Hello, friends, and welcome back to your weekly Linux talk show. My name is Chris.
My name is Wes.
And my name is Brent.
Hello, gentlemen. Coming up on the show today, Linux 6.14 is landing, and there are big improvements in there for gaming, file systems, and a lot more. But you might have noticed there's a Windows feature that's sneaking into Linux kernel. We'll dig into what that's all about. We'll round out the show with some great boosts and feedback, some picks, and more. So before we get into all of that, let's say hello to our virtual lugtime appropriate greetings, Mumble Room. Hello, hello.
Hey, Chris. Hey, Wes. And hello, Brent.
Hi. Hello. Got a handful of you in the on-air and a handful of you in the quiet listening. The Mumble Room, of course, is always going when we get our stream up and running on a Sunday morning. And a big good morning to our friends at Tailscale. Tailscale.com slash unplugged. Tailscale is the easiest way to connect your devices and services to each other, wherever they are. It's modern networking the way you really, really want it, protected by... Great for companies, great for self-hosters.
It's secure remote access to your production systems, your mobile systems. You can combine complex network infrastructures into a flat mesh network. Maybe you got multiple VPSs, you got some on-premises stuff. All of that can come together, and it's really, really fast. It's easy to deploy. It's zero config. It is a no fuss VPN. And then there are things like ACLs and additional feature sets that you can build on top of like integrated in with your company's authentication infrastructure.
So when you really need to take it to the next level, tail scales there and it works like nothing else. But the real magic is if you go to tailscale.com slash unplug, they're going to give it to you for free for up to 100 devices and three users, no credit card required, not a trial. That's the starting plan. And like me, you'll start using it and it will change your network. game. It is really something else to now have everything on my own tail net and nothing on the public internet.
All my devices sync over my tail net. We access and manage all of the Jupyter broadcasting infrastructure over our tail net. And even some of our public traffic is actually tunneled through our tail net to infrastructure on the back end and then displayed to the public. It really is very powerful. That's why thousands of companies like Instacart, Hugging Face, and Duolingo have switched to using Tailscale. And so many in our audience love it too.
So try it for yourself or for your business, but support the show and get started by going to talescale.com slash unplugged. Go there, get it for free on 100 devices. Talescale.com slash unplugged. Brace yourself. We have 25 days until LinuxFest Northwest, if you can believe it. Now, of course, the schedule is live. We'll have a link in the show notes. Linux Unplugged is going to be live at 10 a.m. in HC108 on Sunday.
Live from LinuxFest. And I just heard from Noah that he'll be joining us from the Ask Noah program. And also Wes has a talk after the live show in the same area, maybe not the same room but the same building.
Definitely the same conference.
Same conference, same building I think but maybe upstairs Linux powered payments running the Lightning Node with Nix Bitcoin at 1.30pm.
Yeah trying to just share some of the stuff we found useful if you do want to engage with the Lightning Network probably the best way to run a node at least if you want full control is something like Nix Bitcoin powered by Nix and Nix OS Also.
You can just come say hi to us because probably a lot of us will be there. We're going to have to figure that out because there's no way we are done with the show by 1.30. I mean, we might be done streaming. There's no way we're done with the show by 1.30.
Yeah, maybe it publishes a bit later. It publishes after my talk.
I don't even know what we're doing. It's going to be crazy, but we've got 25 days to figure it out. And we'd love to see it, LinuxFest Northwest. I'm feeling like it's going to be a good weather year. I'm hoping. We'll see.
Yeah, we've earned that after last year.
Is it like a seven-year cycle?
I think it's a two-year cycle. I don't know. We'll see. This will be the trendsetter.
Yeah, if Brent doesn't do his cloud seeding like last year.
Stupid lobes. Weather lobes.
Hey, it worked well. I think we got good seed funds from that one.
Well, speaking of funds, I just want to take a moment up front in the show and thank our members and our boosters who have been supporting the show. The show really is trying to do something unique in this space. We're trying to make the highest quality product we know how. We aspire to make a magazine's worth of content in nearly every episode, every single week, for free.
And we're doing it for a market, the Linux listener base, that historically has failed to make this kind of thing sustainable via sponsorships alone. And of course, we seem to prefer to do things the hard way. So we're not leaning into where the money is these days. YouTube, clickbait drama topics. We're using an open platform with an open standard that doesn't have an algo that pushes us in front of you.
And one of the things we would like to do in the next couple of months is raise some funds to buy some gear, some headsets that we can use on our LinuxFest Northwest coverage and the BangBus trip to rescue Brent's new BangBus, which we expect to be doing live shows from the van on the road. And you know our standards are fairly high, and I've been researching this for weeks.
I've asked around to people's opinions that I respect to see what they suggest, and I've landed on a set of headphone-microphone combos that I would like to buy for the boys. It's the Sennheiser HMD-26 dual ear broadcast headset. Now, these things are not cheap. They're $500 each, and they don't even come with their proprietary cable, which is also another $100.
So it's not that I come to this conclusion easily, and I have ordered one for myself already, just so I can get some experience and test it. You know, these are really in the sweet spot for a quality that I think we could accept on air that are under $1,000. And we should be able to EQ and process them. I'm going to talk to Editor Drew on his opinion about that. And we hope to make them sound decent. And I'd love for each of the boys to
be able to get one too. So it's around $1,300 we're going to raise. And I'm comfortable buying them with BoostSats or something like that. So if you would like to help support us getting that gear and getting us ready for the van trip and for LinuxFest Northwest, please consider boosting the next few weeks episodes. We do come to the audience from time to time with these kinds of things. Not always, but with fewer sponsors out there, this is kind of how we have to finance these types of things.
The show goes on if we don't reach our goal, but it is about trying to make the show as good as possible on the road. Not just from an audio quality standpoint, but I'm also looking for something that we can reasonably manage in a small backpack gear kind of thing. Something compact, on the go. And again, like when we're at a booth at LinuxFest, something where we don't have to have a bunch of microphone stands and all of that.
And if you think about it from a gear piece, it's a microphone, it's the stand, it's the cable, and then it's another set of headphones. And if you can slim all that down into one device, it makes it just a lot easier from an operational standpoint.
Yeah, investing in simplicity, which makes it cheaper in a variety of ways, less gear to bring around, less complex, and hopefully makes it easier for us to be able to do this at more places or more events.
Yeah, so if you'd like to, that's probably about 1.3 million sats. So if you'd like to boost in for the next few weeks and have a go to that, just note it in your boost, and we'd appreciate that. And of course, the members, your support keeps the show going there, and we really appreciate that as well. And the show will go on if we don't raise the funds, but I wanted to put it out there as it's a goal of ours, and we'll have more if it works out.
And thank you, everybody who does support the show, in either way you do, either through treasury, through time, or through your talent. We really appreciate it. Well, just as we're sitting here recording, Linux 6.14 is cooking, and it is a major release. It's 2025, it's got some barn burners, and this particular release is going to be in the next Fedora and Ubuntu releases, as well as a bunch of other distros. So it's an important release, and it's a banger release.
Also, should we just note at the top that it was the rare event where, in a way that didn't matter at all, Linus Torvalds released a day late, I'd love to have some good excuse for why I didn't do the 6.14 release yesterday on my regular Sunday afternoon release schedule. I'd like to say that some important last-minute thing came up and delayed things. But no, it's just pure incompetence.
I love it. After all this time.
Because absolutely nothing last-minute happened, and I was just cleaning up some unrelated things in order to be ready for the merge window. And in the process, just entirely forgot to cut the release.
That's all right.
Yeah, it's still, it turns out, a great kernel.
It really is. and we're going to get into some of our favorite improvements and updates. There's a lot more to this kernel, and we could do a three-part episode series on it. So please feel free to nerd out with the links in the show notes. We have, I think, links to some of the best coverage. But let's start with the one that's close to our heart. Wes is running it on his laptop right now. BcacheFS saw a major step towards stability in Linux 6.14.
Yeah, well, it was kind of the catch-up release, right? So there was a good, lots of good stuff in 6.12 and then 6.13 due to the code of conduct ruling, Kent could not contribute. So we didn't see any updates in 6.13. So now we're getting a lot of good stuff in 6.14. In particular, there is hopefully the last planned big, major, expensive, but still automatic and required on-disk format upgrade coming with 6.14.
So it just happens automatically in the background. But if you have like a giant array, that could be some work to do.
So you might not even notice. Yeah.
In the past, I have not for just running as my root of S.
Okay. So with that comes major scalability enhancements.
Yeah. And that's the thing, right? A lot of this stuff is as, you know, the file system gets used in actual scenarios and can't go through bug reports. You kind of find out, oh, it turns out we wanted to slightly tweak this structure or a lot of times it's adding points where they can keep more information or more statistics or metrics or like information or back pointers.
A lot of times where you kind of have more of the structure preserved to make it faster when you're trying to do some repair operation so that you don't have to like go scan as much data.
To that point, they say in here there's a dramatic speed up for FSCK.
Yeah.
And they tested it on. Do you see that in there?
Yeah, right. 10 petabyte file systems. They.
Have 10 petabyte file systems out there running bcachefs right now.
Yeah so i don't think my like 500 gig root fs is really gonna sweat that.
So good to know.
Yeah good to know and it's just if you do ever get that it's a good reminder right like if you're not willing to test on disk format upgrades uh with giant arrays then bcachefs is not yet for you which is totally fine it's still uh experimental but also this is good good signs i think of um you know signs of stability going forward they were hitting places where, Once you get to a point of not, at least not even expecting on disk format changes,
there still could be some if like a major issue needs it. But this is the first step towards it being more trustworthy for like actual serious data arrays.
Very much so. I should have looked more into this. I just noticed in my notes here, the default runtime has been set to self-healing enabled.
Yeah, there's been a lot of self-healing work going on, right? There was already ways that you could trigger that. We don't yet have a full scrub, but we do have self-healing.
Uh and coming in 6.15 it's already merged upstream i think in 6.15 we should see, full scrub support it's already also in the upstream userland utils but i think in a future release so again right you need you need the self-healing capability at all having it happen when it detects things on the fly and then from there you can have scrub that just sort of walks the whole file system to trigger that checking it's a bones release
maybe not one that you'll see a ton of changes from a user perspective but that's a sign of where we're at with the infrastructure it's.
It was interesting so like the moment the 614 window opens kent he does a poll that is just like here's syncing everything up and then here is here's all these improvements too so it's really fascinating just to read through his description of all of it and to see him just right back at it and i wouldn't be surprised if by the end of the year or before the end of the year, the experimental flag is removed.
Yeah, I like that as a 2025 prediction.
Could be.
A little low-key predict.
I can't remember if I actually made that in the actual predictions episode.
Do you have the lock it in sound somewhere?
All right, well, the file system that's actually shipping right now that a lot of folks are using is ButterFS. And there has been some really good improvements landing in ButterFS, including three new RAID 1 read balancing methods, which improves how reads are distributed across devices. So you have rotation... Which actively rotates and reads to keep all devices engaged. That's the preferred default going forward. You have latency, which optimizes for unstable or failing device transports.
And then there's dev ID, which gives you manual control over which device handles the reads, which I could definitely see workflows where that would be useful.
Yeah, absolutely. And maybe even just as like an operational thing for, you know, while you're having issues with the disk you're going to replace or something.
And of course, not just like changes to RAID 1, but also other improvements to ButterFS have landed, which includes support for uncached buffered I.O. And here's a bit from kernel newbies. They say this release adds support for optionally sending buffered I.O. Whose pages will be dropped from the page cache once the data is read or written. The reason for this feature are fast storage devices that can and do fill the RAM with too much page cache that will not be needed.
With this feature, it is possible for the read and write data and to drop it from the cache without facing the disadvantages and complexity of dealing with direct I.O. Man, I wish I had that problem. Where my disk I.O. was so fast that I was overfilling my RAM with stuff I didn't need.
But isn't it great that the file systems in the kernel are attempting to keep up with the latest and greatest hardware? Or at least, you know, a version ago of that?
Mm-hmm. Mm-hmm.
I think there's actually been some work generally around on cache buffer I.O. In the latest couple of kernels to just make that better.
Okay, one last file system thing. You had to get the file system things out of the way because these probably really impact the most of us, some of these features. And the one that really sees some love is actually your Fuse file systems. They're going to get a big performance boost via IOU ring.
So this is a fascinating story. A gentleman who works at DDN Storage, he had a talk at the 2023 Linux Storage Summit and also at BPF Summit, and he talked about discovering an odd issue when really taxing his MVME disk. He says, I was looking at an MVME driver and was wondering why it was able to avoid the bottlenecks, why it wasn't or was, I'm not sure, able to avoid the bottlenecks I was seeing. Oh, it was. It was because it was using Ioring, Uring, but it was in the wrong direction, he says.
It was going from user space to the kernel, but what Fuse really needed was for it to go the other way. Can you help me make sense of that, Wes?
Yeah, I think they were investigating issues with Fuse and noticed that the NVMe driver wasn't having the same problems.
Okay.
And the way it was using IOU ring, which is a new I.O. Mechanism in the kernel that makes it really fast and cheap to be able to have the kernel, you know like fill a section of memory with data say coming off of a network driver or something i.
Think our last kernel deep dive episode we did go into that a little.
We did yeah and then user space can read from that and then avoids a whole bunch of context switching between the two which is always expensive uh so now we're having some of these techniques being able to be applied, to fuse to speed up basically whatever you're doing with fuse which is a whole bunch of stuff because Fuse is, as a reminder, file systems in user space.
So you can have a user space process managing things, but the kernel and other programs, crucially, treat it like it's a kernel file system, right? And this is where you can have stuff like SSHFS. And there's a whole bunch of very, very cool Fuse file systems.
But because you have to have, right, you're talking the kernel API for file systems, but then the kernel has to go and talk to the user space thing managing the Fuse file system, that's a ripe area where you need a whole bunch of data exchange that can be expensive if you're doing context switches but with iou ring, It can be faster.
You avoid some of that. One of the interesting insights to this is this could have broad ramifications for all types of Fuse file systems. We use one that's essentially a WebDAV Fuse file system for managing the files on our next cloud. And this could be really nice to see something we've considered generally a little bit slower than a real native file system actually get a nice little boost here.
And we're seeing IOU ring, which we talked about before, really kind of come in and change the game, which is fascinating to see it kind of get rolled out in different areas. And I think the discovery here is probably going to be applicable in other areas. Like Fuse is where they're starting, but I bet you this idea ends up kind of spreading out in other places.
Yeah, usually it does.
Now, that is all the file system stuff. Great job, everybody. There's more in there, of course, including there's a fascinating history with Fuse. There has been a couple of goes at replacing Fuse to try to solve some of these problems. And a couple of different attempts at like fuse two and a different type of fuse all together to specifically try to solve these problems they never really got adoption and then we kind of came along and solved it at the kernel level i.
Have good news and bad news we're actually not done with file systems.
Oh we're not no okay because.
Xfs gained enhanced real-time device support.
Oh you're right you're right you're right xfs got some love this cycle Yeah, and it's, you know, I think this is particularly worth noting just because last year we covered the XFS maintainer swap, or maybe it was a little bit longer than that. And that maintainer that stepped down was a long-time really trusted maintainer. And the person coming in, we weren't quite sure what direction they would take the project. And what we've seen is just impressive development after impressive development.
So the XFS developer Derek Wong explained on some of the real-time reverse mapping and reeling support that's landing in this patch series. He says, Christoph and I have been working on getting the long-delayed port of reverse mapping and reflink to the real-time device into a manageable shape. With these changes, the real-time volume finally reaches the feature parity with the data device.
This is the base for building more functionality into XFS, such as the zone storage support that Christoph posted about last week.
I think it continues the story of, I mean, just XFS being a great, you know, if you don't need the kind of things that things like Butterfess and ZFS and Bcache address, XFS is just a great file system. And they're really pressing to keep up with, in the ways that its architecture supports, It's an impressive subset of modern file system functionality.
Okay, now that is all of the file system stuff. And this next category is, holy crap, there's a lot of AMD stuff landing. But first, we should acknowledge the very important addition, at the kernel level, support for the Microsoft Copilot key on modern laptops.
Hey, did you know that was the thing?
There you go, Brent. And so if you get a framework now and, you know, some of those new frameworks come with a copilot key, don't worry, it's going to work.
So is this in addition to the existing Windows key?
Yeah, it's on the other side now. You know, so you got the on the other side of the keyboard on the right side, I guess. I don't know what it does on Linux. Maybe you can just map it to something.
Maybe it'll trigger your home assistant for you.
Sure.
I think it should be your compose key. But if you don't know what that is, then maybe it's not useful for you.
All right. Well, let's talk about the plethora of AMD GPU driver updates. New DRM Panic supports. You get the fancy blue screen now with AMD devices.
Yeah, we talked about that a whole bunch. But as a reminder, to take advantage of some of that new functionality, there needed to be per graphics driver support. So now we're finally seeing that on the AMD side. It was already in a bunch of like the base ones and the Intel side.
We're seeing early support LAN for the RX90 Zero Series GPUs and beyond. And the big one is the AMD P-State driver update, which makes big improvements for laptop power management and efficiency for AMD systems. But there's an interesting twist in here.
The maintainer of the Linux power management subsystem, is an intel employee uh and you know this release was just a lot of amd code and so the intel guy, is the one that is you know actually going through this and submitting it upstream and this is what he wrote for the patch for 614 he writes the majority of changes are cpu frequency updates which are dominated by amdp state driver changes like in the previous cycle Moreover,
changes related to AMD P-State are also the majority of the CPU power utility updates. Indeed, there are some pieces of new hardware support, like the addition of the Clearwater Forest processors to support the Intel idle. And there are some other improvements for other Intel chips in there. But the vast majority of the updates for this 613 and 614 kernel have been from AMD. And this Intel maintainer is doing his job, doing a great job,
is submitting all of this upstream. And I just find that to be a beautiful irony of free software development.
It also makes me think, I mean, A, there's probably a lot of nice benefits for these employees of diverse companies working together on some of this stuff. Maybe there's insights or, you know, learnings from each other. But also, like, in a way, just open source sort of erodes some of the, artificial's too strong of a word, but I'm going to say artificial boundaries imposed by, like, the corporate structure of things.
Like, here are a bunch of CPU nerds who are now nerding out together to make all of the CPUs work better.
Seeing what common problems they're solving and also both actively contributing to these release cycles, Intel and AMD, and both actively preemptively contributing bits before the products are even in the marketplace, which is so critical when you think about the time delay. So Linux 6.14 is going to be in the most recent Fedora and Ubuntu, but it'll be quite a while before 6.14 is in any of the LTSs or RHEL.
So when they get that groundwork in there now, people that take advantage of the in-between distribution releases get to use the drivers, but it takes years before it's in the enterprise-grade stuff.
And so it's really incumbent upon them to try to get this stuff in as early as possible and that's why before they've even finished all of the driver work they're getting the primitives in now as early as they can so that way the next cycle they can just get you know a little bit more in there and then, hopefully get it out to end users as fast as possible we'll see but that's the that's the interesting game that they both have to play and both AMD and Intel
just by watching this do seem very committed to that process they're getting it done. 1Password.com slash unplugged. That's the number 1Password.com and then unplugged all lowercase. You want to go there to support the show and to learn more about 1Password. Here's how I like to think about this. I have a question for you. Do your end users always, and I mean to always use the IT approved applications, services, and devices?
I don't know if any users do. I think there's just so much out there now. There's so much getting pushed at them from a consumer level from a this is going to change how you work level. So how do you keep your company's data safe when it's sitting in all these unmanaged apps or devices? That's where 1Password comes in. Again, it's 1Password.com slash unplug. They have the answer to this question, extended access management.
1Password extended access management helps you secure every sign-in for every app on every device because it solves the problems that traditional IAMs and MDMs just weren't built to touch. You secure every app, every device, and every identity, even the unmanaged ones, like when I was a contractor coming in. You go to 1Password.com slash unplugged, check it out. This really does change the game. Now, 1Password's famous, of course, for their beautiful password management.
Now they've taken it up to the next level, from that award-winning product to extended access management. They're really bringing everything they've learned, including their best practices. You know, they undergo regular third-party audits. They have the industry's largest bug bounty. And it's really great to say that 1Password exceeds those standards that are just set by other authorities and whatever the letter of the law might be. They go above and beyond. I love to see that.
It's really the first security solution that brings all these unmanaged devices and apps and identities under your control. That's the big deal. That's the big shift. So go check it out and support the show. Go make sure every app, every device, every identity is secure, even the unmanaged ones. Go to onepassword.com slash unplugged. That's onepassword.com slash unplugged.
Now, as with many previous kernels, we do see a little bit of gaming love in Linux 6.14 as well. Looks like the Linux 6.14 kernel will officially include the completed NTSync driver, which is a major improvement for Linux gamers using Wine and Proton, especially via Steam Play or SteamOS.
An NTSync feature in our Linux kernel?
Yeah, that's right. It's kind of neat, right? Now we have a specific synchronization primitive enabled inside the kernel specifically to support compatibility with Windows applications.
Very fascinating. They say here, The driver enhances the emulation of Windows NT synchronization primitives, which are critical for accurately replicating Windows behavior and improving performance in many Windows games running on Linux. Though NTSync was introduced in Linux 13, it was initially incomplete and not functional. 6.14 marks its full implementation and usability. As for the motivation behind the work, well, it's clear in the benchmarks.
The gains vary widely depending on the application and the user's hardware, but for some games, anti-synchronization is not a bottleneck, and so no change will be noticed. Some other games, though, improvements 50% to even 150% are normal.
Yeah, okay, so there is some nuance here as well. Right now, this is specifically helpful for, or going to be, for Upstream Wine, and it is the newest and shiniest and most compatible with Windows implementation, but there's been earlier attempts, specifically eSync and fSync, which have basically been gaming-focused mechanisms to enable improvements to emulating games specifically.
And since I think like kernel 514, we've had Futex 2 in the kernel, which Proton is patched Upstream Wine to take advantage of. But I think Upstream Wine does not, partially because it was kind of gaming focused and there were some thoughts that it was kind of a hack and didn't really like the specific interface being created for Wine and Proton.
And it doesn't solve it for all games, like solves it for some of the games, which is why some of these games won't see an improvement.
And the root of this is, okay, so in the Windows world, you've got the Windows NT kernel. It's pretty different than a Unix-y, Linux-style kernel influenced by stuff like VMS. And there's a whole bunch of objects, and you have handles to these objects, which if you've ever done Windows programming, you've seen handles freaking everywhere, right? And if you're a game or just a program, often you need to wait for things, right?
So like in a game, it might be wait to be told that like the frame I wanted to draw is done painting or copying to the GPU memory or whatever it is. Or I need to wait for a keyboard or a mouse event for the user to do my next thing. Or maybe a network package shows up telling you what the enemy participant did and how to update their game state.
And so there's oftentimes where you've done your immediate work and so you want to say like, hey, wake me up when any of these things happens and then based on what it is, I know what to do next to advance the process.
And due to the way Windows is designed, they've got some pretty nice APIs for this, including wait for multiple objects so you can kind of just hand it a whole smattering of things that you want to wait for regardless of it's like a file handle or a network thing or some status update to a particular process or thread, and Windows will set that up for you and while Linux has things like ePoll and other mechanisms to specifically wait for things like file handles there's various.
Equivalencies of course because right ultimately windows and linux make your computer work in roughly the same way but the apis are different enough that like originally wine had to have kind of like a helper thread or process that could sort of translate between the windows system call in user space and then implement it under the hood with the right kernel features and as you can imagine like that's not going to perform super great some of it was fine
for some things it did work but like it was just going to hit bottlenecks especially with games and so that's where we've had multiple attempts with esync and then fsync which has been the most successful and widely used with proton but now finally we're getting ntsync which is it prioritizes compatibility so it's like the best translation in the kernel for this style of synchronization primitive but it also looks like it has some quite nice performance
gains so especially for upstream wine you're going to see those big numbers that you're seeing reported if you're doing proton it there's still it has not yet adapted in theory in the future they'll probably switch to ntsync and there'll be some slight performance improvements and compatibility improvements.
And like from a practical standpoint for end users they're going to get essentially a new device a new virtual device right like a slash dev slash ntsync device.
Yeah and that's uh the mechanism for how things can sort of go request that and wait on it from the kernel yeah.
And each each time each process opens up the device they get a unique instance but it's shared across threads which is so they get like their own id but there's some shared space there too.
Yeah and that's usually the case right so for processes they have different memory space and threads get to share so that way you can have multiple threads that are doing things but they can all be they can like share waiting or locks or use that use that for a shared synchronization primitive to like communicate between threads.
Sort of like the fuse iou ring improvements we talked about earlier in the show This NTSync that is definitely targeted at gaming right now seems like there would be other complex Windows applications that would benefit this. From this. And so while it is initially useful for gaming, I could see in a year or two, we've discovered some other application that didn't work before is now working because of this too.
Yeah. Right. In the big picture, it's just like a big, nice step up for our Linux's ability to be compatible with Windows applications through Wine. It's just cool because, I mean, there's already, again, there's already been some of these kinds of efforts, but just the idea that the kernel is being adapted to support this non-emulation user space thing of wine that somehow seems to work surprisingly well.
That's just that takes a lot of coordination and planning and years of effort to get across the finish line. And like, you know, folks had to make the case that, hey, you know, people are trying to run programs not even intended for the operating system you designed, but can you make this change, please?
That we're going to sell on this proprietary hardware.
Yeah, right.
But it does show you that if you've got good code that works and you're willing to do the work and you show up to the Linux community, kernel community with you know proof of receipts they're probably going to take it uh they're they're they're really pretty much down to take most things if you're doing the maintenance you're doing the the legwork you know like yeah.
And you have you can point to real you know use cases or people who want it.
This is how samsung got a samba server in the kernel west jeff.
Makes a good point too and this is definitely true right through the history of computing games have often been one of the best ways.
To sort of.
Like push the limits for how to make these things happen.
I agree i agree and And that's why I think we'll fix it for gaming. And then we'll discover some other complicated application.
Just in time so Brent can switch back to Lightroom.
Now, we were thinking, and we'd like to hear your thoughts, about maybe doing a gaming episode next week. Nothing crazy. Maybe just one or two games, and then also a bit of a hardware update. So even if you're not a gaming person, there'll be some hardware talk in there, too. So if you have thoughts on gaming on Linux, a game you're loving right now, particularly those that support co-op, send them in to us. Boost them in, or go to linuxunplugged.com slash contact and let us know.
Because we've been thinking about doing a gaming episode literally for about two years. So maybe we'll do it next week. We've got a couple of things. I got a new device in-house. Remind me to show it to you after the episode. It's pretty slick. I want to talk about that next week. So that's been on our mind. It's been a long time. The gaming episodes tend not to be super popular. So if you don't want us to do it, you can let us know too. I don't know. It's like, it's always hit and miss.
You know, if we really got ahead of it, we could do like a gaming meetup. If we were playing a game, people could join.
Well, okay. How about this is a goal. We come up with a co-op LAN style game that we could try that we could use at Linux Fest. And we demo it. So we would love some suggestions for that because I literally don't know.
Maybe producer Jeff could help.
The other thing to kind of just sort of think on for Linux Fest is where we, oh, maybe in the room where we do luck. Maybe we could do it in the room. That could be interesting. Anyways, something that's been on our mind, new hardware and all of that. So let us know your thoughts there. And I don't think anything's ready yet to try out this new NTSync thing. I wish we had an opportunity to test it with a game. But when one does, I'm already going to be on 614, so I'll be ready.
Did you know your ad could be right here, right now? I could be talking about your company. Your Linux Unplugged podcast is looking to book a sponsor, and we're running a special for the second half of Q2. Now, you know we're kind of picky who we work with, which has limited us, but if you're listening to this show, well, that's a good start.
Chances are you could be a great fit. So if you want to reach the world's largest Linux audience on a show that never misses a week and has been going for nearly 12 years strong, Reach out to me, chris at jupiterbroadcasting.com. I'd love to work out a deal with somebody who's in the community that has a business or a product or a service that they'd like to reach. We'll make a pretty good deal. And I think it could be neat. I don't know. Chris at jupiterbroadcasting.com.
That's all you need to know.
Well, we got a special little piece of mail here from Grant who asks a little for some advice, let's say. It says, greetings, Chris, Wes, and Brent. I keep hearing you guys talk about this Nix thing, so I'm finally going for it. I was wondering if you have any tips for Nix OS configs on a laptop specifically. From what I research, so far most of it has to do with power optimization while on battery.
Since using Linux on a laptop is literally Linux Unplugged, maybe another question is what are some community recommendations for Linux on a laptop? Love to hear what the best Linux community around, Linux Unplugged, has to say about this. And thanks in advance from Grant.
That's a good question, Grant. I feel like I'm going to be in the laptop market towards the end of the year. It has been interesting living on Asahi on the MacBook and very limiting. And I eventually want to get something, but I want a very nice laptop experience. Something real smooth. Framework's a contender. Of course, I'm going to look at System76. I'm going to be looking at ThinkPads too. But I think Grant's asking two questions here. Number one, what is a great distro hardware combo?
And, like, what should he look into for Nix to pre-set up his Nix for the best chances on that hardware? Like, I know Brent has looked at a config specifically for framework laptops.
Yeah, well, there's the NixOS hardware community repo. So that's one place you can pull in stuff.
That's the thing.
You know, because, right, a lot of it is going to be making sure you're fully enabled, whatever settings or kernel parameters or driver options for your specific rig because after that honestly you know the linux world's gotten a little bit better i've been pretty pleased with the power management options at least in things like plasma and gnome very.
True and even in the last couple of releases there's been some nice improvements.
So there are you know various uh demons and things like power top and tools to monitor but for the most part i'm not currently fussing with those as much as i used to yeah.
I just leave mine in balanced mode for the most part and then power saver when I need it and boost when I need it. I just, what I'm looking for is, it's Brent, he has the framework with, I think it's the 12th gen MOBO, right Brent?
Yes, it is.
And you have this problem that I witness all the time when you're around where you close up the laptop and then you come back to it a couple hours later and it's hot.
Yeah, I have to say like maybe two out of 10 times. I guess that's one out of five. I will pick my laptop out of my bag and it'll be roasting. The rest of the time it kind of behaves as it should. I've been trying to figure out what is the combination of what I'm doing that causes that to happen. Now, Jeff, I know you've struggled with this kind of thing before, and you probably have a bunch of advice for me. But it has, I would say, gotten better.
So there's constantly tweaks happening where Wes suggested in the Nixos hardware repo to try to solve some of these common problems. But it seems, I don't know, suspend is just in not so great a place as it was like, let's say, on an old ThinkPad laptop from back in the day. So it is something that's getting better, and there are some tweaks you can do, but I got to say, like, as a warning, part of the reality is sometimes it just doesn't go so hot.
PJ, I'd like to hear your thoughts, but I would think if he's on a pretty modern kernel and he's using a 12th gen system, which has been out for a few years, shouldn't that be best? And it's all Intel. Shouldn't that be best case scenario?
I think the framework specifically does have S3 sleep and that's one of the biggest problems that we're seeing with modern laptops is Intel and AMD have kind of pushed away from that and it's up to the motherboard, really the upstream, you know, laptop maker to put it back in and everything defaults to S5 sleep, which doesn't fully shut down.
With the frameworks i've seen that you need to go into the bios and turn off usb power because all the little dongles will just sit there and suck suck power down when it's trying to sleep even even if it is an s3 so you might be seeing that but it's a problem my lenovo does it my last hp did it it's all s5 these things aren't actually going to suspend a ram they're just basically shutting the screen off tiny.
On the other end though you're on a framework using fedora.
Yes um fedora is just switched from using ppb ppd to toondy and it's really nice because you can nest profiles so you could have the uh default balancer laptop profile and then make your own little changes on top.
Yeah, okay. So that's, you know, maybe the advantage to having a distribution, tune that stuff for you. Something Brent could look into, I suppose.
On NixOS, I see there's a power management option, which has stuff like CPU frequencies in here, power top, you know, all kinds of stuff. Power profiles, daemon's an option as well. So there's various things in NixOS too.
Boy, I'd love a great experience on my next laptop. Something really, you know, hassle-free, works with distros, just great, has some decent little performance for maybe some basic gaming and sleeps.
I was going to say, in general for laptops, I think we've all been pretty happy. I mean, it's not laptop-specific, but happy with things like ZRAMP.
Yeah, yeah. I think there are also compromises to be made. If I could get it to boot super fast, I could possibly live without sleep, I suppose. I don't love that, but that could always be a route I could explore if it meant I could have a fantastic laptop that I really love, that might be a compromise I could make.
By fantastic, you mean loud fans, right?
Yeah, no, I do not, sir. I do not. Yes, it is, and we start with our baller booster this week, who is the tech geek with 45,000 sats. And he just sends along, hey, JB team, keep up the good work. Well, thank you for the value.
Turd Ferguson boosts in with 41,088 sats. Oh, this is a triple turd.
Oh, gosh.
Boosting the 2i challenge.
Uh-oh envy.
Top it's not just for nvidia did you know intel amd and even at least sorta the m1 and m2 gpus are supported.
Envy top huh it sounds like it's for nvidia okay.
Another trip here brent's traveling uh we could just if it was a website is brent.
Traveling it'd just be yes yeah uh.
Pack trippy the 2i love child of ping and trace route.
Sounds like you could have used.
This just the other day actually.
Do we know about trippy i.
Don't know if we do.
It's a 2e huh let's go look this up trippy uh a network diagnostic tool oh my god it's a riff it's a riff on the clippy logo too which is funny oh i like this i like this a lot brent i think you could use that i could use this actually what's.
It written in.
The question oh look at that did you know this already no.
I just wanted to know.
Yeah 99.6 rest it's.
Not 100 but i'll take it.
Yeah it's pretty close so that's trippy all right that's good um now do we have you said there was three of them.
Yes i did there's one more worried you might get bored during the two-week challenge try tty solitaire now with mouse support.
Tty solitaire all right solid.
You know they told you your flight would have wi-fi it doesn't but you already downloaded tty solitaire turd.
Ferguson coming in with some good two-y picks. I'm feeling the two-y energy. I'm feeling like this challenge is going to happen, boys. I'm feeling it.
Are there any two-y games we can combine these two things?
Well, there's the Solitaire, I suppose. I suppose.
You know, on my flight to this here continent, there was someone the entire flight that played Solitaire, and I could not believe it, but maybe...
With cards?
No, no, no, no, no. On their phone.
My wife could do it with cards. She brings a deck of cards in her purse.
It's a great idea to have.
We were out and about last night, and we had a half hour to kill with the shuttle, so she just busts out the cards. She just goes to it. Great battery life.
Wakangaroo Paradox came in with 30,000 sats across three boosts. Great coverage of the rust drama recently i really appreciated the objective breakdown of events and your take on it also loved the deep dive sorry brent he says i think oh.
You know brent you know hold on brent's.
Got a reputation.
No i was trying to speed right past that one so you wouldn't have a chance to play it, shakes fist at you both he continues here i think you could sprinkle these once in a while and that could be enjoyable for people who aren't super technical good way to learn things too also plus one for the tui challenge i'm all for spending as much time as possible in the terminal versus the browser i think a tui challenge with a point penalty for
mouse usage could be a fun spin oh love the planet nix coverage could not justify flying from europe to go for a couple days over 1,000 euros total for that trip. Can we expect you guys to make the trip to NixCon EU, maybe, in the coming years? These usually take place October-ish in Germany or around there. I imagine the price would be an issue, but, you know, the beer here is worth it.
Yes.
The truthful answer is, I imagine when my kids are a bit older, I'll probably be traveling on longer trips, where I could go somewhere I could spend like two weeks, or a week.
You know that would be really great but that'll probably still be a couple of years maybe every now and then you know i never want to say never you never know right maybe wes could talk me into something i think you guys can be pretty convincing when you put your minds to it thank you for the boost congo it's good to hear from you the musa came in with 5 000 sats firstly i finally got around to say in my own bitcoin note with nick's bitcoin hey that's great to hear and i'm using AlbiHub,
thanks to Wes and the community for providing a good starting point to do so. I am personally excited for GNOME 48, as controlling the Orca screen reader under Waylon has taken a big step forward. I've wanted to get XOR off for ages now. Come May and Nix2505, I may just be able to do that. That's great to hear.
Yeah, that's rad.
Yeah, there is some nice changes there. And Wes, look at you, helping people get AlbiHub up and going.
Yeah, and props to folks like the muso and draca who uh have taken the leads on getting the module going i got a base flake happening but also i've been negligent so this is a great reminder uh i'll go try to see if i can spend some time in that repo adversary 17 boosts in with 5 000 sats, The Dynamic We Don't Even Knows are pretty funny.
Oh, yeah. The members version somehow got set up with Dynamic AI inserted ads for all kinds of crazy things.
We got a good kick out of them the whole show.
Thank you.
Of course, the only proper song to play on the Clinton sexy saxophone is Careless Whisper.
That was true. Yeah, that would have been, yeah, that's true. Nice to hear from you adversaries. Thank you.
Well, Brooke Loves You sent in 3,535 sets. Hey jb have y'all tried kd plasma mobile yet i'm daily driving that thing on my minis forum v3 with fedora and it's pretty awesome i recommend giving linux on a tablet a try if y'all haven't in a while or even at all keep up the good work and thanks for making my work travel more enjoyable this.
Is a great suggestion brooke and you're kind of reading my mind, this Friday I was browsing Amazon looking to see if there's like quote-unquote Linux x86 tablets or even Linux ARM tablets.
You know, I do this thing where I have tablets around the home and I want to have more in the studio to like control home assist in different areas and why have Android on there when I could have Linux on there and it would give me an excuse to play with Plasma Mobile properly which I've only ever done in emulators.
And it's probably more realistic. I was going to say we pressured Brent because it seems like his kind of thing.
Yeah, yeah, you're right.
If you're already doing it. Or we get double cores.
Well, I could take my learnings and use that to pressure him.
Or Flash's stuff and set them up.
With me, if you get me in right, you know, the right mode, I'll do it in a Flash. But with Brent, it's going to take like two years, you know, before Brent gets an Android tablet that runs Linux.
Well, you know, only one of us carries two devices in their pockets. So I think, you know, having a backup built in, you're more suited for this thing.
Thank you for the boost, Brooke. Appreciate it. Ambient Noise comes in with 3,930 sats. And they say, sorry, this took so long, but here is my write-up on how to control the front gate via MeshTastic and MQTT. Oh, yes.
Awesome.
This was one of the things we wondered would be, could you do IoT control with MeshTastic? He sends us his GitHub repo, which we will include in the show notes, and it's instructions to create an MQTT controlled input and output board that runs its own logic on low power radio instances. And you can use it to control your gate.
Neat.
That is so...
A whole bunch of comments in here too.
I'm saving that. I am saving that. That was really great. Thank you for sending that in, AmbientNoise. And well done. Well done. Thank you, everybody who participated, including you SAT streamers. 25 of you stream SATs. We stacked a humble 22,923 SATs. Not too bad for those of you just sitting back and enjoying the show. Thank you very much for doing that. Then we combined that with all the people that generously boosted in with the message. We read the ones above the 2,000-sat cutoff line.
And we stacked 156,476 sats. Not our strongest week. I'm not going to say it's a total loss, though. But, you know, it could be better. It could be better, especially with some of our ambitious plans we have coming up for the rest of the year. So if you'd like to get in on the fun, there's a few ways to do it. But you've got to get some sats. Strike is a great way to do it. in over 110 countries, River in the US, Bitcoin Well in Canada.
Get those sats, send them over to Lightning to a podcasting 2.0 app like Fountain. The workflows also in Fountain are really simple too now. They're getting easier and easier to just do it within all one app. And it just keeps getting better. You can send your message in and we'll read it on the show above 2,000 sats. Thank you everybody who supports this episode and all 31 of you who boosted in and supported episode 608 of your Unplugged program. It really does mean a lot to us.
We appreciate it. And of course, Big shout out to our members who keep us going, too. You are all fantastic and much appreciated. Well, moving right along, gentlemen, moving right along, we have a pick that may help some of you break your grammarly habit.
Oh.
Have you wondered what are you going to do to get over your grammarly habit?
Finally, take back your freedom from big grammar.
Yep. Well, that's where Eloquent comes in. It runs as a local service in the background to do local offline, I guess, proofreading. It uses a language tool server, and it's available as a backend instance to Firefox, LibreOffice, and others. It is GPL 3. It does also have a little standalone desktop GUI where you can paste your text in and then it gives you a Grammarly-esque pop-up to do spelling.
Neat.
No, if people have never used Grammarly, they have no idea what I'm talking about, but it's sort of like grammar and spelling and sentence structure all in one.
And it's a good service, but to do this, they're reading everything you write and they want to integrate into everything constantly and the fact that you could actually run this as a back-end instance for firefox which i have not done well it looks really great like that is the perfect combination for me um and it is also available as a flat hub.
Well and this backing tool is cool language tool open source proofreading software.
Yeah english.
Spanish french german portuguese polish dutch and more than 20 other languages.
Yeah.
Finds many errors that a simple spell checker cannot detect.
There's a collection of tooling out there right now, like this language tool, and Piper from the Homosystem folks in Whisper, which do text-to-speech and speech-to-text, and then they have their Wyoming protocol, which lets it communicate. All of it is open source, all of it running on Linux, and we're not doing anything with it. It's really weird.
Maybe that's some show projects we need to get ahead of.
You can actually even train your own voices with it. You can have it model on your own voice and generate stuff. It's just right there for the taking. And some project just needs to integrate it into some sort of application or something. I mean, it's very useful in Home Assistant.
Right, of course.
But Piper alone and Whisper alone just feels like there's a lot of potential there. Anyways, today we're talking about Eloquent, and we'll have a link to that in the show notes should you want to check it out.
I also see it looks like there's some stuff that does makes like a language or an LSP implementation on top of language tool so you can get this kind of stuff in your VS code too baby.
And don't forget we're trying to raise some funds for headset for the boys so if you got any tips for gaming or games you love or co-op games something we can do at Linux Fest do boost in and support the show with a contribution towards the headsets and a message we could read we really do appreciate that and then, we'd love to see you at Linux Fest in just 25 days.
Link to the schedule and all of that is in the show notes as well and you know where you find that get ready for this linuxunplugged.com slash 608 that's all you need to know linuxunplugged.com slash 608 and of course we love it when you join us live it just gives it a vibe and we make it a tuesday on a sunday it begins at 10 a.m pacific of course we have it in your local time at jupiterbroadcasting.com slash calendar.
If you're enjoying a podcasting 2.0 app, I try to get it marked as pending 24 hours ahead of time. So you can just tap play in your podcasting app of choice. It's pretty cool. It's pretty nice. That mumble room gets fired up too. It's cranking now that we're going and details for that are jupiterbroadcasting.com slash mumble. Last but not least, if you're looking for a little something extra, go check out The Launch. It's on the JB site and weeklylaunch.rocks. It's a brand new show.
Thanks so much for joining us. See you next Sunday.