

are a bit older, I'll probably be

44:43

traveling on longer trips, where I could

44:46

go somewhere I could spend like two

44:48

weeks, or a week, you know, that

44:50

would be really great, but... That'll probably

44:52

still be a couple of years. Maybe

44:55

every now and then. You know, I

44:57

never want to say never. You never

44:59

know, right? Maybe Wes could talk me

45:02

into something right to this my arm.

45:04

You guys can be pretty convincing when

45:06

you put your minds to it. Thank

45:09

you for the boost, Congo. It's good

45:11

to hear from you. The music came

45:13

in with 5,000 sets. You're supposed. Firstly,

45:15

I finally got around to say it

45:18

at my own Bitcoin with Nick's Bitcoin.

45:20

Hey! I am personally excited for Ghanom

45:22

48 as controlling the Orca screen reader

45:25

under Whalen has taken a big step

45:27

forward. I've wanted to get X-Orga off

45:29

for ages now. Come May and Nix

45:31

2505, I may just be able to

45:34

do that. That's great to hear. Yeah,

45:36

that's right. Yeah, there is some nice

45:38

changes there. And Wes, look at you.

45:41

helping people get Elby hub up and

45:43

going. Yeah, and props to folks like

45:45

the Muzo and Jiraka who have taken

45:48

the leads on getting the module going.

45:50

I got a base flake happening, but

45:52

also I've been negligent. So this is

45:54

a great reminder. I'll go try to

45:57

see if I can spend some time.

45:59

that repo. adversary 17 booths in with

46:01

5,000 sets. Put some macaroni and cheese

46:04

on there too. The dynamic we don't

46:06

even know is pretty funny. Oh yeah

46:08

the members version somehow got set up

46:11

with dynamic AI inserted ads for all

46:13

kinds of crazy things. You got a

46:15

good kick out of them the whole

46:17

show. Thank you. Of course the only

46:20

proper song to play on the Clinton

46:22

sexy saxophone It's careless whisper. That was

46:24

true. Yeah, that would have been, yeah,

46:27

it's true. Nice to hear from your

46:29

adversaries. Thank you. Well, Brooke loves you,

46:31

sent in 3,535 cents. Allegedly. Hey, J.B.

46:34

Have y'all tried Katie plasma mobile yet?

46:36

I'm daily driving that thing on my

46:38

minisform v3 with Fedora, and it's pretty

46:40

awesome. I recommend giving Linux on a

46:43

tablet a try if you'll haven't in

46:45

a while or even at all. Keep

46:47

up the good work and thanks for

46:50

making my work travel more enjoyable. This

46:52

is a great suggestion, Brooke, and you're

46:54

kind of reading my mind this Friday.

46:56

I was browsing Amazon looking to see

46:59

if there's like quote-unquote Linux X86 tablets

47:01

or even Linux arm tablets. You know,

47:03

I do this thing where I have

47:06

tablets around the home and... I want

47:08

to have more in the studio to

47:10

control home assist in different areas. And

47:13

why have Android on there when I

47:15

could have Linux on there? And it

47:17

would give me an excuse to play

47:19

with plasma mobile properly, which I've only

47:22

ever done in emulator. It's probably more

47:24

realistic. I was going to say we

47:26

pressured Brent, because it seems like his

47:29

kind of thing. Yeah, yeah, you're already

47:31

doing it. Or we could double course.

47:33

Well, I could take my learnings and

47:36

use that to pressure. Or flash's stuff

47:38

and set them stuff and set them

47:40

up. With me, you could get, if

47:42

you get me in right, you know,

47:45

the right mode, I'll do it in

47:47

a flash, but with Brent, it's like,

47:49

it's going to take like two years,

47:52

you know, before Brent gets an Android

47:54

tablet that runs Linux. Well, you know,

47:56

only one of us carries two devices

47:59

in their pockets, so I think, you

48:01

know, having a backup built in, you're

48:03

more suited for this thing. 930 Sats.

48:05

Yes, it is hammer time. And they

48:08

say, sorry, this took so long, but

48:10

here is my write-up on how to

48:12

control the front gate via Meshtastic and

48:15

MQTT. Oh, yes. Awesome. This was one

48:17

of the things we wondered. Could you

48:19

do IOT control with Meshtastic? He sends

48:22

us his get-up repo, which we will

48:24

include in the show notes, and its

48:26

instructions to create an MQTT controlled input

48:28

board that runs his own logic. on

48:31

low radio, on low power radio instances.

48:33

And you can use it to control

48:35

your gates. Indeed. That is so. A

48:38

whole bunch of comments in here too.

48:40

I'm saving that. I am saving that.

48:42

That's really great. Thank you for sending

48:44

that in Ambient Noise. And well done.

48:47

Well done. Thank you everybody who participated,

48:49

including you SAT streamers. 25 of you,

48:51

stream Sats. We stocked a Humble 2923.

48:54

Not too bad for those of you

48:56

just sitting back and enjoying the show.

48:58

Thank you very much for doing that.

49:01

Then we combine that with all the

49:03

people that generously boosted him with a

49:05

message. We read the ones above the

49:07

2000 SAT cutoff line. And we stacked

49:10

156,476 sets, not our strongest week. Smoke

49:12

if you got them. I'm not going

49:14

to say it's a total loss though.

49:17

Good, good. But you know, it could

49:19

be better, especially with some of our

49:21

ambitious plans we have coming up for

49:24

the rest of the year. So if

49:26

you'd like to get in on the

49:28

fun. There's a few ways to do

49:30

it, but you've got to get some

49:33

Sats. Strike is a great way to

49:35

do it in over 110 countries. River

49:37

in the US, Bitcoin well in Canada.

49:40

Get those Sats, send them over the

49:42

lighting to a podcasting, 2.0 app, like

49:44

Fountain. The workflows also in Fountain are

49:47

really simple too now. They're getting easier

49:49

and easier to just do it within

49:51

all one app, and it just keeps

49:53

getting better. You can send your message

49:56

in, and we'll read it on the

49:58

show above 2000 Sats. and all 31

50:00

of you who boosted in and supported

50:03

episode 608 of your unplugged program. It

50:05

really does mean a lot to us.

50:07

We appreciate it. And of course... Of

50:09

course, big shout out to our members,

50:12

who keep us going too, you are

50:14

all fantastic and much appreciated. Well, moving

50:16

right along, gentlemen. Moving right along. We

50:19

have a pick that may help some

50:21

of you break your grammarly habit. Have

50:23

you wondered what are you going to

50:26

do to get over your grammarly habit?

50:28

Finally, take back your freedom from big

50:30

grammar. Yep. Well, that's where eloquent comes

50:32

in. It runs as a local service

50:35

in the background to do local offline...

50:37

I guess proofreading, it uses a language

50:39

tool server and it's available as a

50:42

back-in instance to Firefox, library office, and

50:44

others, it is GPL3, it does also

50:46

have a little standalone desktop gooey where

50:49

you can paste your text in and

50:51

then it gives you a grammarly-esque pop-up

50:53

to do spelling. No, if people have

50:55

never used grammar, they have no idea

50:58

what I'm talking about, but it's sort

51:00

of like grammar and spelling and... sentence

51:02

structure all in one. And it's a

51:05

good service, but to do this, they're

51:07

reading everything you write, and they want

51:09

to integrate into everything constantly. And the

51:12

fact that you could actually run this

51:14

as a back-end instance for Firefox, which

51:16

I have not done, but looks really

51:18

great. Like that is the perfect combination

51:21

for me. And it is also available

51:23

as a flat hub. Well, and this

51:25

backing tool is cool, language tool, open

51:28

source proof reading software. Yeah. English, Spanish,

51:30

French, German, Portuguese, Polish, Dutch, and more

51:32

than 20 other languages. Yeah. Finds many

51:34

errors that a simple spell checker cannot

51:37

detect. There is a collection of tooling

51:39

out there right now. Like this language

51:41

tool and Piper, from the homo system

51:44

folks in Rusper, which do text to

51:46

speech and speech to text. And then

51:48

they have their Wyoming protocol, which lets

51:51

it communicate. All of it is open

51:53

source, all of it running on Linux,

51:55

and we're not doing anything with it.

51:57

It's really weird. Maybe some show projects

52:00

we need to get ahead of. You

52:02

can actually even train your own voices

52:04

with it. You can have it model

52:07

on your own voice and generate stuff.

52:09

It's just right there for the taking.

52:11

And some project just needs to integrate

52:14

it into some sort of application or

52:16

something. I mean, it's very useful and

52:18

home assistant. Right. Of course. Piper alone

52:20

and Whisper alone. Just feels like there's

52:23

a lot of potential there. Anyways, today

52:25

we're talking about eloquent and we'll have

52:27

a link to that in the show

52:30

notes. Should you want to check it

52:32

out. I also see it looks like

52:34

there's some stuff that does, makes like

52:37

a language, or an LSP implementation on

52:39

top of language tool, so you could

52:41

get this kind of stuff in your

52:43

VS code too, maybe. And don't forget,

52:46

we're trying to raise some funds for

52:48

headset for the boys, so if you've

52:50

got any tips for gaming or games

52:53

you love, or co-op games, something we

52:55

could do at Linux Fest in just

52:57

25 days. Link to the

53:00

schedule and all of that is

53:02

in the show notes as well.

53:04

And you know where you find

53:06

that? Get ready for this. linuxunplug.com/608.

53:08

That's all you need to know.

53:10

linuxunplug.com/608. See you next week. Same

53:13

bad time. Same bad station. And

53:15

of course, we love it when

53:17

you join us live. It just

53:19

gives it a vibe and we

53:21

make it a Tuesday on a

53:23

Sunday. It begins at 10 a.

53:26

Of course we have it in

53:28

your local time at Jupiter broadcasting.com/calendar.

53:30

If you're enjoying a podcasting Tudoro

53:32

app, I try to get it

53:34

marked as pending 24 hours ahead

53:36

of time. So you can just

53:39

tap play in your podcast and

53:41

have a choice. It's pretty cool.

53:43

It's pretty cool. It's pretty nice.

53:45

That mumbleroom gets fired up too.

53:47

It's crank and now that we're

53:49

going, and details for that are

53:52

Jupiter broadcasting.com.com. Thanks so much for

53:54

joining us. See you next Sunday.

54:40

You