Untitled Linux Show 186: Accidental Honeypot

Summary

Episode Summary: Untitled Linux Show 186 – "Accidental Honeypot"

Release Date: January 19, 2025

In Episode 186 of the Untitled Linux Show, host Jonathan teams up with co-hosts Jeff and Ken to delve into a myriad of topics essential to the Linux and open-source community. Despite the absence of their regular co-host Rob due to illness, the trio delivers an informative and engaging discussion covering security vulnerabilities, desktop environment updates, support for legacy software, and exciting developments in hardware standards.

---

1. Rsync Vulnerabilities and Security Implications

The episode kicks off with an in-depth analysis of critical vulnerabilities discovered in rsync, a widely-used file synchronization tool.

• Ken highlights that rsync version 3.4 addresses six significant vulnerabilities present in versions 3.3 and earlier. Among these, the most alarming is a heap-based buffer overflow vulnerability in the rsync daemon, which could allow attackers to execute arbitrary code on affected servers.

• Jonathan elaborates on the severity, stating, “[06:37] Jonathan: That's real bad.” He references his security column on Hackaday, emphasizing that if rsync 3.3.0 or earlier is exposed to the Internet without proper safeguards, it poses a high-risk zero-click exploit opportunity. With a CVSS score of 9.8, the vulnerability is nearly as catastrophic as possible.

• The discussion further explores mitigation strategies, where Jonathan advises, “[02:24] “…for the vast majority of us, it’s either turn it off or update it. And that kind of needs to happen yesterday. It’s bad.”

This segment serves as a crucial warning for administrators relying on rsync, urging immediate updates to safeguard against potential exploits.

---

2. Plasma 6.3 Beta and Upcoming Features

Shifting focus to desktop environments, Jeff provides a comprehensive update on Plasma 6.3, the latest iteration of KDE’s desktop.

• Jeff shares, “[17:16] ...the 6.3 kernel is expected to be released. And there are some really interesting things in there...” He discusses the reduction of high-priority bugs from three to one, improvements in UI scaling, Night Light color accuracy, and enhanced keyboard backlighting controls.

• Jonathan adds his excitement about the upcoming features, mentioning patches related to HDR support in Firefox, which, although not immediately noticeable, signify substantial advancements in display technologies.

The conversation underscores the ongoing refinements in Plasma, promising a polished and user-friendly experience upon final release.

---

3. Tuxcare Extends Support to Microsoft .NET 6.0

In a significant move within the open-source ecosystem, Ken introduces the news that Tuxcare is now offering extended lifecycle support for Microsoft .NET 6.0.

• Referencing an article by Christine Hall, Ken explains that although Microsoft .NET 6.0 reached its end-of-life on November 12, 2024, Tuxcare provides security patches for vulnerabilities, filling a critical gap for enterprises reliant on this framework.

• Jeff and Jonathan discuss the implications, noting that supports like Tuxcare’s are becoming essential as companies seek to maintain legacy systems without the overhead of upgrading or risking security breaches.

• Jonathan remarks, “[31:58] * ...NET is so the metric. It’s at least for me that is my metric for whether something is actually completely open sourced…*”

This development highlights the evolving landscape of open-source support, especially concerning proprietary technologies integrated into open-source environments.

---

4. OpenSUSE Tumbleweed Embraces Wayland with LXQT

The episode then explores desktop environment advancements with OpenSUSE’s latest Tumbleweed release.

• Ken discusses the introduction of LXQT 2.1 with experimental Wayland support, allowing users to experience a lightweight desktop with the modern Wayland compositor.

• Detailed installation steps are provided, emphasizing the need to manually install Wayland compositors such as Kwin, Wayfire, or Sway to complement the LXQT session.

• Jeff expresses his preference for KDE, highlighting the personalized customization it offers, while Jonathan appreciates the balance between functionality and aesthetics in KDE’s design.

This segment is particularly valuable for users interested in exploring alternative desktop environments while leveraging the benefits of Wayland.

---

5. PCI Express 7.0: The Next Generation of Connectivity

A substantial portion of the discussion is dedicated to the advancements in PCI Express (PCIe) 7.0, a critical hardware standard for high-speed communication between components.

• Jeff breaks down the technical aspects, explaining that PCIe 7.0 aims to achieve a theoretical throughput of 512 gigabytes per second using PAM4 signaling, which quadruples the bit representation compared to the traditional NRZ signaling used in previous generations.

• Jonathan questions the practical applications, pondering, “[59:48] what kind of devices does this even make sense for? Right…” The hosts agree that while current consumer hardware doesn’t fully utilize PCIe 5.0’s bandwidth, the leap to PCIe 7.0 will primarily benefit enterprise and industrial applications, including networking, memory pooling, and high-performance computing.

• The conversation also touches on the challenges of compliance testing for PCIe 7.0, highlighting the complexities introduced by the new signaling technology.

This insightful dialogue underscores the forward-looking nature of hardware development, anticipating future needs in data-intensive applications.

---

6. Fedora 42 Introduces Official WSL Images

In a nod to developers who operate across both Linux and Windows environments, Jonathan announces that Fedora 42 will offer official Windows Subsystem for Linux (WSL) images.

• Jonathan explains, “[71:03] now has approved the shipping of Fedora Linux WSL images. And so that means that starting with Fedora 42…” This allows users to install Fedora directly from the Microsoft Store, enhancing flexibility for those who require Fedora’s environment within Windows.

• The hosts acknowledge that while this may not be widely utilized among all listeners, it represents Fedora’s commitment to supporting diverse development workflows.

This addition expands Fedora’s accessibility, catering to a broader audience that integrates Linux tools within Windows-based systems.

---

7. Command Line Tips: Managing Audio Devices and Printer Queues

The show also offers practical advice through command-line tips, enhancing listeners' system management skills.

• Ken introduces the PipeWire Device Reservation Utility (pw-reserve), demonstrating how to manage audio devices effectively. He walks through command examples and screenshots, showing how to reserve and release audio resources, which is particularly useful for users dealing with multiple audio interfaces.

• Jeff covers the lpq command, a tool for monitoring printer queues. He outlines various options, such as specifying different users (-U) or remote servers (-H), enabling users to efficiently track and manage print jobs.

These segments provide actionable insights, empowering users to optimize their system configurations through the command line.

---

8. Proton 923 Release and Gaming Enhancements

Concluding the technical discussions, Jeff highlights the release of Proton 923, a critical update for Linux gamers utilizing Steam’s Proton compatibility layer.

• He mentions improvements in Wine, DXVK, and fixes for Battle.net integrations, enhancing the gaming experience for titles like BioShock Remastered.

• Jonathan notes the benefits for the gaming community, appreciating the continuous advancements that make Linux a more viable platform for gamers.

This update signifies ongoing support and optimization for gaming on Linux, reinforcing its growing viability in the gaming sector.

---

Notable Quotes

• Jonathan on rsync vulnerability severity:

  “[06:37] Jonathan: That's real bad.”

• Jeff on Plasma 6.3 improvements:

  “[17:16] Jeff: ...there's a handful of really cool things in 6.13 and here in a bit we're going to talk about what's coming in 6.14.”

• Jonathan on Fedora’s WSL support:

  “[71:03] Jonathan: ...you'll be able to jump on a Windows machine, go to the Microsoft Store and say install Fedora...”

• Jeff on Proton 923 release:

  “[87:12] Jeff: ...Proton 923 was released. It's, you know, updates wine to the bleeding edge...”

---

Conclusion

Untitled Linux Show 186 delivers a comprehensive overview of pressing security issues, desktop environment innovations, extended software support, and cutting-edge hardware developments. The hosts provide valuable insights and practical advice, making this episode a must-listen for Linux enthusiasts keen on staying informed and enhancing their system management skills.

For more detailed discussions and weekly updates, listeners are encouraged to visit Hackaday and support the show through Club Twit.

Transcript

[Jonathan]

Hey folks, this week we're talking about rsync, which you really need to go update. We're talking plasma, the kernel, 6.13 release, and all the fun things coming in 6.14 and some interesting information about Fedora on Windows of all places. It's a lot of fun. You don't want to miss it. So stay tuned listeners, as we go.

[Sponsor]

Into a new year, we all have a lot on our plates. There are backpacking trips across Europe to plan, personal best to crush in the gym, and capsule wardrobes to create Good thing. Our sponsor, NerdWallet is here to take one thing off your plate. Finding the best financial products introducing NerdWallet's 2025 Best of Awards List your shortcut to the best credit cards, savings accounts, and more. The nerds have done the work for you, researching and reviewing over 1100 financial products to bring you only the best of the best. Looking for a balance transfer credit card with 0% APR? They've got a winner for that. Or a bank account with a top rate to hit your savings goals? They've got a winner for that too. Know you're getting the best products for you without doing all the research yourself. So let NerdWallet do the heavy lifting for your finances this year and head over to their 2025 Best of Awards at NerdWallet.com awards to find the best financial products today. I'm no tech genius, but I know if I want my business to crush it, I I need a website now. Thankfully, bluehost makes it easy. I customized, optimized and monetized everything exactly how I wanted with AI. Within minutes my site was up. I couldn't believe it. Plus, Bluehost keeps me protected 24. 7. No more worrying about hackers. I just relax and focus on my customers. So whatever your hustle may be, secure your online presence in 2025 with Bluehost. Head to bluehost.com to start now.

[T-Mobile]

At T Mobile, we'll give you four free 5G phones and four lines for only $25 per line per month with eligible trade ins. And no, it's not a contest. It's every day for a limited time. Everyone's a winner on America's largest 5G network.

[T-Mobile Disclaimer]

Minimum of 4 lines for $25 per line per month with autopay discount using debit or bank account. $5 more per line without autopay. Up to $830 off each phone via 24 monthly bill credits plus taxes, fees and $10 device connection charge. 4 well qualified customers contact us before canceling entire account to continue Bill credits or credit stop and balance on required finance agreement too Bill credits end if you pay up Devices early ct.

[Jonathan]

Podcasts you love from people you trust this is Twit. This is the Untitled Linux show, episode 186, recorded January 18th. Accidental honeypot. Hey folks, it is Saturday and you know what that means. It's time to geek out about Linux. It's the Untitled Linux Show. We've got news and open source and Linux and hardware, all kinds of fun stuff going on. It's not just me. We've got Jeff and we've got Ken. Our other regularly scheduled co host is down with a cold and we think it's kind of lame that he's not here, but we'll I guess, give him a bit of an excuse for being sick.

[Jeff]

I don't know, I've. I've got a bit of a cold, so if I sound a little funny this week or I have to pause to cough or sneeze or something, that's why. But I'm. I'm here because I'm. I'm powering through it.

[Jonathan]

Yes, you're just. You're more dedicated than Rob is, obviously.

[Jeff]

Yeah, I wasn't gonna say that.

[Ken]

Jonathan and I are both recovering from the colds that came through here.

[Jonathan]

Well, yeah, very cold came through here. Goodness. So, Ken, you've actually got the first story and you sort of stole it from me. But I'm not bitter. Tell us what is.

[Ken]

You covered it yesterday.

[Jonathan]

I did. Tell us what is up with rsync and then I'll maybe fill in some color commentary afterwards.

[Ken]

Well, I'm hoping you will, because I'm just going off of what Bobby Barsoft reported, which is that rsync version 3.4 is coming out addressing six vulnerabilities that rsync versions 3.3 and earlier had. I'm not going to go through because I don't think you want to hear me repeatedly saying something like CVE 2020 dash 12.084 every time I before giving a description of it. But that. That is the first one for a heap based buffer overflow flaw in the rsync daemon. That's the first of them. The second one has a flaw in the rsync daemon that appears during file checksum comparisons. The third one is a flaw in rsync that could let a malicious server read files from a client's machine that sound good. Then the fourth one is a path traversal flaw in rsync that allows a malicious server to write files outside the intended directory. The fifth flaw is a flaw in R sinks and Boris put this in Quotes Dash dash safe dash links option that fails to verify Test nested symlinks not tested nested the last one that Boris reported about and this was CVE 202412747 rolls right off the tongue. This one has a race condition in Rsync simlink handling where it can bypass its default link skipping. Now Bobby also reports that rsync version 3.4 includes some noteworthy changes, such as move the FreeBSD continuous integration pipeline to GitHub actions it silence compiler warnings about unused variables, thereby reducing clutter in the code. It also Upgraded capital P Opt 1.19 for Enhanced Command line parsing and improved consistency. Now, I do recommend reading Bobby's article if you want the link to Rsync's changelog or if you need other details I haven't covered and I believe Jonathan included a link to what he had pulled up for what he wrote about it.

[Jonathan]

Yes. So this was the headline story from my this Week in Security column over on Hackaday and it's that first vulnerability in rsync specifically. So the disclosure text has a bit of I called it nightmare fuel and here it is. An attacker only requires anonymous read access to an rsync server, such as a public mirror, to execute arbitrary code on the machine the server is running on. That's real bad.

[Jeff]

That's scary.

[Jonathan]

Yes. So if you have rsync 3.3.0 or before and it is running as a server and it is exposed to the Internet, it's potentially vulnerable and potentially a zero click. Just connect and you get to execute code on it. It's kind of the worst case scenario. This clocks in at a 9.8 CVSS, I believe, which 1010 being the worst. 9.8 means it's almost as close as it could possibly be to worst case scenario and what it is. This is a it doesn't necessarily require authentication, Right. This is particularly important if someone is offering public files over rsync. That doesn't happen very often, but it does happen from time to time and it is the handling of the checksum links. So when rsync is moving files back and forth, it sends data and then you send back a checksum which is I've added all the bits together in a special way and here's the answer I came up with. Does that match the answer that you have? And if the two match, then you know, you transferred the data over Correctly. The problem is that when you send one of those checksums, part of the information that you send is, and by the way, here's how many bytes I'm sending you and you can say, by the way, I'm going to send you 15,000 bytes. Not literally. I'm making that number off the top of my head and it still gets copied over a 16 byte buffer. And so it just goes right past the end of the buffer. Writes over other important things. And for those of you, for those of us that follow security issues, you understand that writing past the end of the buffer is how you get remote code execution. So it's not good, it's not good at all. For rsync. If you've got public rsync servers, your hair is on fire with this. Red Hat has a little bit more to say about it. I've got that linked in my article as well. The important thing though is that there is not a mitigation for this. You just need to update. And so when we talk about like in this sort of context of mitigation, that is, there's nothing you can do in your configuration file. Like you can't go, you can't go into configuration file and easily say, just don't allow, you know, SHA256. So sometimes with these sorts of vulnerabilities, it'll be that you can just go say, don't allow SHA256 checksums and everything's fine. Not the case here. You can, from the reading I did, you could compile a copy of rsync so you could compile the older so 3.3 rsync and leave out the checksum. But you can't do it in configuration. So, you know, for the vast majority of us it's either turn it off or update it. And that kind of needs to happen yesterday. It's bad.

[Ken]

If I remember correctly, there is a flag that you can use on the command line to turn it off.

[Jonathan]

I believe that was a compile time option in what I saw because Red Hat, I'll click the link, we'll see if they've changed it. But mitigation, Red Hat recommends filtering untrusted connections to turn your firewall on. Additionally, systems which only need to provide remote rsync access to users, users with known identities can enable authentication. But systems that provide anonymous read access to hosted files via rsync, such as mirror hosts, do not have reasonable mitigation options available. So that is. That is Red Hat's latest on it.

[Jeff]

When, when Was this released?

[Jonathan]

January 14th is when this became public. Last modified January 16th.

[Jeff]

Okay.

[Jonathan]

And the. See if I can quickly find. When.

[Jeff]

I was kind of curious. Mine, mine hasn't updated to 3.4 yet. My system hasn't updated 3.4 and I would think that'd be one of those. That would, they would, they would push through pretty fast.

[Jonathan]

Yeah, you would. You would hope so. The, the nice thing about this is most people are not running rsync Connect, like the rsync server exposed to the Internet and you know, doing it with anonymous access enabled. That is not something that hardly anybody is doing. Although I say that according to the Shodan, which is sort of the Internet search engine that doesn't search for websites, it searches for services. Apparently there are like 664,000 thousand rsync servers that are available on the Internet.

[Jeff]

Wow. Yeah, I mean, I'm in that first boat of I don't have an rsync server. I don't have any of that, so it doesn't matter.

[Ken]

I was just kind of curious how fast it updated as a backend for a lot of backup utilities.

[Jonathan]

Yes. Even at that, though, you can do it like over SSH or over vpn. And I think it is less vulnerable there, particularly with over ssh. If you require authentication first, like if your rsync server is always set to require authentication first, then you're not vulnerable to this. Well, you are, but only two people that can authenticate. Right. Which is a very different thing than you're vulnerable to this from everybody on.

[Ken]

The Internet unless you configured so everybody can authenticate.

[Jeff]

Yeah, I use rsync, but it's all local from machine. Just in my little home network. I don't ever use anything out on.

[Ken]

The Internet, just from one local directory to another.

[Jonathan]

Yeah. So your exposure to this, Jeff, is. It's only a problem. Assuming that's the case and you haven't accidentally exposed rsync out to the Internet. Your exposure to this is limited to if someone is already on your network, which is a very different problem, you still need to update it as soon as you can. But that's an extremely different problem from this is exposed to everywhere. And the moment that someone publishes the actual exploit for this, all of these systems are going to get attacked.

[Ken]

Right. I know one utility that uses rsync in the background is rclone, which I use for doing some of my backups to my Google Drive.

[Jonathan]

Right, Right. But again, it's. That's. That's going to be the same scenario. You're not, you're not exposing this as a server to the Internet. So that, that's not.

[Ken]

That I know of.

[Jonathan]

So go, go use. I think, I think Gibson still has shields up, right? If you're. If you're worried about your. Your local network, go use shields up. The two ports that you want to check are 873 and 8873. And if those show closed or stealth, then you're good to go as far as your local network goes.

[Jeff]

So everyone's gotta go on there, which.

[Ken]

I probably should do anyways just to make sure I don't have anything open.

[Jeff]

That's what I do.

[Jonathan]

I accidentally left SSH open one time to an open WRT router that had a very, very weak password on it. Thankfully, it was virtualized inside of this. So I accidentally made a honey pot because it was virtualized inside of another machine. And I had it forwarded through and I'm pretty sure I had it firewalled off from the rest of my network. And so somebody did get into it and I think I was able to pull some logs and see, and it was like, try this, try that. Nothing worked. And they just kind of threw their hands up and walked away and, ah, it's a dumb little embedded machine I can't do anything on. I still blew it away, wiped it out, and, you know, panicked just a little bit because somebody did in fact, SSH into my network. But it's a long time ago.

[Ken]

Deep breathing.

[Jonathan]

Yeah, yeah, yeah. It's one. It's one of those, one of those clench moments. Like, oh, no, this could be really bad. Oh, it's not as bad as I thought.

[Sponsor]

Listeners, as we go into a new year, we all have a lot on our plates. There are backpacking trips across Europe to plan, personal best to crush in the gym, and capsule wardrobes to create. Good thing. Our sponsor, NerdWallet is here to take one thing off your plate. Finding the best financial products. Introducing NerdWallet's 2025 Best of Awards. List your shortcut to the best credit cards, savings accounts, and more. The nerds have done the work for you, researching and reviewing over 1100 financial products to bring you only the best of the best. Looking for a balanced transfer credit card with 0% APR? They've got a winner for that. Or a bank account with a top rate to hit your savings goals. They've got a winner for that too. Know you're getting the best products for you without doing all the research yourself. So let NerdWallet do the heavy lifting for your finances. This year and head over to their 2025 Best of Awards at NerdWallet.com awards to find the best financial products today.

[T-Mobile]

AT T Mobile we'll give you four free 5G phones and four lines for only $25 per line per month with eligible trade ins. And no, it's not a contest, it's every day for a limited time. Everyone's a winner on America's largest 5G network.

[T-Mobile Disclaimer]

Minimum of 4 lines for $25 per line per month with autopay discount using debit or bank account $5 more per line without autopay up to $830 off each phone via 24 monthly bill credit plus taxes, fees and $10 device connection charge for well qualified customers. Contact us before canceling entire account to continue bill credits or credit stop and balance on required finance agreement to bill credits and if you pay UP devices.

[Jonathan]

Early ct mobile.com all right Jeff, let's move on from security to Plasma Yeah.

[Jeff]

I know, I know it seems like we talk about KDE a lot, but that's because there's a lot going on as we get ready for the Release of Plasma 6.3. You know, if you remember, last week we talked about how Plasma 6.3 just hit the beta stage in preparation for its final release and we're talking about it again this week because of the number of bugs that have been fixed in just one week. Nate Graham has released another blog on the Planet KDE website which is of course linked in the show notes. But you know, let's get right into the meat of it. As of January 18th there's one very high priority bug which is down from three last week. And the very, very high priority bug is described as, and I'm quoting here, sometimes occurs on X11 with compositing turned on causing a black lock screen when using the Breeze plasma style. But controls are all there and remain interactive. So not the most smooth flowing sentence. But you know, while it is considered a very important bug, it's only affecting people still on X11 and not on Wayland. Now the 15 minute Plasma bug count is currently down at 23 from 36 last week and in general 136 KDE bugs of all levels have been fixed in the last week. In addition to all the bugs that have been squashed, there's also new features and polish that's gone into the ui. For example, scaling has a couple of improvements such as Kwin now being smarter about choosing a default scale factor for devices with small screens. It won't choose A scale factor too high to be impractical. And Kwin's automatic scale factor chooser now also chooses a scale factor rounded to the nearest 5%, no longer to the nearest 25%. Night light is now going to have the correct colors when using a custom color profile or an ICC profile. If your keyboard has a button to toggle the keyboard backlighting on and off, that button now works on the lock screen. And when a panel defloats and causes its pinned open widget pop ups to also defloat, the pop ups defloatiness animations are now beautifully synced up, so everything just looks good now. And for those who run several desktops, when switching between virtual desktops using your meta, alt and scroll keys, it now goes in the correct direction you'd expect when you're using a reversed or natural scrolling I'm not going to dive into the notable bug fixes listed in the article because there are just so many of them, but if you wish to know more, there are links to the full change log so you can see all the changes that have happened in the 6.3 beta so far by following the link in the show notes. And you know, by the time Six.3 comes out, it should be a very polished product and I'm really excited to see it.

[Jonathan]

Yeah, very cool. I the guys may have noticed that I whipped around to look at something I have. I'm running pretty much. I don't think I'm on 6.3 yet, but I'm on the latest KDE and there is a there's a patch bouncing around for Firefox that's supposed to get HDR working. So earlier today I was fiddling around with that and finally got it compiled and I'm now at the point where YouTube thinks it's playing HDR content. It's not. I don't think to my eyes it does not look like it, but it thinks it is.

[Jeff]

So like we're this close guys, and so everybody remembers Jonathan does have an HDR monitor.

[Jonathan]

Oh yes, the HDR TV went out. Yeah, I bought the smallest OLED TV I could find and it's a great monitor. It's a great 4K monitor.

[Jeff]

Well, actually there's not much difference. And the only real difference between a monitor and TV these days is a TV is still required to have the the old analog input decoder box in it. And so you can still hook up your old Atari 2600 to a modern TV because of that little decoder box they have to stick in there to Call it a tv.

[Jonathan]

Yeah. You know what I have found? The. The other difference is on a TV, they put some smart OS on it, like Google TV or, or what's. What's the other?

[Ken]

WebOS.

[Jonathan]

Yeah, WebOS is what this one is. Yeah, yeah. If you want to get a TV that does not have one of those, you get to pay about three times as much and get a commercial display tv, I think is what they call it. If you want a dumb TV that doesn't have any smarts and therefore is not spying on you, you get to go pay for a commercial display tv. Yeah. And that actually is one of the reasons why monitors like 4K HDR monitors are so much more expensive than the TVs are, because the TV manufacturers are making some of the money back by having their smart OS on it.

[Ken]

By selling. Selling your information.

[Jonathan]

Yes, that is. That is part of what happens. Yes. If we're going to be frank about it.

[Jeff]

Yeah, I have a smart tv, but, you know, I never let it talk to the Internet. I actually run a separate, like Roku or something that. So that the TV hasn't seen the Internet. It doesn't. Nobody knows it exists on the Internet as far as the manufacturer's aware, because don't let it talk. Don't let it have the network.

[Jonathan]

Yeah, yeah. All right, let's see here. We've got more stuff to talk about. Oh, yes. So this weekend is a very interesting weekend in kernel land. It is expected. And so my coverage on this comes by way of Michael Darabold on Pharonix, who is like the kernel whisperer in the media. He is the guy that knows how to read the tea leaves and tell you what's coming with the kernel. And so we are going to recap his work here. So this is the weekend that 6.13 is going to release. And there are some really interesting things in there. More Support for an AMD Epic, but the AMD 3D V cache optimizer is finally going to land in 4.13. More stuff for the Raspberry PI. And so he calls up better performance for Super Pages or Big Pages with driver support. For the 3D driver. There's a couple of other. He talks about the tipping point for rust in the kernel, and I went on a. A bit of a crusade a few weeks ago to find out. And there are actually like working real rust drivers in the kernel. It was so that I wasn't sure about whether or not it was actually being used. And there are some. So depending upon what hardware you have, you may actually be running some Rust code. But you know, more and more is landing and we have more and more people writing code in Rust that's kind of controversial. Not everybody appreciates that and I can respect whether it's controversial. There are some real challenges with adding another language to the kernel that is a very real thing. But anyway, and then there's just your normal kind of churn of new things that are happening between the Scheduler. The NVMe 2.1 specification has finally landed some other things like that. So there's a handful of really cool things in 6.13 and here in a bit we're going to talk about what's coming in 6.14. And there is at least one thing in 6.14 that I am particularly excited to see happen, but some good stuff in 13 as well.

[Jeff]

Or at least a subset of what's coming in 14.

[Ken]

Well, yeah, which of the options are dropping tomorrow do you look forward to seeing? What do you mean of everything that's dropping with 6:13 tomorrow?

[Jonathan]

Ah, the one that I will probably have the most, the most chance of working with directly is the better performance for the Raspberry PI. I will probably do another, you know, vanilla Raspberry PI install of say Fedora and get the 613 kernel on it. That'll probably be the one that I touch and really see the benefits of first.

[Jeff]

So yeah, personally I don't know if there's much coming that is going to have really any big. I mean there's, there is some like, you know, new AMD CPU features, things like that that might be kind of cool, but I think, you know, realistically a lot of times for the average person, you're not going to really notice that much unless there's a specific piece of hardware or specific bug or something that applies to you that it's going to fix. If your kernel is running fine, you probably not going to really notice anything unless maybe if you're running benchmarks, but just day to day computer use.

[Jonathan]

Yeah, or if you're running like the latest generation of graphics card, you may see a bump there. Although anymore a lot of that code happens in mesa, which is not actually part of the kernel, that's part of your user space.

[Ken]

I hope nobody tries going to 6:13 that's using a riser fs file system.

[Jonathan]

Well, I would hope that nobody is using Riser FS that is not aware that it is being phased out. And that's the previous version of Riser FS2, right.

[Jeff]

Yeah, it's an old one.

[Jonathan]

Yeah.

[Jeff]

The latest version is actually out of the kernel.

[Jonathan]

Yeah. Yeah. So if you're, you're not running that one by accident.

[Jeff]

Yeah, yeah. And we've talked about in the past where it's kind of gotten surpassed by, you know, it was, it was really cutting edge when it first came out, but right. Over the years things have kind of equaled it or surpassed it. So it's, it's not the groundbreaking file system it was at one time.

[Jonathan]

Yep, Yep. All right, so let's move on. What's up, Ken, with Tuxcare and dot.

[Jeff]

Net channeling Rob actually no.

[Ken]

Was reading an article by Christine Hall. She wrote about continued support for For Microsoft's Net 6.0 from Tux Care through Tux Care's endless lifecycle support program. Well, bad news is Microsoft Net 6.0 reach end of life on November 12, 2024. Yes, last year or last November. So not that long ago. Fortunately, Tux Care is one of several companies offering aftermarket support for open source operating systems and other essential platforms after its developers move on. Yes, Microsoft is moving Tux Care service will supply security patches for any vulnerabilities in Microsoft Net 6.0. According to Christine, it's becoming more common for sysadmins and DevOps teams to rely on services like Tux Care rather than go through the often frustrating and always time consuming and costly ordeal of upgrading a process that often breaks essential secondary components which then have to be upgraded or fixed. Also, and I do apologize if I pronounce mispronounce this, but Joaquin Korea, a technical evangelist for both Tux Care and its parent company Cloud Linux, points out that supporting a product for Microsoft is something new not only for Tux Care, but for any of the other companies that offer aftermarket support for expired open source projects. Quoting Joao, that's somewhat of a departure in that we are now touching on proprietary Microsoft space, which we hadn't touched before. At the same time, it's an open source project and we have lots of experience with those types of projects and maintaining them today. You have to admit it, Microsoft is a very large contributor to open source space. They have a presence on multiple projects. They have a very distinguished presence on the Linux kernel directly. So yeah, they're essentially a considerable player in the open source space now. Christine feels it's still pretty much business as usual and mainly just underlines how much the open source landscape has changed since the open source ought to or 2000s when Microsoft was public enemy number one at least in far circles.

[Jonathan]

Yeah.

[Ken]

Despite this, whenever Redmond is positively mentioned in an open source context, the need to defend treating it without rancor definitely surfaces. Many open source developers were pretty much jubilant when Microsoft open. Sourced. Net 6.0 and have been relying on. Net for years now. This makes the reasoning behind supporting it post end of life the same as it would be for offering extended support for any open source platform. Especially since Software written for. NET 6.0 will also need to be ported, which means even more time and more effort to fix something that isn't really broken. I recommend reading Christine's article for more details on this decision by Tux Care and her thoughts on providing aftermarket support for expired software.

[Jonathan]

Yeah, that is very interesting that Tux Care is caring for. NET now, but it's an open source project. In fact NET is this open source. It's this much open source. You can just install it with DNF and Fedora. Now that's like. That's. That's kind of the metric. It's at least for me that is my metric for whether something is actually completely open sourced and not patent encumbered is can I just get it in a vanilla installed Fedora?

[Ken]

Which means it's probably installable on RHEL as well.

[Jeff]

Yes.

[Jonathan]

Yep.

[Ken]

So those enterprise users are probably definitely going to take advantage of Tux Cares extended life cycle support.

[Jeff]

I'm guessing the later versions of. Net there's some API breakage in there so that it's not fully backwards compatible.

[Jonathan]

I assume so. I am not a. NET developer, so I can't speak exactly to that. There's another question though, and this is based on an assumption, but is Red Hat also doing extended. NET support? If they have support for any of these releases of Red Hat Enterprise Linux that have those. NET packages, they would as part of their support contract also provide extended support for it beyond what Microsoft does. So maybe it's not just Tux Care. Maybe there's a couple of companies doing it.

[Ken]

Yeah. The article mentions one or two other names, but makes you wonder REL is actually providing the support themselves or if the May might be reselling somebody support.

[Jonathan]

That is very possible. Or you know what may even be happening is not so much that they're reselling, but it might be that. Let's see who's the other one that it does really, really long support. It's. It's Susie Lenno. Susie Linux Enterprise suse.

[Jeff]

Suse.

[Jonathan]

That's what I said. Sles. I believe they are stepped up to come crazy amount of support, like 10 years of support or something. So it may be that both of those companies have gone to Tux Care and said, hey, you guys are really good at this. Why don't we pay you some money and you watch out for vulnerabilities. But there are, because those packages are now available, there are going to be several different places that sort of have, have some responsibility to make sure that nothing too terrible happens to it.

[Jeff]

Yeah, makes sense. And I can see a large company like Red Hat just saying, oh, you know what, it's just easier and cheaper for us to just pay somebody that's already kind of doing this to. We will kind of license the support.

[Ken]

And partner with them.

[Jonathan]

Right, right.

[Sponsor]

Listeners, as we go into a new year, we all have a lot on our plates. There are backpacking trips across Europe to plan, personal bests to crush in the gym and capsule wardrobes to create. Good thing. Our sponsor, NerdWallet is here to take one thing off your plate. Finding the best financial products. Introducing NerdWallet's 2025 Best of Awards. List your shortcut to the best credit cards, savings accounts and more. The nerds have done the work for you, researching and reviewing over 1100 financial products to bring you only the best of the best. Looking for a balance transfer credit card with 0% APR? They've got a winner for that. Or a bank account with a top rate to hit your savings goals. They've got a winner for that too. Know you're getting the best products for you without doing all the research yourself. So let NerdWallet do the heavy lifting for your finances this year and head over to their 2025 Best of Awards at NerdWallet.com awards to find the best financial products today at T Mobile, we'll.

[T-Mobile]

Give you four free 5G phones and four lines for only 25 per line per month with eligible trade ins. And no, it's not a contest. It's every day for a limited time. Everyone's a winner on America's largest 5G network.

[T-Mobile Disclaimer]

Minimum of 4 lines for 25 per line per month with auto pay discount using debit or bank account. $5 more per line without autopay. Up to 830 off each phone via 24 monthly bill credits plus taxes, fees and 10 device connection charge. 4 well qualified customers. Contact us before canceling a entire account to continue build credits or credit stop and balance on required finance agreement to bill credits and if you pay UP.

[Jonathan]

Devices early ct mobile.com all right Jeff, you're gonna, you're gonna Jump the gun. I didn't, I didn't see this until just now. You're gonna talk about something in 614. It's okay. That's not the thing that I'm super excited about for 614, but this one does sound interesting.

[Jeff]

Yeah, I'm. Well, I'm kind of covering a narrow band here, so if, and if you're a listener to this show for any length of time, you know that I like gaming. And because of this, I cover a lot of the graphics news, both hardware and software. And, you know, it's looking like we're going to have a new kernel release tomorrow, January 19th, you know, and as Jonathan said, the 613 kernel is expected to be released. And when the kernel, you know, when the kernels released, candidate, you know, is now deemed stable, the merge window opens for the very next kernel, which is of course going to be 6.14. I bring this up because people are already queuing up to have their code pulled in during the merge window. Now I'm going to cover some of the graphics items that are getting pulled into 6.14 and some code that's kind of near graphics. There's probably a lot of near graphics here. And I say near because, for example, with this pull, David Aryle of Red Hat is pulling up an AMD XDNA driver which is brand new. And this driver supports the Ryzen AI NPUs under Linux. NPU stands for Neural Processing Unit, which is what AIs are based on neural networks. And the really short version is the code helps synchronize, synchronize hard, helps specialized hardware compete AI tasks at a very accelerated rate. So this is little bits of silicon that are specially designed to do these compute tasks, and which is kind of what a GPU is. It's a very specialized piece of silicon for special compute tasks. So this, this though, is going to be AI and it's going to greatly increase the computing you can do in AI. More AMD GPU news. The kernel, graphics drivers, adding DRM panic support. You know, this is also kind of referred to as the Linux blue screen of death, hearkening back to the windows when it goes awry. You know, everybody talks about the blue screen of death, synonymous with crashing. And by tying into the kernel better when something happens, you'll actually see the error state as it happened and not rely on a kernel dump which has to be configured to happen depending on the settings, you know, and when it was, how it was compiled. Not all kernels will give a core dump in this case, when something happens, you'll be able to get a QR code error message to allow for better troubleshooting and pinpointing of the problem as it exactly happens. So you there are some thoughts that, you know, the core dumps, while they can tell you what happened, you know, there's a lot more sifting through versus at the time of problem, it's a much more pinpointed error message. The driver is also getting cleaner shader support for rdna 2 graphics cards, dcn 3.5 improvements, which is the instruction set for rdna 3.5 cards, which would be the strict point cards. There's also Updates and various rdna4 and graphics12 improvements. So I'm summarizing the fixes that are going into the driver. But in the article in the show notes, there's a further link that goes deeper into all the improvements that are landing, or ideally landing in the 6.14 kernel. Just because you're on the pull list doesn't mean you're actually going to get in. And we've seen this in the past. But you know, when it's coming from people like Red Hat or you know, Intel's AMD's, you, you have a very high likelihood that if they're putting it in for the merge, it'll make it in. Speaking of intel, intel is getting some love as well in the form of their XE and i915 drivers, which are bringing SR IOV space pf scheduling priority. SR IOV stands for single root IO virtualization and the pf is physical function. This will basically this allows you to take a single PCIe device and you make it look like several PCIe devices by virtualizing the actual hardware. So this would be, for example, if you have a device that supports, you know, you have like a graphics card that you can kind of break into like say two graphics cards and then you could have two users actually using the graphics card at the same time. Enterprise hardware supports this more of the time, but they don't like to have their consumer cards a lot of times do this. But there's other applications as well. So it's now coming through the intel driver. The next generation Panther Lake chipset is also going to get support for ultra high bitrate uhbr mode for DisplayPort with Thunderbolt in its alt mode. So uhbr10 allows for 40 gigabits per second bandwidth and uhbr20 gives 20 gigabits per second bandwidth. There's also several other graphics improvements for the intel code. Take a look at the article in the show notes for more details as I've skipped over a lot of, you know, graphics and improvements, other chipsets that I haven't talked about. So lots of stuff I didn't mention but all the details can be found there and they've got several links to sub articles which go into greater detail, even greater detail of the changes that are coming. So basically we're set to get a lot of graphical and AI goodness in the 614 kernel.

[Jonathan]

Yes, some really fun things there. What I wanted to talk about I will go ahead and pull up is the NTsync feature and that is something that is in the kernel itself and it is supposed to land in 6.14. It looks like now with 6.14 we're sort of reading the tea leaves right, because these things have not been merged yet but. But it's larable looking at the various branches where people track this stuff and into sync is the it's sort of a primitive from Windows that has been added to the kernel specifically for Wine and that's kind of a minor miracle that the kernel guys allowed it, right. Like that's sort of the thing that in the past they've sort of not appreciated. I think it helps a lot that Valve is pushing this and Valve is doing some other neat things. Anyway, if you click through on the articles listed you've got the percentage of improvement on various games with all of the NTsync stuff running and you've got like dirt 3 has a 678% frame rate improvement. It goes from 110 to 860 frames. So there's just some. There's some big gains to be had and it is a big enough gain to be noticeable. Absolutely. Across the board. So I kind of Predict that with 6.14, assuming that distros turn this on like it's probably going to be a compile time option, but assuming that distros actually turn this on and get it to people. 6.14 is going to show a huge performance improvement on a bunch of different Windows games running through Wine, running through Proton. So it's going to be really cool. It's going to be a big deal. There's a few other things that are in here that are really, really interesting. The AMD GPU DRM panic support for the blue screen of death. Yay. We have more blue screen of death in Linux. And of course yes, I'm being sarcastic, but at the same time that can be useful more than just so it's better than Your system just absolutely freezing, or it may be better than your system going to just the console to show you what's going on. You get a little bit more information about it. And there's other updates in here. There's a new RISC V processor, the one that the Banana PI development board is built around is now supported in the kernel. That is pretty neat as well. And then a bunch of sort of smaller things that, like we mentioned earlier, unless you're hitting this bug or you have this hardware or it's a thing that you specifically want to be able to do, it probably doesn't make that much difference to you. Just sort of acknowledge that it's a bunch of little fixes and little updates and little improvements all around. But between the GPU stuff that Jeff just talked about and the NTsync support, I think 6.14 is going to be really cool and really exciting. I'm looking forward to it.

[Jeff]

And 6.14 is the version that Ubuntu is planning to ship. Yes, 25.04.

[Ken]

I'm looking forward to the Chrome OS UCSI driver.

[Jonathan]

I saw that as I. As I flipped through here. What exactly is UCSI.

[Ken]

USB C Platform policy? It's used, featuring the. Used on Chrome OS with Chromebooks, featuring the Chrome OS embedded controller and supporting a platform policy manager. Basically, UCSI is the USB Type C connector system software interface.

[Jonathan]

That's the sort of thing that doesn't sound necessarily too exciting, but you might see people doing some really interesting things with it in the future. Those are the sort of drivers that give you really interesting abilities when plugging.

[Ken]

Things in, like using power and audio and video, potentially all on the same cable.

[Jonathan]

Yep, Yep. Potentially.

[Jeff]

You know, we both kind of touched on the blue screen of death there, but we probably should mention that, you know, we mentioned core dump. And if anybody doesn't know, a core dump is where your kernel is basically crashing and it writes a whole bunch of information to your disk so that ideally you can go in there and parse through it and figure out why your kernel crashed. But it's kind of just a big. Literally a dump. So there's a lot of kind of rooting through it and digging to figure out what's. What's going on.

[Jonathan]

A dump of core memory, which. Core memory is a term from way back in the old days when we were talking, you know, when, when mainframes roved the land.

[Ken]

Now do you think we'll see bcache fs merges this with 614 there.

[Jonathan]

There was also yes. Something with bcash fs where they're doing round robin raid 1, which you know, has the potential to make raid 1 on bcash fs a decent bit faster. And we will see.

[Ken]

Was it BCASH FS or btrfs? Butterfs.

[Jonathan]

Oh, that was Butterfs. I know what you're asking about, Ken. Okay. I've not seen any news on that, so.

[Jeff]

Well, it should. It should, yeah. For people that don't remember we talked about this two or three episodes ago where the. The. One of the main programmers for BCASH FS kind of got in trouble on the current kernel Linux mailing list and they denied his merge request. But it was only supposed to be for the 613 kernel and they were going to open it back up to that person again in 614. So it should be available for them to put in their code to get merged. And a lot of times when they open the merge window, people don't always have everything queued up ahead of time. So just because we haven't talked about it, we don't see it listed yet, doesn't mean that it's not going to get submitted. They have a week to submit their.

[Jonathan]

Code, up to two weeks, I think in some cases. Yeah, I think it's two weeks.

[Jeff]

Okay.

[Jonathan]

I think it can depend on how it lines up with other schedules. Like if a major holiday is going to be right in the middle or right at the end of the merge window, it'll get adjusted a little bit in either direction, but I think it's usually right at two weeks.

[Jeff]

So I guess bottom line though is just because we don't see it listed so far doesn't mean it might not appear right.

[Jonathan]

Right.

[Ken]

Because you've got a few weeks there for them to get it in, we hope.

[Jeff]

Yeah, yeah.

[Jonathan]

So, all right, well, let's move along and I think up next is some Open SUSE Tumbleweed news. And it's yet another Wayland win.

[Ken]

Yes. And Jonathan, this week Marius Nestor wrote about the OpenSUSE project's latest announcement, which is OpenSUSE Tumbly Tumbleweed Rolling release Distro now comes with Wayland support for those who want to install and use the latest LXQT desktop environment. Now, LXQT 2.1 is the first release of the lightweight desktop environment to introduce an experimental Wayland session through the implementation of a new component called ready for it. LXQT-Wayland- Session. Doesn't that sound great? The Wayland session supports Several compositors, including LabWC Kwin Wayfire Hyperlin, Sway, river, and Niri According to Marius, at least since OpenSUSE doesn't offer a dedicated live ISO image with the LXQT desktop, you'll have to install it from the repositories on top of your existing environment or perform a fresh install. According to Marius, when performing a fresh install, you must choose the Generic Desktop option during the installation, then select LXQT by clicking on the Software section in the installation summary before starting the installation. After you have LXQT 2.1 installed, you then install LXQT Wayland Session manually and Marius provides a command in the article in his article that would let you just do that as a one liner. After installing the LXQT Wayland session package, you must also install a Wayland compositor. Now, the OpenSUSE LXQT team doesn't currently offer any recommendations for a default Wayland compositor for LXQT. They do suggest Kwin for the best possible support LabWC for hardcore LXQT users, Hyperlin, Niri, river or Sway for those who want a tiling desktop or if you like lots of desktop effects, wayfire, Marius chose LabWC and said it works nicely. Now I do want to remind you again, this is an explanation experimental Wayland system, so you may find some of LXQT's components are not Wayland compatible. If you want to try LXQT on Tumbleweed, then read Marius's article for the detailed installation instructions. So Jeff, should I switch my Tumbleweed installation from KDE to lxqt?

[Jeff]

Yeah, I'll be honest. I'm a KDE fan. I have used older versions of LXQT and various other ones and GNOME and kde. Just I like the custom custom customizations that you can do, even though I don't get really deep into custom customizing my desktop. But it's just kind of what I like, you know, like an old pair of old pair of shoes that just fit just right. You know, it's like, oh this is nice.

[Jonathan]

Nice thing about KDE is that somebody comes along every few weeks and resolves it for you. So you know, it's it fits just right, but it's nice and new and shiny at the same time.

[Jeff]

True, very true.

[Ken]

Yeah, I've got to move back into Tumbleweed and find out which version of KDE its own now.

[Jonathan]

Yeah, I'm probably not ever going to go to lxcute. Just not, not my, not my thing either. All right Jeff, let's talk PCI Express. This is kind of one of your, one of your babies, maybe in more ways than one. But what's, what's new? What's the, what's the published news on PCI Express 7 these days?

[Jeff]

So this story, as kind of Jonathan was hinting at, isn't exactly about Linux. But I did run across this article on Pharonix and you know, since we cover hardware once in a while, I thought this might be an interesting topic. So today we're going to talk about PCIe Express version 7, which is a new revision, but it's a 0.7, which means it hasn't been fully released yet. But it's still scheduled to be released sometime later this year, you know, sometime in 2025. They haven't given an exact date. Now, you know, I'm sure there's several of you thinking PCIe 7 I only have PCIe version 5 in my computer. Now, that's even assuming you have a very recent motherboard. You know, we haven't even seen PCIe 6 and it's already, it's already been out and ratified for a while. You know, I think it's about a coming up on a year in a few months. What the heck is going on? Well, to start from the beginning, the PCIe SIG group is a consortium of different people and groups who work on different versions of the PCIe Express standard. They're the ones that set, set everything and the goal is to have a doubling of data every time they release a specification. For example, if you, if you take full, if you take the full by 16 width of our current PCIe version 5, we can do 128 gigabits a second. Now that's theoretical throughput. So take that with a grain of salt. Like anything, there's a little overhead. You never get the true theoretical maximum, but you get an idea. PCI Gen 6, which has been released but we don't have any hardware for it yet, is set for 256 gigabytes a second. And PCIe Gen 7 and those of you who are adept at math have already calculated it out, is going to have 512 gigabytes a second again when using the full by 16 width. Now this sounds good, but doesn't answer the question about why we don't even have gen 6. Now the short answer is the verification of the standard. So once the standard is released, that's all well and good, but then there's hardware which is designed and created to be able to actually test the standard in the real world. Now to fully test the compliance, you need things to plug into the hardware and you need programs on things like oscilloscopes and other test devices which can test the compliance of the standard. But since nothing exists to test the software against, it's a process where two things are being tested against each other and the data verified to see if things are behaving like they should. Now, PCI Gen 6 and 7 brings also something some new signaling technology called PAM4. Gen 5 and below use signaling called NRZ or non return to zero. This is the kind of signaling that most people think of when they think of binary communication. The lowest voltage normally is 00 volts or close to. It represents a digital bit of zero and the high voltage represents a bit of one. There's a little more to it than that, but that's the basic concept. While signaling in NRZ can represent 1bit PAM4 in the same amount of time, assuming the same frequency can represent four bits. It does this by having four defined voltages instead of having two defined voltages in the signal. So if we keep a simple example and have a voltage range from 0 to 3 volts, 0 volts represents bits 0, 0, you know, 2 bits. Because it takes 2 bits to represent 4 numbers, 1 volt would represent the binary digit 01 and so on. Now, what this does is complicate the compliance testing. Not only does it have to be backwards compatible to handle the NRZ encoding of older cards and drives, but it also has to handle PAM4 signaling of the new cards and drives. So the biggest jump in compliance testing is from generation 5 to generation 6, since gen 6 is the first one with PAM4 signaling. Now, the PCI SIG group delayed initial compliance testing for a few months in 2024 for Gen 6. So the compliance testing for Gen 6 is going on right now. So we should see Gen 6 hardware hit the market soon, probably maybe in the fall or something. You know, it's not that far away. But Gen 7 is scheduled to start the compliance testing in 2028. Now, the PCI SIG group has delayed the Gen 7 testing until then, but they didn't state why. Now we're probably going to see Gen 7 devices, and I'm just guessing at this around 2030, because it takes about 12 to 18 months, possibly up to two years from when a standard's ratified before you actually see it. Now keep in mind that not much can fully saturate a Gen 5 slot at the consumer level currently. So like all our big graphics cards and everything, and you Know, some drives can kind of get close, but sustained, not really. So it, for the, for the average consumer level, we're not really running out of bandwidth on the PCI bus. Not yet anyway. You know, so I just want to give a little status update. I saw the article, thought it would be fun to talk about and kind of go over everything and you know, I tell you what, you know, the future for computing keeps getting better and better and you know, I just keep looking forward to it. So it's always exciting times.

[Jonathan]

Yeah. So one of the things that I was going to ask you kind of just touched on it. We can dive a little bit more into it. And that is like, what kind of devices does this even make sense for? Right. Like one of the things with PCI Express 5 I think is that hardly anything exists that could use that much speed. And so like the market, even in 2030, the market for things that can do like a full X16 of PCI Express 7, it's got to be tiny.

[Jeff]

Yeah, it's very enterprise industrial type stuff. So like networking, you know, things like that. If you can use some of the like. In Linux we've talked about how you can use networked memory for computing as well. So you could have big memory pools that you could have multiple servers working on so that you don't have to put all the memory into one server. You can have it into a central pool that all of them kind of share based on their load. So it's going to be basically drives, networking, memory things and even drives. It's not so much like a single drive, it would be a large array of drives and you could have striped or large banks of, you know, something that holds a lot of NAND memory that's really fast, or even some of the, like the CXL type devices that can really pump out some data.

[Jonathan]

And so this is, this is something that comes to my mind is we're probably going to see the ability to do more with fewer lanes. And it's not so much that, especially for consumers, there's probably not going to be very many x16 devices that can fully take advantage of it. But with this much bandwidth per lane, you could do a lot more with an X4 device. Oh, definitely. That seems like that might be something that we could go to in the future. You know, you get into this kind of this ludicrous speed and I don't, maybe GPUs become x8 by default instead of x16 and you see more x4 GPUs that are pretty Much fully performant.

[Jeff]

And maybe it even goes back to the SLI type or cross the crossfire.

[Jonathan]

Yeah, possibly.

[Jeff]

Where now you can have several slots that are by 8 or by 4 that are full speed because we have PCIe 5 right now. So PCI Gen 6 if you have a by 8 slot that's going to be as fast as your PCI Gen 5 x 16. And like currently right now the fastest card is a 4090 and it will just barely saturate a PCIe Gen 4. So it's nowhere near filling up a Gen 5.

[Jonathan]

Yeah.

[Ken]

Could you see anybody coming out with a PCIe connected RAID device that you just plug into your computer computer and slap the NVME zone?

[Jeff]

You could. It would have to be something kind of striped because you're still limited at the. At the memory cell level. There's a lot of caching and logic that goes in. And as even memory speeds increase a lot, there's a lot more pre fetching and things like that going on. Because the actual memory cell hasn't increased in speed that much. Yeah, it's, it's all the, the anticipation of what you might need so you can preload it. So as you're streaming data out of the device it looks like, wow, this is reading really fast. It's way ahead of you setting up reading cells and so it can just keep pre fetching all that stuff. So it looks extremely fast. But there's, there's still some device physics that are. Things are improving, but not at the rate that advertised memory speed is.

[Jonathan]

Yeah, makes sense.

[Ken]

So basically 2035, before we see any consumer devices like video capture cards.

[Jeff]

We'Ll see some drives. Whether or not they can fully use it or not, that could be something. Because remember when we had early PCIe Gen 5 drives, they were actually running at Gen 4 speeds. They kind of said, Hey, I support Gen 5 but I'm just going to kick down to Gen 4. Yeah, you know, there was a little.

[Ken]

Since that's all we've got.

[Jeff]

No, I mean it was a Gen 5 drive on a Gen 5 slot. But the, but to have the silicon prepared to actually communicate all the time at that speed, they didn't really have it. So you can put in a little logic that says okay, we're going to say we're Gen 5, we're going to talk a little bit at Gen 5. We're going to kick into Gen 4 for the actual data transfers and the rest of the chips can't support that speed anyway. So it doesn't really matter.

[Ken]

We're capable of gen 5.

[Jeff]

Yeah, well, we do enough for the advertising so we can check the box.

[Jonathan]

And then you mentioned oscilloscopes to be able to do some of this testing of this at this sort of data rates. That's not just any oscilloscope.

[Jeff]

No, no, there's some, you know, I, I know there's, I think it's about, and this is just advertised off the websites. I think it's about 100 gigabit is where they're currently at. And those types of oscilloscopes, they're big, they're expensive, they take like 220 type power or 240. They take a lot of electricity and they give out a lot of heat. Because to get that to speed up like that, it's, it's, it's takes a lot of power to do that. And a lot of times there's a lot of multiplexing in the background where you can't have one chip reading this. You have multiple chips and then they're all reading different sections of the signal and then it puts it all together in the background. So that's the multiplexing where, okay, I can't have chip A and chip, you know, read this signal. So I'm going to have A, B, C, D. And they're all staggered a bit so it's sampling at that really high rate. But by the time chip A is ready to sample again, all the other ones have fired. And then there's logic in the background. Okay, okay, we captured all this. Let's stitch it together and transform it into what your actual waveform is. And kind of a side note too, at that frequency you can't just like, oh, I'm going to take a probe and touch down. And the, the signaling dynamics at that speed are crazy. I mean the, the, the basic is the signal equation is really large. When you're running at say, you know, 1 megahertz, you know, most of that stuff in that equation, it's almost zero. We can ignore it. This is, this, this is almost like one. So we just add this in. It's, you know, it's, it's a lot of stuff simplifies at that speed. All that stuff matters. And it lot, a lot of considerations have to go in. Even, even the angle that you measure at and, and at that speed it's all like coaxed cable and things like that. Because, because probe tips generate too much RF and they pick up too much signaling noise.

[Ken]

How much noise did you add to the circuit by touching that probe too much.

[Jeff]

Oh, oh, way too much. And it, and yeah, it's. And depending where you touch and how you touch. So there's some high end probes at much lower frequencies that they have a gap between them. Say we'll say 2 millimeters. I'm just making something up. Well, they're adjustable and when you readjust that gap to touch your trace because the trace is slightly wider than what you have. So they, they're adjustable so you can open them up a little bit. You have to recalibrate them to handle that signaling because it changes. And when you touch down, you don't touch down at a 90 degree, you touch down at an angle. And there, there's a lot of specifics that go into it at that speed. And, and those type of scopes anyway, I mean they're wicked. You know, you're over a million dollars and then you have to have the software to run with it and I got to have the compliance package and I got to have all this other stuff to support it.

[Jonathan]

Yeah. You know, you're definitely up there in the realm of call for price.

[Jeff]

Yeah. Well, anybody buying those, it's probably not as much call for price as they've already corporate has negotiated.

[Jonathan]

Yeah.

[Jeff]

A price and discount and, and you have like a price sheet that the, the corporation has of. And you know, even big corporations, they're not. No, not everybody has these. They're pretty shared because of the expense and very few people have a true need for that kind of high end speed. But when you're, you're doing PCIe, you know, Gen 6 and not only are you reading signals, you got to be able to unpack it and you, you could do that manually, but it just would be very tedious and time consuming. So they make compliance packages that you can plug in and it reads the signals and says, oh yes, this is what it, the bits that are coming out. What it actually means and it can handle, you know, is your slew rate at the proper timing so that it, when your drive or your motherboard is saying, hey, what bits are here? Has it reached the state it should or is it still maybe climbing to its position?

[Jonathan]

Yeah.

[Jeff]

Give you bad data.

[Jonathan]

One of these scopes have to be able to over sample the speed that the circuit is running at to be able to actually get some of that fine data about slew it. Because if you're sampling at the same frequency that your circuit is running at, you do not get slew information. You just essentially get binary information and you actually need Some analog information to be able to look at slew rates and what the signal actually looks like.

[Jeff]

Yeah, the general rule of thumb is you want 10 times the sampling rate of your actual signal to be able to get a true accurate picture of your signal. Now you can get slower than that if you're not super particular about the slew rate. Not really trying to define your highs and lows and your slope, but when you're really, really digging in, 10x is kind of the, the industry standard rule of thumb.

[Jonathan]

Yeah, yeah.

[Ken]

So getting back out of the nitty gritty of the hardware, how quickly could I run the Linux distribution on Windows subsystem for Linux on Windows running in a virtualbox on a Linux system 2 gigahertz?

[Jonathan]

Well, I believe that was a convoluted segue to my last story. Yeah. And that is that in Fedora 42 we are going to have official Fedora images for the WSL for the Windows subsystem for Linux. And this is pretty neat for those of us that are Fedora fans that occasionally have to touch Windows. I don't have to very. I guess I do for customers, but I don't do hardly any on my machines. But this is something that hadn't existed and now has been voted upon by the Fedora Engineering and Steering, AKA Fesco, and they have approved the shipping of Fedora Linux WSL images. And so that means that starting with Fedora 42, you know, once it all percolates through the Microsoft Store and all that, you'll be able to jump on a Windows machine, go to the Microsoft Store and say install Fedora and it will give you a Fedora image inside of wsl, which is really pretty cool. And it's neat to have another option out there. Ubuntu has been there for a long time, I think, and there's been a couple of others, but less, less Red Hat based. Red Hat, sort of from that, from that line of, you know, from, from the Red Hat line. Few fewer of those distros have been available and now Fedora is going to be out there for a very bleeding edge distro on wsl. And so that's gonna be cool. I'm probably not gonna make much use of it, but it's nice to know that it's there. We need Rob here to talk about this one. Rob's the guy that would actually use this.

[Jeff]

Yeah, not me either.

[Jonathan]

Rob or David. Yeah, David. David might make use of something like this as well. He does floss weekly from Windows.

[Jeff]

I mean Goodness hurts my heart a little bit. It does tear my eye Nerds Listeners.

[Sponsor]

As we go into a new year, we all have a lot on our plates. There are backpacking trips across Europe to plan, personal best to crush in the gym, and capsule wardrobes to create good thing. Our sponsor, NerdWallet is here to take one thing off your plate. Finding the best financial products introducing NerdWallet's 2025 Best of Awards List your shortcut to the best credit cards, savings accounts and more. The nerds have done the work for you, researching and reviewing over 1100 financial products to bring you only the best of the best. Looking for a balance transfer credit card with 0% APR? They've got a winner for that. Or a bank account with a top rate to hit your savings goals? They've got a winner for that too. Know you're getting the best products for you without doing all the research yourself. So let NerdWallet do the heavy lifting for your finances this year and head over to their 2025 Best of Awards at NerdWallet.com awards to find the best financial products today.

[T-Mobile]

If you love your phone but not your carrier, just switch to T Mobile. You can keep your phone, keep your number and we'll help pay it off up to $800 per line. You can also use our savings calculator to compare our plans and streaming benefits against Verizon and AT&T. So switch and keep your phone, keep your number and keep more of your moolah. @t mobile.com up to 4 lines via virtual prepaid card.

[T-Mobile Disclaimer]

Allow 15 days qualifying unlock device credit service port in 90 plus days with device into eligible carrier and timely redemption required. Card has no cash access and expires in six months.

[Jonathan]

All right, let's move to some command line tips. We're gonna let Ken go first and he's still on a pipewire kick. I do believe I've got quite a.

[Ken]

Few to go through. Still though, as I dig into them deeper, I'm finding I may end up having to diverge into some of the other commands pertaining to audio, MIDI and video. But this week I am covering the PipeWire Device Reservation Utility. Or you can just simply type pw rs r e s E R V E. Now you can use the ALSA Mixer. Thank you Jonathan for covering that in episode 39. And he also covered another one, a play in 120. You can use these commands to identify your audio cards or devices since this is what you're using the PW Reserve with now. If you have just a single card. It should be numbered zero. Now let me go ahead and bring up my screenshots here so you can see what I'm talking about. Here we go. And let's go back to the beginning here. I was going through them earlier, make sure I had them all. But with the first screenshot, it's showing me with the two standards, Dash dash help and dash dash version options. Dash dash Help, of course, gives you how you would format the command on the command line and what options you could use with it. In the case of PW Reserve, in addition to the help and dash dash version, you have dash dash name or dash N. This is the name you use to reserve the device you're going to talk about. It could be Audio zero, Midi zero or Video zero, or one or two, depending on how many devices you've got for that type. You can also give a application name that you want to reserve, which is. But the default here is PW reserve for the application name. You can set a priority, it defaults to zero. You can release by doing a dash R or dash dash release, basically that requests a release when the device is busy. And you can do dash m or dash dash monitor. Now this will only monitor the device, it doesn't try to acquire it. But if we move on to screenshot two, I've brought up my Pulse Audio volume control with the configuration tab shown showing all my audio devices. And you'll see on the terminal that I've got it displayed. Also for those of you all listening, it's PW reserve space dash in space capital A for the audio zero. And then screenshot three shows the response from that and how I did actually get it to come up it with the first command it just ran and said audio device, Audio zero is busy. Use dash R to attempt to release, which I did right after. And it did release it. It gave you information that it was requesting to release on audio zero reserve available on org.free desktop.reservedevice1 audio zero. If you'll look follow the link I've got on PW reserve. You'll see that it has information about how they standardize the reservation names. But if you look down at my Pulse Audio volume control, you'll see it's only listing three devices. Now let's go back to the second screenshot for a minute and you'll see that the one for my HDMI output to my Sanyo Table TV is there, but now it's gone. Audio zero reserved that pulling it out of a pulse Audio so that I could use it and say jack. Then we move on to screenshot four, where it shows me doing a signal interrupt or Control C and it releases audio and it comes back. And I did some playing around and I found out that the HDMI is my audio one. The my Behringer preamp is a PCM 2902 audio codec is audio two, or excuse me, audio one. It's 01 for the Behringer. Two is the family 17h 19h audio controller. And three is my C920 Pro HD webcam, which, if you'll notice here, I've got it actually disabled in my Pulse Audio configuration so that it's not used by anything at all.

[Jonathan]

So I'm wondering in thinking about what this is going to be used for. If you didn't want to have that disabled, but you wanted to have it just completely ignored, could you just run PW reserve, like as part of your boot and make that device just totally go away and disappear and never show.

[Ken]

Up again, or in a script that you're running before you set up?

[Jonathan]

Yeah, seems like that could be one of the uses of this.

[Ken]

That would be very interesting because I'm still having to. I tried setting it up to auto start with KDE in a terminal, but it's not recognizing the last week's command. The pulse pipe wire pulse Nash pulse. So I'm setting that up in a separate terminal that I open up and run. And you can see where I was opening and closing it as I was playing around both with it and this at the same time.

[Jonathan]

Yeah, interesting. All right, very cool. So up next, Jeff, you've got a command I'm not familiar with here. What is this?

[Jeff]

Lpq, which is going to be pretty basic here, but stands for Linux Printer queue. And it's just like it sounds, it'll let a user see what's in the printer queue. So there's. There's not many options. This is a pretty basic command. There's an interval command which will or option that will let you have a continuous query going at whatever interval you specify. So you can monitor your printer queue at whatever time frame you desire. There's a dash L which will give you a long or more verbose version of the output, a dash U to specify a different username. So if you want to look out, look at what specific user has in their queue. Dash H is for specifying a remote server so you can look on other servers to find out what's in their queue. And you have like a dash A that reports all the jobs or the jobs on all the printers connected to your machine. Of course, you know, you can string these together so you could say, oh, I'm going to look at all the print queues on this remote machine or on the user on this remote machine, that kind of thing. You know, if you have more than one printer connected to your machine, the Dash P option will let you specify a specific printer. So it's the basic, you know, you, it's let you show what you, you would kind of think your options would be. And that's basically about it. There's a link in the Show Notes, the man page for lpq, so you can do more reading and keeping up, keep up on your printing in line. But I pretty much gave you everything the man page has. So now, now if you're in the console, you can keep track of all your printing.

[Sponsor]

Fun.

[Jonathan]

Yeah. All right, I've got one. It's not exactly a command line tip, although there's a lot of command line work that you will probably use if you go to, if you go to do this, and that is obs. But not the Open Broadcasting System, the Open Build Service. And one of the projects that I'm a part of, we were kind of struggling to figure out how to do Linux builds. So we've got, you know, we've got Ubuntu builds running now through a PPA and we've got Fedora builds running through a Copper copr. But then there were some other, some other places that we wanted to be able to build stuff. And one of the ones that really was getting us was doing builds for the Raspberry PI, because particularly the 32 bit Raspberry PI is kind of its own weird little thing. It's not ARM V8, it's not even ARMV7 technically. It's like ARMV6, which is a sort of obscure 32 bit hard float ARM system. And so trying to find a service that will do builds for you for that in particular was a challenge. And come to find out OpenSUSE runs the open Build service and they will do builds for Tumbleweed, they will do builds for Raspberry PI OS and they will do builds for Fedora, because two of those three things at least are RPM based. I think that's why they do those. And then nobody else was stepping up to do Raspberry PI builds and so they, they stepped up to do that. I think there may also be a couple of other targets that you can, that you can hit with it, but it's a super useful build service. Let's see. Arch, Kiwi, Simple Image, appimage, Flatpak, Debian, rpm. So basically everything, everything you could think of, just about that you would want to build for the open build service will do it. It's free for open source projects. You can jump in there and make it happen. It was a little bit easier to work with than the PPA service.

[Jeff]

Yeah.

[Jonathan]

If you need the. If you need to build and distribute packages, open build service is great. So add that to. And again, that's sort of niche. I know, but add that to sort of your, your tool. Bag of tricks. If you're ever in a position where it's like, man, I just, I need to get builds done for the Raspberry PI. Well, there you go. They're the guys to do it for you.

[Ken]

So does this mean you're going to cover one of the other open SUSE tools next week?

[Jonathan]

Which one are you thinking of?

[Ken]

I'd like to see you try to do a cover of Yost.

[Jonathan]

Oh yeah, yeah. Yet another. What does that stand for? Something tool. Yet another setup tool. Yeah, yet another setup tool. We could look at that. I've never played much with Yost. There are a couple of other tools out there that are similar to it.

[Ken]

The one that really sounds interesting. Interesting as Kiwi.

[Jeff]

Yeah.

[Jonathan]

I can look into OpenSuse's stuff, see if there's some more of those command line tools that we want to mess with. It sounds like fun. See what we can do about that. All right, well, we have done our news and our command line tips. I'm going to let each of the guys get in the last word if they want it. We'll start with Jeff. What do you have?

[Jeff]

I have a link in the ending notes. So sorry, no poetry corner today. Just wanted to say for all you gamers out there, glorious egg roll, Proton 923 was released. It's, you know, updates wine to the bleeding edge, you know, fixes some battlenet broken updates, update VDK or DXVK, the latest git updates, VKD3, Proton, DXVK, Dash Nvapi, you know, a lot, a lot of upstate stream fixes. And the link in the show notes also will show you the other Proton fixes. There's you know, some BioShock Remastered fixes and launcher fixes. So for those of the. Those of you that enjoy Proton gaming, glorious egg rolls out there. And with that being said, I hope everybody has a great, great Linux week coming up.

[Jonathan]

All right, Ken.

[Ken]

Well, as we're moving into tax season. I just want to remind everybody, start backing up.

[Jonathan]

Oh, taxes. Thank you for that.

[Jeff]

I said have a happy week. You kind of. You kind of. Yeah, kind of rained on that parade there, Ken.

[Jonathan]

I'm just saying.

[Ken]

Oh, you still got three months to before you have to get it in by.

[Jonathan]

Yeah, I'm already starting to get emails though, messages and things in the mail.

[Ken]

I actually was downloading some of the tax forms I needed last month.

[Jonathan]

Yeah, my wife just today was like, we need to make a tax folder for this year. Yes, we do. All right, thank you guys for being here. I will plug Hacker so behind my Yaquik console, which I have popped down because I've been messing with Firefox, there is the Hackaday website. You can find my stuff there. Plus weekly it records on Tuesdays and goes live on Wednesdays. And then we've got the security column goes live every Friday morning. Enjoy both of those things very much. Would love you to come check it out. And if you've not checked out Club Twit, someone sent me a message the other day like where's the video feed for the Untitled Linux Show? Well, that is a part of Club Twit and you should check it out. It's about the price of a cup of coffee per month and it's the way to support this show and the Twit network. And we sure appreciate those of you that do. And we appreciate everyone that watches and or listens to the show. Those that get us live and on the download. And hey, we will see you next week on the Untitled Linux Show. Foreign.

[T-Mobile]

If you have a locked AT&T phone, we're here with bolt cutters. T Mobile will help pay off your locked phone and give you a new 5G phone for free. All on America's largest 5G network. Visit t mobile.com carrierfreedom via virtual prepaid.

[T-Mobile Disclaimer]

MasterCard in 15 days. Free phone up to 830 via 24 monthly bill credits plus tax and a 10 device connection charge. Qualifying port in trade and service on Go5G next and credit required. Contact us before canceling entire account to continue bill credits or credit stop and balance on required finance agreements. Do you have no credits? And if you pay off devices early.