LINUX Unplugged - 354: Microsoft FINALLY Gets It

Episode Date: May 20, 2020

Windows is getting more competitive by adopting core Linux features, so we cover the latest Linux-inspired additions to Windows. Then review the new release of Pi-hole, sort through recent PINE64 updates, and read your feedback. Special Guests: Alex Kretzschmar, Drew DeVore, Neal Gompa, and Philip Muller.

Well, Chris, did you hear? Someone posted a native in-kernel patch set for WireGuard to OpenBSD. Is that possible? Yeah, you know, actually, they've been working with Upstream with Jason. They've learned a lot from the Linux kernel implementation, and they've got the right licensing in place. So, yeah, I mean, it hasn't been accepted. I don't know if it ever will be,

but it actually might be simpler than Linux because it turns out OpenBSD already had good crypto. Hello, friends, and welcome into your weekly Linux talk show. My name is Chris. My name is Wes. Hello, Wes. It's episode 354, and I think it's too big of a show. Way too big. We're not being very responsible today. I know. We needed to get one of those vehicles behind us with the sign that flashes that says, oversized show. Just go around us on the internet. There's too much to cover today. We have a bunch of great community

news. It turns out this is one of those weird weeks that only 2020 brings us these days, where Microsoft is going to dominate a large portion of our community news section. Say what? I know. It's not just going to stop there. We have a bunch of other things to get into, including Philip is here from Manjaro. Hello, Philip. Thank you for joining us again. Hello, hello. Hello, hello. So stay on the line there, Phil, because we're going to be talking about some Manjaro hardware here in a moment. But before we go any further, I got to say a big hello to Drew and Cheese, who are also here today. Hello, guys.

Hello. Hello. Well, hello, gentlemen. Cheese, you're looking fabulous today. I like this new beaded sweatshirt you've got here. Did you hand bead this thing? Hand beaded and hand sequined. I'd never seen a hoodie made out of beads before.

It's really pretty impressive. This is one of those times I wish we were a video show. And of course, a huge time appropriate greetings to that mumble room. Hello, virtual lug. Hello. Hello. Who let the eagle in the lug? Oh my gosh. Ridiculous. Well, I didn't want you to get lonely. Yeah, it is nice to have a pet eagle. I got to say, it is pretty, pretty sweet. We have some strange news to start with. Everybody take a moment, prepare yourselves, because today is Microsoft Build, and we have a Build blast of Microsoft news for you.

It starts with a Microsoft executive recently saying, we got it wrong on open source. Microsoft's Brad Smith acknowledges they were wrong in past remarks made by Balmer and other executives. Brad Smith has served as the president of Microsoft since 2015. He held a virtual talk slash fireside chat skis. During that, there was a quote that

Michael Arbel at Pharonix grabbed. Microsoft was on the wrong side of history when open source exploded at the beginning of the century. And I can say that about me personally. The good news is that if life is long enough, you can learn that you need to change. Today, Microsoft is the single largest contributor to open source projects in the world when it comes to business. When we look at GitHub, we see it as the home for open source development. And we see our responsibility as its steward to make it a secure, productive home for developers. How does that make you feel when you read that? Do you feel like they get it? They actually get it this time? I mean, you know, Brad Smith is high up there, and it does seem like their workforce gets it.

So at this point, the executives sort of have to play catch up. You know, all the best talent that they want to hire, those are folks that want to be involved and are involved in open source. And you heard that little business reference in there? I think really what they picked up on is, hey, this is working for our bottom line. Yeah, indeed. So Build, like I said, happening right now, Microsoft's annual developer conference. And there's been a couple of bombshells that applied to the Linux world.

Let's start with the biggest, the one that made me go, holy crap. There was some hints going around about this that a developer on Twitter named Alex happened to mention and notice and took some screenshots of some weird DirectX stuff that seemed like it was being baked into WSL. But he didn't know what was going on behind the scenes. And today we learned that Microsoft, Satya Nadella himself up on stage actually, announced that WSL, the Windows subsystem for Linux on Windows 10, will include GPU compute support and full GUI application support. And there's even some DirectX magic happening here to support certain DirectX calls on your WSL applications. This is full-fledged graphics supports. And their initial prototypes right now are even using Wayland as some of the windowing technology to make it possible to run full-fledged Linux GUI applications with GPU acceleration via WSL. They're going to beat us to Wayland on the desktop.

Oh, no. It turns out how they're accomplishing this is essentially making this technology on par with what Windows applications can achieve, graphics performance-wise. We'll link to the developer post where Microsoft explains all of this. But essentially, for a while now, Microsoft has been baking in para-virtualized GPU support into the core of Windows and Windows drivers, including requiring all of the OEMs to support this. Think about this for a moment. This has been a long-term project, actually in part to support HoloLens, of all things. And launching with WSL2 and WDDM version 2.9, which is part of the Windows Display Driver model,

they are going to expand the reach of GPUs to Linux guests using this pair of virtualizations technology that Windows itself is already using. So it's the same pipeline Windows is using for graphics rendering now being opened up to the WSL. And it's one part Windows software, but it's also one part Linux kernel software. Yeah, they've made a new kernel driver, DXG kernel, brand new, and it basically exposes a little device, slash dev, slash DXG, to the user mode Linux running in your virtual machine there.

And it's got an iOctal that basically mimics the native WDDM stuff that happens on the Windows side, and it sits there on the bus between the virtual machine and the host and just sort of forwards those calls along. So once you're in Linux, I mean, you just get the full capabilities that you would get in Windows. And there's not a whole bunch of throttling.

It's not like Linux has some sort of second tier access to the hardware. It competes just the same as any native Windows application. Yeah, like Bruce in the chat room puts it. So now you can play your Steam games using Proton on Linux on Windows. And as Jill very accurately and acutely puts it, there's really no need for VMware on Windows to run Linux with the full GPU anymore. Applications that are running inside of the Linux environment have the same access to the GPU as native applications on Windows. There's not a partitioning of resources between Linux and Windows or limits that are imposed on the Linux applications. None of that exists. The sharing is completely dynamic

just based on what application needs what. Also sort of interesting from the open source perspective, not really the open source perspective, but just the targeting Linux at Microsoft perspective, is they've added stuff like DXCore and D3D12 to Linux. So this is like DirectX driver code now running on Linux, but it's compiled from the Windows code base, the same code base that they make the DLL for Windows. That's what they make the.so file for Linux on. Although it seems that their kernel,

while it shares some of the same names as Windows components, the kernel driver they've created is a clean room implementation that will be GPO. Nice to see. Now, of course, you know, we mentioned Proton and I had the same thought, but probably what a lot of this is targeted to, especially right now before the full, you know, the GUI stuff comes into WSL, well, it's machine learning. Yeah, those kind of jobs and stuff like that. And I think it's also about reassuring certain OEMs that if you wanted to continue to invest in DirectX, you've got still a Windows Avenue here as long as they're running it under WSL. And they do address the question around Vulkan in this post, and they say they're still exploring how to best support Vulkan in the Windows subsystem for Linux.

Although it seems like they're looking at basically adding stuff to Mesa that implements OpenCL and OpenGL and some of those technologies on top of DX12. So this is like the reverse of the shims that we've been using to get DX implemented on those. They just flipped the script on us, Wes. What? On one hand, it's really nice to see that Microsoft is giving more prominence to Linux as a whole, even in their conferences. is giving more prominence to Linux as a whole, even in their conferences. But my principal worry about this is that it'll continue to suck the oxygen out of further advancing the Linux platform in favor of just, well, Windows is good enough, so we'll just do that.

And then all the crap that we have to run on Linux, we'll just run on Linux on Windows. And it's a very, very dangerous and bad path for the success of free and open source computing. Because if people in large droves actually do embrace this attitude, like I don't want to bring up an old dead meme, but it's embrace, extend, extinguish in a very, very real way. Yeah, I see others in the chat room saying the same exact thing. They're echoing the embrace and extend extinguish. I think it's the other way around, though, is I think this is a dramatic reaction to trying to shore up what has been a pretty

lacking aspect of their quote unquote workstation desktop OS. And I think they've looked at what sold a whole bunch of MacBooks a few years ago when a bunch of people bought MacBooks for development in the cloud, and they're looking at that going, well, a rich command line, a good terminal, and now moving towards the future, machine learning and being able to use the GPU for compute is a big part of what's going to make the future workstation for developers. I think they're making that calculus as a reactionary move. And I think the problem here is, because of the nature of the GPL, because of the nature of Linux as sort of a unifying technology platform

that anybody can kind of build on and then bring their final value to make it a complete product, I think that's an unstoppable force. And it eventually will continue to consume the desktop. It'll just be a 10, 15. It's the multi-20-year, multi-decade saga. But when you zoom this thing out and you go 5, 10, 15 years down the road, it's not a Windows future. This is a stop-the-bleeding kind of move right now. Windows itself and the NT kernel has a limited long-term potential. Windows 10 is sort of the big last hurrah for that platform. And in a way,

what they're essentially doing is just making Linux the de facto runtime. If you target Linux, you can run it on the cloud, you can run it on Android, you can run it on Windows, you can run it on Linux desktops, you can run it on ARM, you can run it on the cloud. You can run it on Android. You can run it on Windows. You can run it on Linux desktops. You can run it on ARM. You can run it on everything that's in everyone's pockets. And I think that is just too big. It's bigger than any one product. It's bigger than any one company. And this is sort of just attempting to become at least competitive with the kind of workstation you get out of the box with Linux or Mac OS. It's almost Microsoft sort of just selling, you know, fancy developer tooling on top of Linux. Like here's this whole OS, but really it's just a wrapper around Linux

to, you know, smooth that experience and let you play games on the side. That's it. We're taking what you know, we're taking what you like, and we're just going to shave off some of those rough edges. Why even concern yourself with leaving the Windows experience? But I don't think it will supersede Linux as a whole because it's just utilizing part of Linux, which has already been so successful. That doesn't mean it's not extremely clever. And I think the company that's most at risk is Apple. This is clearly going after the developer market in a

very aggressive way. And I think it's going at the heart of what made macOS good, because they've taken Windows 10 now and they're doing this iterative update process. They're trying to make it a stable OS with WSL, with all these new features. It's not us Linux users that should be concerned. It's Apple. If this was really the way that, you know, Microsoft was going towards as like they're trying to stave the bleeding and trying to recover a little bit and attack Apple more. Why not invest in wine? Why not build up that part? Like essentially what Valve has been doing to beef up the GPU and the hot pass for supporting video games.

Why not do the same for the rest of the stuff that Microsoft tends to care about, which is productivity and things of that nature? I've been wondering about that myself, if that would be something we start seeing them do. that Microsoft tends to care about, which is productivity and things of that nature. I've been wondering about that myself, if that would be something we start seeing them do. But I think instead, you'll see them just start targeting the platform. I've recently, for the first time, used Teams on Linux. Have you ever tried that? It's not bad. It's actually not bad on Linux. They've done a decent job of creating the Teams application for Linux. It lacks some of the features, like background blur. Right. And it did take them a while to add that support. But you saw it there as a bug,

and I think that's part of their process now, right? They have ways to get feedback about products. And when they saw that, oh, yeah, there are a bunch of Linux users out there who maybe want's not the right word, but need to use our software and want a native-ish version, let's make it for them. I have this chrislass.com slash consulting where I'm doing the podcast consulting for folks. And I had a client who is interested in working with me for a serious chunk of hours, but the requirement to work with them was I had to use Teams because everyone on their team is using that. And I was like, um,

is using that. And I was like, um, okay. I didn't realize that until like the call was supposed to start. And so I was like, I bet this is in the Manjaro repos. And so I did, I did a quick search. There was Teams. I did a yay-s Teams and hit enter. And within 30 seconds, I had Microsoft Teams installed. I clicked the link and I selected my audio source and I was done. And I've used it three times since then and it's worked perfectly fine. I think the only thing I can say to that is, yay. Yeah, exactly. There's also a couple of other Microsoft things before we move on to Manjaro, which I want to talk about next. They also officially today released the Windows Terminal, which was announced at Build 2019 and has become official with 1.0 at Build 2020. This is their open source terminal application where you can have in one tab the Windows command

line and then in another tab any Linux distro you want. And you can have four or five tabs open with different Linux distros in each tab. It's a very well done piece of software. It's something Wes and I were very impressed with when we tried Windows 10. I hope it doesn't get too good or it might be better than our native terminals. I mean, it's already graphically accelerated. It's got a ton of customization options and, of course, first-class emoji support right there. You know, it's one of these new breed of sort of Windows apps that have been reimagined in this Microsoft loves open source future. So I imagine there's going to be a lot of new workflows going on. Well, you and I met a couple of the individuals behind the project at WSLConf,

and the people behind the project, from what we could tell, they came to Microsoft after this new open source kind of attitude. So they came with this whole fresh perspective. I mean, I've known people who've worked at Microsoft for a very long time, because it's here in my backyard. I've had family and friends that have worked there. And there was an old way of looking

at a Windows first world, and then there's been a shift since then, and they joined the company since that shift, and they came in with that open source first mentality, and then they built this terminal. And it's just a fascinating story. But last but not least in all of this, because there's actually even more, Microsoft announced a built-in package manager for Windows now. And this one's interesting.

The company today announced the Windows Package Manager. I knew they would do this. I felt like Windows needed this. It's a new native package manager that'll be built into Windows 10. With the new WinGet, or Winget if you want to be silly about it, command, Windows Package Manager automatically gets the latest version of an app you want to download, validates its authenticity, I mean you want that too, and installs it on your machine. So as an example, you can install the latest version of that Windows Terminal we were just talking about just by running winget, one word, install terminal. And that just pulls down the latest version from the Microsoft Store and sets it up for you.

Yeah, so when I read this at first, I was like, oh, so it's a command line front end to the Windows Store? Womp womp. That's not very exciting. But no, the company will also allow users to download apps from third-party repositories, which I think is the key thing to actually making a package manager useful. Yes, right. It's a little more generic tooling in that sense and not just tightly tied to their proprietary infrastructure. Yeah, a command line front end of the store is boring, but a legitimate package manager that also can pull from the store? Okay, that's interesting. The Windows Package Manager supports Windows 10 version 1709,

so you need to build 1709 or later, and it will come pre-installed with Windows 10's desktop app installer once that reaches version 1.0. It seems nice for automation, too, you know? We'll tie this into some of your scripts and setups, and instead of relying on third-party tools, why don't you just do it all with the Windows stuff built in? This speaks to the larger point that you made

about Microsoft looking at the landscape at macOS, maybe. What sold a bunch of MacBooks? Well, Brew sold a bunch of MacBooks. So let's just copy Brew and have our own thing. I think that's it. I think what you're seeing here with this package manager is Microsoft looking at Linux workstations and Mac workstations and saying,

what works, how do we do a modern workstations and saying, what works? How do we do a modern take on that and build that into Windows? And where I sit as a Linux user, it feels like it takes four freaking ever for this stuff to land in the operating system that should have been there for years. But to Windows users,

this stuff's coming at a rapid pace. They said it from a totally different perspective. Definitely big changes. I mean, it's interesting just from the velocity aspect to see how much stuff Microsoft's able to get done. And in the new Windows 10 world of, you know, frequent and smaller updates, actually push this out to people. I think as Linux users, we're going to have to get comfortable with a new reality where there's a lot of people on Windows that are playing in our space. reality where there's a lot of people on Windows that are playing in our space. And we need to be inviting and understanding and help them when they're ready to make a transition to a fuller Linux system. Because if we're lucky, we can set a good example, we can make it look like a friendly

place, and we can encourage them to move to a richer, fuller Linux ecosystem. I mean, that's just it. All right, you got Windows from your, you know, enterprise-supplied workstation at work, and oh, hey, you were able to start playing around with this Linux stuff. Well, now when you go home, maybe you want to try putting some stuff on a Raspberry Pi and doing home automation. We've got the tools for you. part of the greater community and further the cause. And I think it won't be for everyone. I mean, every user of the system won't be thinking like that, but we'll be able to get some of them. All right. I know that's a lot of Microsoft news in a Linux podcast, but today's build, and that concludes our build blast. And we move on now to some new hardware, the Manjaro Book AMD Ryzen, running a Ryzen 5 series, which is a 3600, I believe. Phil, help me sort this out, because I was thinking when Manjaro announced a Ryzen system,

it was going to be the 4000 series, but it looks like we have maybe an interim release hardware here. Well, it's like you have a normal desktop CPU flagged into a laptop, so you can even change the Ryzen later on if you have a normal desktop cpu flagged into your laptop so you can even change the ryzen later on if you have a different cpu simply swap it to a powerful one oh so it's this is like one of those workstation killer laptops something like that is a standard m4 cpu and you can simply swap it out there's a powerful uh cooling system so it works So we have a smaller version, but you can also put it in Ryzen 9 up to 3,900 12-core. So we've got a whole range of systems here. I mean, we've had you on recently talking

about Intel-based systems. Now we have an AMD Ryzen system. There was talk recently about a Ultrabook-style Manjaro system that would also be Ryzen-based. So how do we keep all of this straight now? Well, we have several vendors. So this one is the commuter shop from Belgium. And they said, hey, we are looking also into the Ryzen plane and said, okay, let's go for that. And we said, yes, why not?

And we promoted for them. And so are they doing the bulk of the heavy lifting for getting the OS to run on there? Or is it because it's at this point, the hardware is so well supported, there's not a lot of heavy lifting to do? Well, we test with them. The main developer of their shop is also using Manjaro.

So he is testing with us and we give him some pointers to get his hardware supported. So yeah, it's a win-win situation that he is so active in distributing Manjaro on his machines. So I see kind of a split now in the strategies because some distros have one laptop and it's like the KDE Neon laptop. There's one machine, you can get it.

It runs there. When they update it, they update it infrequently, but it's that one box. Manjaro seems to be going the approach of working with multiple vendors with multiple different laptops. Are you concerned at all about that getting confusing? Normally not. We have Tuxedo laptops where we have the complete branding. Then we have Manjaro Computer from Belgium, which is also doing some kind of branding for us and having more vendors than only one.

You have more options. And with this vendor from Belgium, you can sell also to the US and Canada without the taxes. So we get a discount of 21% of each automatically. Oh, that's pretty significant, actually. So you are, in a sense, trying a different model here. And I couldn't be more for it. I think it's great to see some just stick with one piece of hardware and they iterate it infrequently and there's that choice. And then you have the Manjaro approach where there's several vendors which might work better for different customers with different machine options.

I absolutely, for a very long time and long time listeners of mine will know this, I was a fan of the Battlestation laptop. I would haul these hunkers to my clients, and then I would just plug in, and I would have the power I needed of a desktop in a laptop. These days, and I think kind of like yourself, I'm more of portability first because, you know, I end up on a plane or I'm carrying a machine around home and work every day. So it sounds like there'll just be multiple options for different types of users. Well, Phil, I got you here. Let's just do a quick update on the Manjaro project itself. I saw a bunch of updates come out recently for my workstation. How are things in the land of Manjaro generally?

It's pretty good. We're working on getting the updates for the 20.0 releases out. So the first point release went out there as well. And we found some issues, we fixed them, and we hopefully will also update the new KDE packages as they come out soon. I love that. I'm running Plasma on my Manjaro workstation, and I just love it. I told the story recently on the show, but it's so nice having one package manager for everything. Even my stupid printer drivers were in the AUR. And then you combine that with the curation that happens at the project level on top of Arch. I was initially skeptical, but as an end user, I have just absolutely loved it.

And I'm going to keep an eye on these hardware choices. Will you let us know when the 4000 series Ryzen box gets close? As soon as some vendor has a 4000 series, then I'll let you know, of course. Oh, I'm excited already. Yeah, I mean, everybody is. Everybody is. And truthfully, in our testing, the Intel 10th generation CPU-based systems are also very competitive. So it's a good time.

There's lots of choice. Well, thanks for coming on, Phil, and updating us on this. It looks like a serious machine for anybody who needs workstation performance from a mobile rig. Yeah, and the 10th generation of Intel, we will have some new laptops soon, also from the UK and Germany. So stay tuned for those as well. Thank you, sir. Now, on the other end of the price spectrum, but also well-supported by Mangero, are the Pine devices. And recently, as in since the last show, that's my definition of recently. I'll take it.

Yeah. The Pine tab had a pretty significant update with dates on shipping and also other new features. Unless I missed this, and Cheesy, maybe you can correct me. new features. Like, unless I missed this, and Cheesy, maybe you can correct me, the last time I saw information about the PineTab, I didn't see this keyboard attached with it. Well, yeah, they've always had the idea of a magnetic keyboard in play for the PineTab. I think it was in November when they were initially kind of talking about it, and there were dev units out there. I know Marius from UbiPorts had gotten one.

The magnets on the keyboard case weren't very strong. And so they were having issues with those. So I think they re-engineered where those were placed and used stronger magnets. I mean, it seems like a great little tablet to get it kicked off. There's already some videos out there of the pine tab running Ubuntu touch, you know, from UbiPorts with the five, six kernel in it. Lima graphics drivers, seems like a solid little machine, 6,000 milliamp battery, a full size USB 2.0 socket does have two gigs of DDR33 ram which is a little low but it's definitely not a killer for a small tablet like that 10 inch 720 ips display 64 gigs of emmc flash storage but you can expand this and i thought that this was super cool so they have an optional m.2 adapter that you can open up apparently with one screw you can take the back off the tab uh toss in this m.2 adapter that you can open up. Apparently with one screw, you can take the back off the tab,

toss in this M.2 adapter. So you can add in your NVMe, SSD, even an LTE module. We're talking about LoRa modules, a lot of interesting stuff. Yeah, that really stood out to me, Cheese. It just because, you know, I mentioned a little bit like some of this is for aimed at personal users, but some of this is aimed also maybe at commercial applications like LoRa support, which is that low bandwidth but long range wireless protocol. So wouldn't a PineTab, especially at that price point, be something just flash on your like commercial thing, give it to your employees, they've got it there in the truck with them when they need to go service something? That'd be awesome.

Precisely. And the one thing that does seem to be kind of a hiccup with that adapter though, is that you can populate it with two modules. So you can populate it with the LTE module, for example, or a LoRa module, which isn't quite out yet and say an SSD, but you can only use one at a time. But I mean, you got to think that this is a $99 price tag with a $20 keyboard upgrade. That's really a pretty awesome deal. There's a lot of cool stuff that actually came out of this update. So the Pine phone itself, they're going to be adding a Qi wireless charging that attaches via the pogo pins. They're exploring adding another 5,000 milliamp battery to the back of the case to give you more runtime. The keyboard that they've talked about for the Pine phone has kind of been put on hold,

but that's going to be coming around eventually whenever this kind of COVID thing has died off. And I think that really comes down to they need some in-person visits, which you really can't do right now with these factories. So until that opens back up, you know, they won't be kind of going down the keyboard route for a little bit. A couple of things jump out at me. First of all, $20 for a magnetic backlit keyboard is crazy. And when you compare that to the $300 entry price for the new Apple iPad keyboard, the other thing I really like to see is adding Qi wireless to the Pine phone is a big deal because there are millions of Qi wireless devices already out in the market made

by some of the biggest manufacturers in the mobile industry. You know, they've said that there's going to be compatibility there. So any Qi wireless devices that you already have or charging mats that you already have will work. It's a genius idea that they added the Pogo pins so that you can easily, you know, make these additional devices that you can attach to the Pine phone. As far as software, I tried the latest UbiPorts. I still don't have a SIM for it. It is getting better. In the video they released, it looks like UbiPorts was designed for a tablet. I think that really kind of crosses that convergence area where everyone's been talking about. The tablet itself is also supposed to

support display out. I don't know if that's USB OTG, how exactly they're going to handle that, but it should be interesting. I got a question for you. You said that they have the same quad-core A64 system on a chip. Is that just essentially the Rock Pro 64 in there? Well, that's simple. They use the same CPU as in the normal Pinebook, the original, the 1080p. same CPU as in the normal Pinebook, the original, the 1080p. So you have the same CPU in the phone, in the Pinebook original and in the tab. And with that, you have the possibilities to have the same hardware and us as vendors of the software, we don't have to release multiple

kernels. So it's easier for us to have one kernel for the Pine devices and to work on the phone on a tab and on original. Yeah. I bet as a developer, you're very pleased to see that. I want to be the guy here for a moment who just says, you know, set phasers to stun on your expectations. It's not going to be a performance powerhouse. It's not going to shame the iPad Pro in that regard. But in price, customizability, tinkering, future expansion, and potentially your preferred workflow, this could be an unmatched value. I am very excited to see where they take this. I mean, it's just the beginning, right? I mean, this will be the first tryout, and I'm sure it will evolve as a platform as it keeps going.

Very cool. All right, well, let's do a little housekeeping. We don't have a lot in this week's episode. I do want to mention that we'd love to have you join us live. We have a great attendance today, 33 people in the mumble room right now as we go. You can join us on Tuesdays over at jblive.tv, and we convert that to your local time zone with the assistance of robots at jupiterbroadcasting.com slash live.

And that's good because you don't want us doing it. I'll mess it up every time. Every time. We also have the Luplug happening every Sunday. This is how I know some people skip housekeeping because they don't know about this. They show up in the Telegram group and they're like, isn't there some sort of lug thing? Yeah, the lup lug. Sundays at noon Pacific.

But we got that calendar so you can convert that stuff. Noon Pacific, which is the same exact time we do this here show, only on Tuesday. So it's the same bat time, just a different bat day, Sundays. And we use our here mobile. Yeah, easy peasy. And it's a great way to practice your setup if you want to come join us live on Tuesday, which you should. And, you know, Chris, you did mention the Telegram group. That's still happening, going stronger than ever, jupiterbroadcasting.com slash Telegram.

Yeah, some say it's too strong. I say they don't know what they're talking about. So we'd love to see you in there. It's a great way to continue the conversation in between shows, but also get us looking at something. Is there something in the open source world that you think needs our attention? That's a great way to get it in front of us. So there you go. Also, you know what I'm going to plug right here? Linuxunplug.com slash contact. Get your feedback in. Ha! How about that? We love that feedback. How come we never plug that in the housekeeping, Wes? I don't know. Let's add that right now.

There you go. Well, let's just do a quick old update on our Pieholes. Piehole 5.0 came out, and we run Piehole here in the studio. I'm sure some of you run it yourselves, but those that are not familiar, Piehole is a network-level ad-blocking software that is predominantly designed to run on a Raspberry Pi. There is where you get the name, obviously. But because we're maniacs, we just run it on our big old x86 server in a container.

Old school. Old school. Well, actually, kind of, sort of. I don't know. Or new school. I don't know. What new school. I don't know. What's old is new again. So version 5.0 has landed.

And I thought, well, let's jump on the old server here at the studio and do an update right before the show, because what could go wrong on the very thing I used to access the internet? And nothing did. It was very boring. I did a Docker Compose poll, and then I did a Docker compose poll and then I did a Docker compose up dash D and waited one minute impatiently and then the new page loaded and I was running the next version. And it was the most boring upgrade in network software history,

which is a good thing in my opinion. Yeah. you know, they do have some notes as a project that this is a one-way upgrade. So if you leave version 4 and you go to version 5, there ain't no going back except for your backups. And yeah, you probably should have some backups. But they also note that Docker is a great way to run this because the upgrades, I mean, they're just a little safer. You don't have to worry about any intermittent state. As long as the configs don't break, well, you've got a time-tested image right there available. As far as 5.0 goes, what I noticed pretty much immediately was improved graphics on the dashboard for like my history charts of what's been blocked and whatnot, and a little bit of more information in log and areas like that. But I think probably the biggest deal for me and why I went ahead with the

upgrade was the custom host names from the web GUI now through local DNS records tab, this where you can assign the host name to an IP address, which I think, Wes, was the reason why we ended up not using Pihole for our DNS. Does that sound right to you? Because they didn't have this feature when we first deployed it? Yeah, exactly. You know, it wasn't quite ready to do everything on the network. We had some other infrastructure options available to us at the time. Now that they have this, well, yeah, why not just do it in Pihole? Because it really does have a pretty friendly GUI, especially if you've been used to configuring things like DNS mask by hand in the command line. Yeah, it is so nice when you just once or twice or three times a year

want to go in and do that kind of stuff. It's really nice just to have a UI that sits on top of tried and true Linux technologies. It's not creating some weird implementation itself. Right. Overall, pretty solid release, though. I mean, the update was done. I got in there and it's one of those experiences where I was like, hmm, I don't notice a lot of differences.

But what I have noticed is an improvement. Per device blocking is quite nice. So you're able to block individual clients. So say you have a teenager that wants to access Facebook between a certain time, you can do that now through the Pihole web interface. Yeah, it's a pretty nice customization. You know, Pihole is one of those systems that by and large, you sort of forget that it's running on your network. But for folks all stuck at home right now with a lot of diverse use cases, that might be causing some problems. Or maybe you don't need some of that blocking on your work computer that needs to get some stuff done, but you do want it to block out those things for your smart TV. Yeah, very true. I recently just did the math, the back of the napkin math,

and I looked at some paid- for blocking services for the kids. I was actually kind of doing on behalf of Angela for her because at the RV I have pie hole, but she wanted something for her network and we were just kind of doing the math. And it's like, if you look at it, if you put this on a Raspberry Pi, it doesn't need to be the latest pie either. So you may have one already, which changes the math, or maybe you just buy a Raspberry Pi 3 for a very reasonable cost, around $35. And then you load this image on there. When you look at it, it's probably a grand total of an hour of investment of your time at $25 to $55 worth of hardware. So when we were done with that back of the napkin math, 12 months or so, you could kind of justify the cost of a subscription service.

But by year two, you were clearly saving money by deploying this yourself. I don't remember exactly how long ago we deployed PyHole on our machine here in the studio west, but I think we're getting close to that year mark now. Yeah, and it's really been a good test that by and large, nothing's broke. You know, it's not that like sites mysteriously fail or all kinds of stuff just hangs when you expect it to load. So I've been pretty impressed. They've improved some other stuff there

just as from the admin perspective too. They've got deep CNAME inspection, which is pretty cool. So you can have multiple CNAME redirects, you know, when you've got a DNS name and it could be a whole chain of them. PyHole can now inspect that chain and determine if any of those are on your blacklist and then block that original request. And you mentioned better logging info. One of those things is more information about why something was blocked. So now in the

GUI, you can go in, find that something was blocked. And if you're curious, like what rule is blocking this, maybe it is something that you don't want to be blocked. It'll tell you just what that rule is. I do like that. I always love seeing other notable features include, quote unquote, more efficient memory usage. Always love seeing that on a new software release because it makes it easier to run it on lower spec hardware, which makes that whole cost calculation even easier. I think there's other things in here for those of us who subscribe to different blacklists. They've improved the parsing of imported blacklists. They have a gravity script that will now show how many domains were not able to be imported from that list too, and then

give you some examples as to why they didn't import so you know where to start from, which is exceptionally nice, especially for people who are doing this casually and don't spend a lot of time managing these types of lists or even trying to block things at a DNS level. Yeah, exactly. It's an interesting blend, I think,

because it is actually pretty nice for the power user. You can get in there and go into the command line. You can customize the DNS mask or DNS FTL, whatever they call it these days. You can customize all that stuff if you need to. We've got some custom configs and host files entry set up. But if you just want to use this from a casual end user perspective, that works great too. Do you struggle with the morality of using something like this, Wes? I've never had this problem with Pihole. But when I use something in my browser, I've gotten a couple of, hey, you know, to use our website, please whitelist us from your ad blocker. That is a little bit harder when you're using a network level blocker because it's just, you got to go log in, you got

to go add the URL. It takes more time than just clicking something in your browser toolbar. There's a bit of a disconnect here when you're using this. Do you struggle with that at all? Because I kind of do. Yeah. I mean, you are just sort of blanketly saying, hey, I want as few ads as possible. And that's pretty great from, you know, the actual end user experience. But we do have to acknowledge that many of our favorite sites on the internet, well, traditionally, a lot of their revenue, that's the things they use to pay for their servers or pay for their staff. That does come from advertising. So it is a good workflow to get into maybe practice once you set this up, like what does it look like to whitelist something?

What are the sites I go to a lot? And, you know, especially right now, it's just a good time if you have the resources available, maybe go subscribe, send some money to those sites that you turn off ads for. That's a good point. Bloomberg just recently did a piece about Piehole, interestingly enough.

And according to their research, Piehole is installed on onlyhole, interestingly enough. And according to their research, Pihole is installed on only 140,000 networks. So we're not talking like, you know, a plague that is impacting the advertising industry. But what are your thoughts on this, Alex, aka Mr. Ironic Badger? Is this one of those situations where end users are essentially robbing the creator of revenue. I mean, I hate to put it in such dramatic terms, but that is kind of what we're talking about here. Yeah, a little bit.

But then on the flip side, you have websites that are designed entirely as ad revenue farms. You know, all these ones that have 10 things about this thing that you don't really care about. They make you click through a 10-image slideshow instead of just writing the article properly. Also, trying to browse some of these websites on mobile is just horrible. And there's no real good way to block ads on mobile. So for me, Piehole or AdGuard Home is just a necessity to get a decent mobile browsing experience these days. I completely agree. You know, in Lady Jupes right now, I'm on a mobile connection. I get maybe,

if I'm lucky, four megabits where I'm parked right now. And having a network level ad blocker that just prevents pre-rolls and prevents certain components from loading on the web page means I put less traffic on the cellular network. I've been subscribing to the New York Times for a couple of months now just to try and support proper journalism. And I think that is the way that we should go about this, not forcing advertisements down our throats, really. Yeah, there is the tracking element, too, of this. And I think, in a way, it's users pushing back, saying this isn't the right way to monetize.

And I don't mean to make this about this, but to your point, the Unfiltered podcast, I've relaunched it recently, and I very intentionally am just crowdfunding. I've committed to no sponsors, no advertising. It is very much not a feasible way to monetize news, in my opinion. I think it's fraught with problems when it requires tracking of clicks, tracking of views, tracking across multiple properties. It's fraught with so many user hostile issues that if you just took a more direct to audience approach would be completely avoidable. just took a more direct-to-audience approach would be completely avoidable. And I'm not trying to justify the use of ad blocking, but I do think in a sense it's users attempting to vote with their wallet. Well, right. I mean, we haven't had a lot of feedback mechanisms, and you can talk to

individual sites, but there's no other real way to go yell at advertisers and marketers out there. And we get that some of that stuff's needed. Sometimes it might even be useful, rarely, maybe, but sometimes. But we've just seen a relentless increase in tracking and fingerprinting. That's it, Wes. Well said. And it would be nice if, you know, maybe a little more than 140,000 of us were using this because it might send a very clear signal. Also, I want to mention this Bloomberg piece is worth reading because the other thing they talk about is the core developer now who's taken over the project recently, well, kind of semi-recently in the last few years. He's got about five to 20 hours, he says, to work on this thing between college and his project. Now there is a flourishing community here, so it's not all on one individual. But the

lead dev for this project, this Pihole project that I think is, it's one of those showcase open source projects that even people who have no idea what Linux or a Raspberry Pi or open source is, you can show them Pihole and they love it. It's being created by somebody who's spending roughly, roughly five to 20 hours, depending on how busy he is at school. I think that's pretty humbling when you think about it. It's a guy who started this when he was in high school, got involved with the project when he was in high school, and now he's in college. And when science classes are busy, he's only got about five hours a week to spend on this thing. Yeah, and I think that does underscore a little bit too

of this is an organic need. It's not some conspiracy to hurt websites and advertisers. It's just end users getting frustrated and doing what they can. All right, well, let's cover a little feedback here on the show. I think that's a pie hole for you. And now let's talk about a Linux Mint success story

from Zachary. He writes, I wanted to share a Linux Mint success story from Zachary. He writes, I wanted to share a Linux Mint success story about my grandmother. She loves it. Her computer is a six-year-old HP all-in-one. It had all of the programs from Windows 8, and she hated Windows 8. Didn't we all? Yeah. You know, it was amazing, Wes. It really was something. The touchscreen worked intermittently. Oh, that's good. Took three minutes to boot. Oh my God. And she had a metered internet connection that Windows did not respect, which she was constantly worried about going over her cab all the time. There were weekly printer issues. Then eventually,

the Windows key deactivated itself. So not the key, but the actual serial key. So she got the Windows watermark saying this is not a licensed copy. She stuck it out as long as she could, but it got horribly slow, unreliable. She desperately needed a new computer. But I convinced her to first try Linux Mint, and she loves it. I do as well because computer problem calls went from weekly to only twice in the last year. Wow. Isn't that the truth?

I've noticed that as well. She likes it and says that it's easier to use than Windows ever was. I even forgot to show her how to shut the computer off, and so I called her and told her how to do it, but she said she'd already figured it out. Can you imagine how proud he must be of his grammar right now? Oh my gosh, I know. He said like, really, you figured it out?

And she just said, yeah, I mean, it has a big red power symbol labeled shutdown. Of course I did. Obvious. How simple. Yeah. Yeah. That grandchild pride is real though. Camille writes in with a question that I get about once a day right now. So I'm I'm very thankful for the opportunity to clear this up. It's about Pi boot. He says, hello. In episode 347, you mentioned putting some info about booting from the USB drive with the Raspberry Pi 4. It's in the show notes, you said, but I can't find it. Help, please. It's in the show notes, you said, but I can't find it. Help, please. Well, we'll link it again. And I'll say Pi 4, and it's that you can run from an external USB drive. Now, you do need the SD card.

You've got to have boot on the SD card. But the process is so simple, and once you do this, the read and write access to the SD card is so limited that it's probably going to last forever. Right. I mean, at that point, it's basically, okay, anytime you boot, it needs to read the inner MFS and the kernel. And then when you update and get a new kernel, then it has to write you back. But that's about it. And what's nice is it's so much faster. And yes, it's over USB 3.0. So a huge caveat,

past Chris would never, ever, ever, ever recommend that you run your system from a USB drive ever. Don't do it. Current Chris has accepted this compromise in hopes that the USB code base has gotten more stable as time has proceeded. And so far, so far, Current Chris has not been proven wrong. Yeah, I mean, USB has been adopted for so many things these days. And I think there are a lot of systems. I mean, I've definitely been guilty of that before, even running Windows off a USB drive once in a while.

So it seems to just work. I hope so, because I'm doing it. The way this works is you take the Pi image, and this may work for other distros as well. This is something I'm going to try. But you take the Pi image for your favorite distro. In this case, I use 2004, but try it with others. It just really requires a more modern kernel. You take that image.

You flash it to the SD card like you always would, maybe using DD or using, I don't know. Do you, do you? Maybe you use Etcher. I'm not judging. Then get ready for this. Ready? Take the USB drive. You plug it in.

You flash the same image to the USB drive. I know, revolutionary. You plug both devices in, you log in, you change the file system ID to match the SD card on the USB drive. And from that point forward, once the kernel loads up from the SD card, it sees that file system label. from the SD card, it sees that file system label. But from what we could tell, essentially, it just, the way that it's hard-coded, it first attempts to boot from USB.

Yeah, it seems to work out that how the kernel chooses its boot devices, you know, they both ended up having the same label, but it just picks the right one to start with. So we didn't even have to erase the label from the SD card. So that means, too, that in the worst case, your USB dies, well, you can still, you know, boot to repair label from the SD card. So that means, too, that in the worst case, your USB dies. Well, you can still, you know, boot to repair things from the SD. Yeah. And we'll link to the guide that I used to do this. And like I said, start about midway down. And it's so much better. First of all, you reduce the wear and tear on the SD card. But, you know, these SD cards are slow.

So slow. So slow. Where I have USB 3 and an SSD it's about as fast as you're going to get in fact I did some benchmarks just to test it because I wanted to get some real numbers and from what I can roughly tell it's about as fast as I could possibly get a Raspberry Pi 4 now without overclocking it you know and just from our own experiments when you SSH into one of those things and it's on the SD card and you launch something, boy, can you tell. But running from the SSD, it just feels like you've SSHed into any old x86 VPS. Yeah, you'd have no idea it's a Raspberry Pi. You really wouldn't. You just think it's a VPS or a physical x86 box and things just run. Yes. And so much stuff is supported on these systems you know

that's all been cross-compiled it works on arm it's great all right well now wes you've got that brand new really nice looking dell xps 13 in your hot little hands for a review and we got some feedback on the like hey include this in the review feedback that we asked for last week it came from cory and he said, for the XPS 13 review, I'd love to hear about using it with a generic install of Ubuntu 20.04. He says, what changes has Dell made to the drivers? Do they submit them upstream? How will this machine run 22.04 in two years?

You know, Corey, thank you for that feedback. Those are all definitely things I'll be checking out. Actually, a little spoiler right here. It came with 18.04, I think, on there and it worked, you know, everything worked out of the box. Wi-Fi was working. I just wanted to test it and I've been loving 20.04. So of course I overwrote it and, you know, stuck that on there. So far, no Wi-Fi yet. I'm going to do some more testing. I'll definitely dig into, you know, what drivers are in use and what you actually need to get this working. But unfortunately, it's like it's still new enough that that's a little bit tricky. Interesting. I'm glad you're trying that little early discovery there. You know, the way my understanding of this process

is, is that Dell is not creating drivers and then creating a bastardized image that ships on the XPS to make it work. They contract with Canonical, and Canonical developers make it work. They generate an image, and they certify it, and they deliver that to Dell, and then Dell ships that on the XPS. What I'm saying here is this isn't Dell like hacking it into an image. It's the Canonical developers making it work. So it has a very, very good shot that it's going to get upstreamed into mainline Ubuntu. And that is part of the arrangement. There is a licensing agreement that Dell pursues with Canonical.

There is a revenue share for every laptop sold, and those support services are part of that agreement that Dell makes with Canonical. The value for Dell is that the Ubuntu developers themselves are doing some of the coding necessary to get things to work like maybe that Wi-Fi or maybe that new fingerprint reader. The nice advantage for end users is you're not having some third-party OEM just slam-jam code into your image where they've hard-coded a PPA and you're good to go. You actually have the maintainers of the distribution that are writing this. It doesn't necessarily mean that makes it always into the mainline ISO,

but it has a very good shot. And it also is a revenue source for Canonical. Yeah, absolutely. And I should note, I was just using, you know, the same random 2004 USB stick I had laying around from our review. So that hadn't been updated. And because I didn't have Wi-Fi during the install, I actually haven't updated the thing yet. So it may be that I do some updates and everything just works. Yeah. And on top of that, you know me, I'd be also interested in other distros as well. You know what you got to do, Wes, is you got to keg-zack into some RAM disks.

As if I won't be doing that already. Yeah, I know. You've got to put at least some of the major ones. I'm sure Fedora32 will be getting a run on there and Manjaro as well. One of Wes's most recent duties was he was put in charge of hiring our new soundboard guy who just sits here. Too soon.

He just did it too soon. Whoa, come on. We talked about this. Wes. Wes, I don't know what to do about this guy. I don't know what to do about this guy. But we do have some picks to get into. It's not the same now.

We've got to get rid of this guy. Maybe one of our picks eventually will be a machine learning model that replaces him. I told you hiring a soundboard guy out of Bangladesh was going to be fraught with latency. I told you that. Yeah, that's true. But doing this for a long time, Wes, that's how these things work. But the price is right. Yeah, you're right.

In fact, I think he's paying us. I don't know how that works. Well, we have a couple of picks. This one was sent in by several individuals. I think the whole internet, in a way that doesn't typically happen, noticed this pick. But we'll give credit to Stephen because he actually used the contact form. Stephen wrote in. He says, I just found this little tool that converts any USB thumb drive to a multi-boot ISO stick.

All you do once it's finished is copy over the ISO and reboot. It can hold numerous ISOs all at once. Also, any normal files you might want just to use it as a USB stick. It's got a GitHub page, and it's undergoing very rapid development. Alex N. on Twitter also sent this in, and interestingly enough, on my own accord, I also found this one and tagged it. Everybody found this,

and it does seem like a super cool pick, Wes. Oh, yeah, absolutely. I mean, way back in the last days, I think you guys talked about some hardware that could do this, a special USB stick that you could copy ISOs on. But talk about easy-peasy. When you're making these USB drives from ISOs all the time. Why even bother with Etcher when you can just use CP? Yeah, it's like the last USB stick you need and you just throw the ISOs on there. And you're right. I had, I think it was a Zalman drive. Yeah, that's what

it was. And that's pretty slick and kind of makes sense that you might do this in hardware. I'm really impressed. I'm going to have to do some digging here in the repository to see how they're making this work and just how slick it is, because they even talk about having UFI support and legacy BIOS both working the same way. It presents you with a nice little menu that finds all your ISOs and lets you choose between them. Holy cow. Yeah. You just now stick the USB disk in, copy the ISO over. And I mean, I love this about my Zalman USB drive, but it would do classic CD-ROM emulation. So it actually show up to the host OS as an actual USB CD-ROM drive, which was just too much. This is much simpler and it just copy them over and you're good to go.

I think this is going to change my Linux distro hopping game forever. The big statement. I think you might be to change my Linux distro hopping game forever. A big statement. I think you might be right. Well, how can we top that? That brings us to the end of this week's Unplugged broadcast. We'd love to get your feedback at linuxunplugged.com slash contact. Please do. And if there's anything else you want to know about that XPS before we do the official review, let me know. It's coming soon. And you can hit me up on Twitter. I'm at ChrisLASWest. At Wes Payne.

And go get yourself some cheese bacon cheesy. At Cheese Bacon. Oh, it's magic. And of course, the show is at Linux Unplugged on Twitter. The network is at Jupiter Signal. Thanks so much for tuning in to this week's episode of the Unplugged program. And I'll see you right back here

next Tuesday. I found the WSL content really refreshing because I got to know the change from within Microsoft when it happened. So I was there with Windows 7 to Windows 8 and Windows 8 to Windows 10. And you really notice the change from Balma to Nadella. And it's like day and night and people are still in the embrace, extend, distinguish mentality. And I really can't hear it anymore. It's just so tiring. So you were an employee of Microsoft? I was a close contractor. It was like subcontracting. I did...

They have lots of contractors. Yeah. So I was teaching sales for the Microsoft employees and for the retail guys and stuff. So from what you witnessed personally you felt like the transition was legitimate internally yeah for sure absolutely it like it was they had they had meetings they had like people flying out to redmond and then like showing them azure and like when the whole canonical stuff happened people started getting really curious. So I did some workshops because I was always Linux curious and did use it privately. And so I did some workshops for the sales guys that had to deal with Azure back then and stuff. So, I mean, I'm from a little country,

like I'm from Austria, but, you know, small changes. Yeah. I feel like my proximity to people at Microsoft gave me a bit of a view into the company that people maybe outside of this area don't have. Not only that, but I will be honest that I, now looking back at it, feel like I was childishly, irrationally angry at Microsoft. In the late 90s and the early aughts, Microsoft was the greatest evil in the technology industry that I perceived. And the technology industry had the greatest potential for influence on mankind. You might even say they were a cancer. my life's work to inform people about alternatives to Microsoft. And so when I had the opportunity to go down to Redmond, because of some content we were covering in Coder Radio, I actually was contacted by some folks at Microsoft and asked to come down there. And when I sat down there and

started chatting with folks that were Gen 2 users and SUSE users and Fedora users and Windows users all around the same table, I went, oh, this is different. This really is like something that's been multiple years. At that point, when I went in, it was like five years in the works. And I just couldn't be intellectually and emotionally honest with myself without recognizing the cadence difference in everything they were saying to me. And the fact that a lot of the employees there were listening to our podcast because they're interested in what Linux users are thinking and doing and care about was also very telling. Because when I walked around there, it was like I was kind of at a Linux conference. Like people knew me and I didn't know who they were. You know, that awkward kind of like, you know me and I don't know you, but you know me. That's weird. And that was the Microsoft

office. And I walked away from that and I brought my wife with me and I walked away. I remember talking to her. I'm like, this isn't the Microsoft that I grew up with. Now we can all just circle up, hold hands, sing Kumbaya and just beat up on Oracle. Amen.