Please be advised this transcript is AI-generated and may not be word for word. Time codes refer to the approximate times in the ad-supported version of the show.

00:00 - Jonathan Bennett (Host)
This week we talk about the new Threadripper hardware and its many, many cores. Bcache FS might get removed from the kernel. We're still watching that one. Wayland's not ready yet, but Network Manager 1.54 is, with some really cool new features around IPv4 forwarding and more. The Steam survey almost hits 3%. There's a lot of stuff to cover. You don't want to miss it, so stay tuned.

00:27 - Leo Laporte (Announcement)
Podcasts you love From people you trust. This is Twit.

00:36 - Jonathan Bennett (Host)
This is the Untitled Linux Show, episode 214, recorded Saturday, august the 2nd. Multiple kernels for the price of one. Hey folks, it is Saturday and you all know what that means. It's time for the Untitled Linux Show. It's where we talk about open source, the Linux, desktop hardware, software, all kinds of stuff. It's going to be a lot of fun. It's not just me. I've got Jeff Massey with me and it is the two of us, but we're going to have a blast today. For those of you watching live, we are getting started even later than usual. I had a headache that was doing a great impression of a migraine, but back at 100% now, thankfully, and we're going to go. The show must go on. So we've got some stuff to talk about see, I was gonna say jonathan's feeling better.

01:28 - Jeff Massie (Co-host)
So that means linux fixes everything, exactly, so exactly oh yeah lots, lots of stuff to talk about uh, you've got a.

01:38 - Jonathan Bennett (Host)
You've got a big review of a big chip, don't you have to start with I?

01:42 - Jeff Massie (Co-host)
I do. I mean it come on it's, it's got to be me.

01:44
I if, if we get, if we got new silicon out, I got to talk about it so this week amd released the new 9000 series thread ripper line of chips the 9960x, which has 24 cores, 48 threads, for about 1500. The 9970x, 32 cores with 64 threads, for almost almost $2,500. It's $2,499, but we'll round it $2,500. And finally the 9980X, which has 64 cores and 128 threads. Now there's also the Pro series of chips which have been released, which have even more cores, more PCI lanes and use a different motherboard, so you can even get more memory on them. But we're not, we're not going to talk to the, the pro series, the, the thread ripper series we're talking about. They target the high-end desktop desktop market and the pro series targets workstation. So they're kind of the midpoint between super high-end desktop and, you know, like the epic server, full enterprise so. But they're, they're a lot more expensive, they take different motherboards, they, you know they're, they're a class unto themselves. So we're, we're only dealing with the high-end desktop. So michael larable over at pharonix got his hands on a couple of the chips. Well, because AMD sent them for testing, the 9980X and the 9970X chips to be exact, and he benchmarked them against the 7000 series of chips, the Threadripper chips I mean, and he even included a 3990X Threadripper. So that's going back a few generations and he even included a 3990X Threadripper. So that's going back a few generations. Remember there was no 5000 series Threadripper. Also included is a 9950X3D, a 9950, and an Intel Ultra9 285K were also included, because that's again high-end desktop and you're kind of crossing over to where you need more cores. Now there's no Xeon, there's no EPYC CPUs in this test. Again, same way with the Pro chips. They're in a whole different class and those are aimed at enterprise users and it's not the same use case. Ubuntu 25.04 was used along with the 6.14 kernel and each chip was run at the maximum rated memory channels. So if it was dual channel, quad channel and max rated frequency of the memory, so this is not holding back to a standard. This is how the chips. You can get them to work as you would run them in your system.

04:27
Now I want to state this up front If you're going to get a Threadripper chip, there are two basic reasons you need it. The first is going to be you need more cores. Now, they're not going to be faster cores than, say, a 9950X, there's just going to be more of them. So anyone thinking this is going to be great for playing games no, this is going to be wrong. Yes, they will play games, but they're not going to be at the top of the heap, as most games don't use all the cores available to them and the single core performance is very important to gaming. And the single core performance isn't as high on the Threadripper chips. When you have more cores in a package you have to have them clocked down a little bit to both heat dissipation and you're kind of limited by the lowest common denominator. And the more cores you have, the more restrictive on the clock frequency you need to be. Now, the second reason you might need one is more pcie lanes or memory. Uh, you know, personally, if I was getting a thread ripper, I would. I would get the lowest level, the, because I would use more pcie lanes. As I have run into that bottleneck in the past, now am I going to get one? No, it's much cheaper to get an NVMe to U.2 adapter and invest in a new than to invest in a new motherboard and CPU. So I don't and I don't really have a real use for the more cores than I have. So I I'm running a 12 core system now and that's more than enough cores.

06:02
Now back to the benchmarking. There were over 200 benchmarks ran and there's a link to the test results in the article which is in the show notes. And there's a lot of compile benchmarks you know, like, for example, compile of the 6.8 kernel with default configs. It was able to. It was compiled in 21.35 seconds for the 9980X. That is getting fast, I mean boy. That's just hauling AI workloads, scientific workloads and media workloads like rendering and transcoding are the rest of the benchmarks. So there was no gaming. It's all a lot of things that really showcase where the cores come out. And for those that like to watch videos, for example Jay's Two Cents he was having problems benchmarking because he got a Threadripper and a lot of his benchmarking software. It was not set up to handle that many cores and so he only had a few tests he could run to truly show what they could do. Because you know, if your software maxes out at eight cores, well, however many you have in addition to that, it's not going to do you any good.

07:21
But looking at the overall results, the 9980X came out on top and it showed a 30% improvement over the 7980X and the 9970X showed over the 7970X a 28% improvement. So that's, I think, a pretty good generational uplift uplift. Basically you're running roughly 30 improvement in a generation. So that's, that's pretty decent. The 3990x came out slightly behind the 9950x which were. That was well behind the thread ripper chips. I didn't do the math but it was probably about 20% to 30% behind the other Threadripper chips, the 9950X, 9950x, 3d. They were just a sliver away from each other. So the 3D was slightly faster, but it's such a small margin that it could have been just rounding error. You know, not statistically significant, you know, I honestly would have loved to see the Intel chip do better this round.

08:35
But the 285 was dead last, and by a fair margin. It was another double digit percentage down from the other chips. It just in these scientific compile type benchmarks it did not hold its own. Uh, power consumption was slightly up on the new thread Ripper chips, but it was just a tiny bit and Michael larableble even said, you know, really it could be considered the same. So basically you can get 30% improvement for roughly the same power.

09:14
There was. There was like a couple Watts difference. But because they're pulling in the triple digits, you know you're you're looking at like a 1% difference. So overall it seems like a great deal to me. Take a look at the article in the show notes for the full details and decide for yourself if you need all the cores, which, for an entire system, do come at a pretty hefty price tag. Don't forget you got the CPU price which we talked about $600 or more for a motherboard, and really, if you're gonna use this, you should get quad channel memory to fully take advantage of this, and that's going to be another nice price tag on there. So you know and that's not counting if you need even more parts because you don't have to have other existing. You know power supplies and things that are going to be able to fully power the Threadripper. You're probably going to be able to fully power, the thread whipper.

10:04 - Jonathan Bennett (Host)
So you're you're probably going to need a dedicated power supply. Just for that, you will probably need a new power supply yeah, most likely.

10:11 - Jeff Massie (Co-host)
So you know, yeah, just keep that in mind. That it's, it's going to be pricey and realistically, that's where a lot of us even could say you know, I compile quite a bit, but you know, for the cost, a few thousand dollars, I can wait another 30 seconds or one minute to just let things go yeah, um, I watched the the linus tech tips video on it as well and, uh, some interesting stuff there.

10:39 - Jonathan Bennett (Host)
but apparently, like they recommend a 1.5 kilowatt power supply for this thing, it's just, it's nuts how much power it can pull. Oh, yeah. I think part of that is when you load it up with all of the you know multiple GPUs in a system, but still it was.

10:55
it was pretty impressive how how much power they said that they wanted it to have and of course then, of course, linus, on Linus tech tips, they they kind of went all out and turned all of the overclocking all the way up to 11 and then did a open loop out of a trash can full of ice water. It was pretty glorious in its ridiculousness.

11:16 - Jeff Massie (Co-host)
But it worked. It's not dumb if it works. I did watch that video as well, and you're right, a lot of times they recommend a huge power supply. Because, you're, you're, if you're using that many PCI lanes, you're either hooking a ton of drives which, okay, isn't going to take a whole lot of power but a lot of times, too, you're going to have multiple graphic cards or things like that, or high speed networking. You're, you know it's, it's built to be worked. One thing, though, is, with switching power supplies, you could run a higher if you happen to have a two 20 or two 40 volt, uh, uh power around you could run like a European style supply.

12:01
A lot of those supplies have a wide uh voltage range and frequency range, so if you need a little more, that's you. You're gonna switch to 240. Unplug your dryer and plug your computer in there yeah, with the realization they might be giving off the same amount of heat.

12:21 - Jonathan Bennett (Host)
You know just drape your clothes over your server and call it a day, yeah, all right. Well, there's something caught my attention this week, and that is the latest update to Network Manager. It's one of those tools that you may not realize you've got running on your computer, but you almost certainly do. It's almost certainly part of your Linux distribution and Network Manager 1.54 dropped this week and it has some interesting things in it, one of which being the IPv4 forwarding setting per device. Now, those of you, those of us, that have done VPNs in the past, you may be very familiar with the IPv4 forwarding settings. I've got a buddy that we've set up VPNs together in various iterations for various reasons, and we have this running joke going that whenever anything doesn't work, it's always have you checked if IPv4 forwarding is turned on? And about half the time that's what it is you need to go turn that on for traffic to actually flow. So I saw this and really perked up. It's like, oh, you can do it in Network Manager now and you can do it per device rather than turning it on globally as a kernel setting. That's very interesting. And then there's some other things in there. There's some IPv6 stuff with the prefix delegation which, if you've got multiple ipv6 addresses from your isp, super interesting to be able to work with that. Um, some other things in there, um, with network manager and wire guard and wire guard and ipv6. Uh, the the nm tui network manager 2e, the NMTUI Network Manager 2E also has support for configuring the loopback interface, which that's not something that you often need to do, but when you do need to mess with it it's nice to have the tools. Some upgrades on being able to change certain connection properties without needing to bring a connection all the way down and back up, which, if you're ssh into the machine you're working on, might be really important. Um, yeah, various things around that.

14:37
And, uh, the one of the last things in this article it's mentioned is the early boot. Network configuration has improved and so, like, network manager now understands the boot firmware table and so it gets a better idea of what devices are available right after boot instead of having to wait, and so that's also pretty interesting. Some neat stuff in 1.54. And again, it's just now released. So you know you'll probably see this in most distros We'll roll out in the next major. Maybe the next major release may come as an upgrade to this one. It just sort of depends upon how aggressive a distro is with doing these in-between major version updates. So something like Fedora I would assume we will get this in a few weeks. Something like ubuntu you may have to wait for um 2510 to be able to get a hold of this.

15:31 - Jeff Massie (Co-host)
It's some neat goodies in there yes, and but you can turn on backports if you want to live a little more dangerously, and which is a lot of the pre-software, the the upcoming 25.10.

15:43 - Jonathan Bennett (Host)
Yeah, I think almost all distros have something like that. It's scarier in some of them than in others because, like in Fedora, the way to do that is just start pulling packages from Rawhide, which is the tip of the spear for the next version. But I've done it. I've run Rawhide. I've run partial Rawhide before. It's always a good time. I have too For the it. I've run rawhide. I've run partial rawhide before it's always a good time.

16:06 - Jeff Massie (Co-host)
I have too, for the same reason wasn't it yeah, I wanted the new uh plasma 6.0 and then uh, but I let mine slip a little too much and I did a whole bunch of updates and that kind of didn't work out so well yeah, yeah, sometimes that causes problems yeah too, too much, much at once, just choked it a little bit yes yes, All right.

16:28 - Jonathan Bennett (Host)
so it is time to talk kernel stuff and we have a kernel release. We have some news coming. So we just got 6.16. We've got news about 6.17 because the merge window is open there and we're starting to see what's landing. But you've got something for 618. We're looking look at a few months off off down the down the road off into the horizon for this one kind of well, we're gonna, we're gonna touch on 617 as well, so you get.

16:56 - Jeff Massie (Co-host)
You get multiple kernels for the price of one. What a bargain. So bcash fs says they're going to drop the experimental label in the 6.18 kernel, which it means then it could be considered stable. Maybe Now. I say maybe because Kent Overstreet and Linus Torvalds don't see eye to eye on some things Now. We've covered issues in past episodes but to catch people up, linus said that he was going to part ways with the Bcash file system in the 6.17 kernel.

17:31
Linus was upset that Kent would be trying to put in new features into the goes in at the pull request window and then the rest of the time in the RC stages you're just fixing any bugs that might have come up with the new code. So you're not supposed to be adding new features at that time. It's just a pull. Everything in RCs are just for polishing and taking away the paper cuts and the bugs. As of the time of this show now, nothing has happened. The code is still in there, but the merge window is still open and maybe Linus won't do anything until after the merge window closes. Now if he does do something, what does that even mean he could push out the code to be under the broken kernel configuration option. So then to get it into the kernel you would have to compile it with the broken flag set. Maybe do another timeout cycle. So meaning in the past, when Kent was pushing Linus's buttons and making him upset the code froze for a release cycle. So no changes were allowed for a kernel lifecycle. So from the pull window to the final release, linus just said no, I'm not accepting code, I'm not taking any pulls. He just wouldn't let Kent have any code change in the kernel. Maybe there's something else that could happen, or maybe things have cooled and they're gonna just let the code stay where it is. Because you know, maybe things have cooled, maybe Kent will behave himself, and yeah, we'll see. Back to today, though, kent said the following "'I've been digging through the bug tracker. "'and polling users to see what bugs are still outstanding "'and it's not much'". So the experimental label's coming off in the 618 kernel. Now that would mean the file system would be stable for this year's long-term support release possibly, which is thought to be the 618 kernel. Now, that's if the kernel timings stay somewhat regular. So that's always subject to change depending on how 6.17 goes. If it goes out the window then the 6.18 won't be the LTS kernel. It's the last released kernel of the year. Is the LTS kernel?

19:59
A few highlights of the code pull going into 6.17 or requested to be pulled. We'll see if it makes it in. Is fix a major performance bug when deleting many files and this was caused by the key cache caching keys well, that's a mouthful that had been deleted, causing certain lockups in the inode triggers to scan previously. They also fixed the IO read no promote counter has been broken out into sub counters. These can be seen with the B cache FS space FS space top on a recent B cache FS tools. This helps when diagnosing why reads aren't coming from the cache. Congestion tracking is now a little bit less aggressive and that controls when we decide to do a promote.

20:48
But Kent acknowledges this area still needs more work. Metadata rights are no longer throttled by writeback. Throttling no cow, probably no. Copyright rights can now be rebalanced, for example background target, background underscore, compression options and almost all recovery passes now have progress indicators. There's also repair improvements. He says we'll now reconstruct missing inodes if we find contents for that inode more than one or two keys, not just if the inode B-tree was damaged similar for dirant to missing inode. And they also fix an in-memory accounting going out of sync with the accounting B-tree when doing accounting updates. Before going read-write, take a look at the article in the show notes for the full details and all the fixes. I just listed a few of those and I edited out some that were a little more archaic sounding. I didn't know if they were super important or not. So definitely check that out, especially if you're a coder or really, uh, playing with the bcash file system and we're just going to wait and see what's going to happen.

22:11 - Jonathan Bennett (Host)
You know, um kind of exciting and we'll let you know what, what the results are and if, if, things change, and keep everybody apprised uh, when was the last time that we talked about file systems in the kernel and called it kind of exciting in this sort of way?

22:29 - Jeff Massie (Co-host)
yeah it, um, but you know that's one thing, though. You don't want an exciting file system, exactly. We don't want this kind of excitement. Yeah, you want boring, very boring, and and personally, even if it goes you know I'm kind of interested in the B cache file system I would probably wait a good year before I even would attempt to get in, to get on it. You know, maybe maybe even longer than that I was going to say you're more aggressive than I am.

22:56 - Jonathan Bennett (Host)
I'm thinking give it five years of stability before you actually trust it for any kind of critical data.

23:06 - Jeff Massie (Co-host)
Well, no, I that's just putting it on like my gaming machine, my, my server, you know, with all my file stuff and all that. Oh yeah, that's not getting tied you know five years or long time. It's staying on ext4 for the foreseeable future indeed I.

23:22 - Jonathan Bennett (Host)
It's funny, I have got a couple of servers that I've moved over to. I think it's xfs is what, like the red hat stuff uses by default, and it's got some nice features, but it's also sort of a pain to work with and I sort of I sort of wish, like man, I wish I just put ext4 on these, it'd be so much easier boring is best, as uh x1011 says.

23:49 - Jeff Massie (Co-host)
Last time we had excitement in file systems, we got reaser file system and boy that, uh that did not turn out well.

23:59 - Jonathan Bennett (Host)
No, no, it didn't all right, uh. So we've got some more kernel news, and that is. I alluded to this briefly, but we got this past weekend the release of 616 and there's some fun stuff in there. Um, new stuff in the no view, excuse me nouveau uh driver for nvidia hardware, although a lot of that new stuff does not yet reclock and so it's still not particularly performant, but it works, turns on, it could fire up a display. Now, one of the other things that really fascinated me was that we got the OpenVPN DCO, the Datto Channel offloading driver, and I talked earlier about uh doing vpns. I used to use open vpn a lot and it's I mean, it's great software. Um, I've just almost entirely gone to wirecard now, uh, and I'm actually really curious whether wireguard can make use of any of this uh dco work or if it's just architected completely differently and therefore doesn't make sense to.

25:09
There's some stuff from Intel that's working on landing and some AMD stuff. There's the AMD GPU user mode queues, which this one is really interesting. This lets a. If I remember correctly, this allows a game that you have full screened to skip the compositor and do calls more directly to the GPU and you might get some improved frame rates out of that. There's your normal file system stuff. The boring file systems do still get fixes and changes. But there's also I remember seeing some additional AMD stuff around ray tracing, as that's getting better and better. A lot of little things like that have landed in the 616 kernel. And then, as I was looking through what all is in 617, I saw an article here about Firewire support being worked on.

26:14
And Firewire is old technology. At this point I think of it rather fondly because I've done quite a bit with it over the years. Between the digital video link that a lot of I think it was Sony camcorders came with, it was actually Firewire underneath. And then I've done Firewire audio interfaces multiple times and I have one in the rack. I don't think it's actually active and connected to the computer because last time I tried it Firewire and Pipewire did not get along very well together. I need to try that again. I think it's better.

26:48
But we got some changes to FireWire and one of those was it removed a tasklet usage and replaced it with a work queue and what that does. I know that doesn't mean a whole lot to people that don't spend time doing kernel stuff. But that gives you asynchronous packet transitions so you don't have to wait, packets can just go. You don't have to wait for things to go one at a time into the FireWire link, which means that you can do low latency, which is super important for things like audio FireWire interfaces. There's a couple of those other things. That was the main one.

27:28
Work queues also support preemptible work items. So again, you do something like a real-time kernel and that will suddenly work better and you can guarantee those low latencies. So interesting stuff going on in the kernel when it comes to Firewire. And you know we've got sort this guarantee of support, or at least a promise of support, from Takashi Sakamoto that he is going to support it in the kernel until at least 2029, which is kind of scary how rapidly that's coming up, but that's four good years of work on it and I doubt it will go away then. Uh, he may just move on to other stuff. But yeah, firewire, it's been around for a long time, still gets used. Still gets used for some of this stuff.

28:14 - Jeff Massie (Co-host)
Oh yeah but you know, in 2029, at that point it might be just a okay, the code just stays there and nothing changes. Well, that's why in the kernel you've got some of the old Sun workstations and things like that that have code in there, because the hardware never changes, so they don't have to really tweak anything. It just stays where it is. So there's no big overhead, it still works and it just kind of sits there for people that need it. Yeah, absolutely. On file systems, though, I did see something that caught my eye too was the EXT4. They're adding the 617. They gain better block allocation scalability. Hmm, gain better block allocation scalability. Hmm, and I didn't. I didn't go deep into the to the story, but uh, looks like uh, block allocation scalability is uh, greatly like 650.

29:17 - Jonathan Bennett (Host)
Improvement in scalability wow, it's a better, better performance on very large drives, yeah.

29:26 - Jeff Massie (Co-host)
Yeah, this isn't something that we're going to see day to day, but on the real high end and certain corner cases, I don't know.

29:35 - Jonathan Bennett (Host)
I've got a couple of different RAID 6 drives that are EXT4 that might see a little bit of improvement from this.

29:42 - Jeff Massie (Co-host)
Well, yeah, but don't expect 600, and you know sure plus no, no, 600 improvements.

29:50 - Jonathan Bennett (Host)
No, indeed, indeed, that is fair, all right. Um, jeff has a story here about small linux projects having big problems and, uh, the the tagline you've got here is when passion isn't enough, yeah, I'll have a book plug at the end of the show. I guess that this very much reminds me of.

30:18 - Jeff Massie (Co-host)
Well, I'm going to start out. This story is an opinion piece and personally, though, I found it kind of mirrored a lot of my thoughts and I thought it would be a good story to have on the show. And especially, we're supposed to be kind of a user's group, kind of feel well, not everything has to be cut and dried facts. So it's just, we're going to throw out a little, uh, color commentary, if you will. So you know, and on this show we cover a lot of different distributions which pop up with, you know, a new idea or a new software packaging scheme or something you know, and and out. I want to say that this is not a dig or any kind of negativity that I'm giving to those startup distributions. You know, just some things to be aware of. And you know, the article says, and I quote this isn't a dig at small Linux projects. In fact they're a big part of what makes Linux so exciting and full of possibilities. The real focus here is on newer users who dive into Linux for the first time, get swept up in all the options and then later realize that maybe they picked the wrong distribution for their needs. Now, for those that don't know, which, I'm sure a lot of our audience does. But just in case somebody's new there, there are a ton of distributions out there and if you want to see some, just go to distrowatchcom, which we've mentioned on the show before, and on the right side of the screen they have a list of the top 100, kind of by popularity. I say kind of because we've talked in the past how the system can be game, so take it with a grain of salt, but bottom line, they always have a top 100. So there's always new ones popping on and some falling off and there's just a lot of distributions out there. Now there are some hidden risks with these small distributions, especially when they're driven by a single person.

32:02
The article goes into more details, but basically someone is really excited because they have a new way of packaging, configuring using software that most distributions don't, or you know, there's some, some unique flavor, they they get excited about something, some particular thing, and they get excited and build a distro and push it out there and you know, then you get some bugs back and while they love making the original spin, they find you know, hey, okay, I can fix the bugs, and then more come in and you know, then they find you know, hey, okay, I can fix the bugs, and then more come in. And you know, then they find that maintaining the distribution is a full-time job and can burn a person out and that enthusiasm can fade away. Even a small team can burn out because just the big workload of maintenance and you know, you kind of, you kind of have the big all right, we release this. You kind of have the big all right, we released this. But the maintenance stuff kind of some people look at it as more drudgery and it's not for everybody. Well, and maybe they tried so. Then that release disappears. Now maybe the original release was they were setting something up in a unique way, and then find out there was a reason others didn't do it that way Because while the idea seems cool, it causes other issues downstream or in other parts of the release and, as mentioned in the article, real life doesn't take a break.

33:25
And all this in addition to what your daily life has and jobs and family and everything, can also cause people to kind of just drift away and that distribution either disappears, it falls behind, doesn't get regular updates. You know, the article also brings up security, both in maybe writing a custom wrapper around another piece of software to make it better, easier to use, maybe for example, packaging software. You know, if the distribution doesn't get a wide adoption and thorough testing, there could be security issues that are not found. Or maybe the distribution is built on top of another known distribution, like, for example, fedora. Say someone built a distribution on top of Fedora 40. For example, fedora, say someone built a distribution on top of Fedora 40. Well, if they're not constantly keeping things up, when Fedora 40 goes end of life and is no longer updated, the distribution built on top of it won't get on any updates. So security issues won't get fixed. They got to stay on top and move to the next revision of Fedora if they're going to have a distribution built on top of it.

34:33
Now I am cutting this article way down from the original as it would be the length of probably about three stories to do here. So I'm not going to touch on some of the other reasons, the issues a person might need to be careful other than to say support can be an issue in the single person or small team spins, and there's also the issue of ecosystem compatibility. You know the excitement which first spawned the distribution was great. It might've focused around a single piece of software and you know, but not had the proper testing for the rest of the system. So while the whatever works, great, they kind of ignored other corners of the distribution and things fall apart at the edges. Now I'm going to let the listeners take a look at the article, the show notes, and see the details and link the conclusion to the article. So, like I said, it's a lot bigger than what I'm giving it justice here. So please go back and check it out, the article. So I said it's it's. It's a lot bigger than what I'm giving it justice here. So please, please, go back and check it out, because there's there's a lot of interesting points in there.

35:36
Now my personal feelings well, I will play with any distribution on a spare drive or something I don't care about you know where. If it crashes, it goes down, I don't care, I'll be okay, you know. Then, sure, I'll run the alpha first, distribution beta, you know, whatever you know, but I don't rely on a distribution unless it has a couple of years at least have regular updates under its belt. That's for me. Bottom line, though know your risk you can take. How valuable is your data you're using for this distribution and be mindful of what you want out of your system. So not to say, don't go play and have fun, just make sure you take calculated risks because you could have something go wrong with the software, it corrupts something, you lose a drive, you don't have support security issues, Just be aware and plan accordingly.

36:35 - Jonathan Bennett (Host)
Yeah. So what you're telling us is that we should make sure and only run distros from big established companies like Intel. That won't let us down. Exactly.

36:50 - Jeff Massie (Co-host)
But they had a notice that they were shutting down and it was advertised. So there was a difference there versus things just stop getting updates or aren't getting full updates. What was the Red Hat CentOS? Centos, there was an announcement. Yes, it stopped. No, distribution is safe from that, but there was forewarning that's fair, so that you were able to make the transition before uh-oh, hey, this hasn't had an update. What's going on here? Oh, everything disappeared. The repositories poofed, or you know whatever.

37:33 - Jonathan Bennett (Host)
Yeah, obviously, taking a little dig there at clear intel and clear linux, um, because there wasn't any. There wasn't any warning giving it, given it just happened. But you're right, they did, at least let people know it was happening, whereas with a little distro it, the maintainer may just disappear and there may not be a notice anywhere. You may not know that it's not getting updated until a year down the line and you finally realize that wait a second, I haven't pulled any updates for a while.

38:01 - Jeff Massie (Co-host)
Well, there was one distribution I remember too. That kind of originally made me wary, because it was I can't even think of the name of it now it was run by a guy named Techstar, Like T-E-X-S-T-A-R, it was like.

38:19
It sounds familiar, but I remember it was there and then it wasn't, and then he came back and then somebody else took it over and then that languished and he was back in it again PC Linux OS, os. Yeah, that's it, pc linux os, and it seemed like it had a little bit of a shaky history there for a while now. The article specifically calls out there are exceptions to every rule, slackware being one of them, led by one one guy, and he's been doing it for a long time. We had its birthday last week. So being small doesn't mean they're going to go away. Just don't put all your eggs in one basket until stuff has been around for a bit.

39:04 - Jonathan Bennett (Host)
Yeah, and you know, there's sort of a broader application of this to not just distros but to open source in general. And so we've seen in the corporate world you see a lot of this, particularly in languages, and so languages like JavaScript and some of those, some of those ecosystems where you end up with a bunch of really small packages and, uh, you may have individual people maintaining some of those packages that get used all over the place. And you know, there's the left pad example. It's the right reason, was the right pad. I think it was left pad, um, but it was the kind of the big one of the first times this happened, big example of it.

39:54
So we got fed up with it. And he's like I want to get paid for my work. I'm working too much, I'm not getting paid, so I'm going to start pulling some of my repositories and broke, you know, a large fraction of the internet by pulling, by pulling left, oh yeah, off of NPM. And it's a problem, right, that you have so many of these little open source projects that don't have any sort of funding model, they don't have any sort of succession model. You know, none of this stuff exists. And it's not a problem that they exist because obviously people can write code if they want to and throw it out there. That's part of the beauty of this whole thing, of the way open source works.

40:36
But you do run into a problem where it starts getting used and it becomes a very important piece of infrastructure and there's nobody thinking about well, is this just one guy? Are we making sure that the guy that's propping up this corner of the internet has his rent paid? It gets back to that XKCD comic, the famous one, where it's like all of the internet, and then you have this one little piece that's maintained by a single guy in Kansas, and when that was written I'm pretty sure that was about NTP, the Network Time Protocol, but there are now countless projects that that's. That could um, could refer to, could describe.

41:15 - Jeff Massie (Co-host)
So well, and I think I think it's a little more salt on the wounds too, because I've seen some stories where they talk about you get companies like google making billions and they're like I can't get just a couple bucks or something or some help here, or can you throw me a bone?

41:38 - Jonathan Bennett (Host)
And it's not even so. With something like that, it's not even that Google is intentionally being evil and like we want to just exploit your work, we're not going to give you any money. No, they just don't know. Like nobody at Google even knows that the dude exists. Right, Like that's the problem. It's just so difficult to even catch somebody's attention.

41:59 - Jeff Massie (Co-host)
Yeah, and I didn't mean to imply anybody was being malicious. I was just meaning that you know T-MyCode's everywhere, but it just kind of works everywhere and it spreads, spreads everywhere and people don't realize the impact. Yeah, absolutely.

42:14 - Jonathan Bennett (Host)
Absolutely All right. So let's chat about the STEAM survey. Rob's not here to do this, so I'm going to do it. Rob, this is what you get when you're off for a week. I cover your stories.

42:34
The July STEAM survey is announced and there's some real cool stuff in there. Um for one, we are all the way up to uh 2.89. Sorry, the july steam survey. A month behind here. Um the july steam survey, we're up to 2.89. Percent of the people running the survey we're runninga linux desktop, which we're we're really edging close to that three percent mark, which is fun.

42:59
Um, there's some big shake-ups in the distros being run. Um, or it may be that we're now collecting data a little bit better, because the percentages by distro the other percentage is now only at 17.58%. That's a fall of like 3.29%, and we've got several distros here that basically went from 0 percent up to so, for example, fedora, fedora, linux 42, kde, plasma desktop uh sits at 1.79 and that is an increase of 1.79. So that makes me suspect that these were just falling into the other category before this. Um, the other Fedora 42, the workstation edition, was at 2.21% Uh, and then Bazite 42 from Fedora 42 is at 3.2% now.

43:59
Quite an increase there too. So Fedora is really holding down its uh, its fair share of the SteamOS, the Steam Linux market. Um, then of course, you've got linux and ubuntu and all, all of the others that you would expect on the list. Um, amd actually fell just a little bit and intel gained about a percent. So amd is sitting at 67.91 percent of the linux gamers and, uh or no, of all gamers, not just linux gamers and genuine intel is now sitting at 32.08. So, some interesting stuff in the steam survey, but, uh, another month of growth of linux users, which we always love to see. It's fun, it's fun to watch it and, uh, like I said, we're really closing in on that, that three percent mark I.

44:48 - Jeff Massie (Co-host)
I bet we hit it especially with windows 10 going and you know the sentiment for windows is just getting worse and worse. And realistically, you know linux now is getting a lot easier. You know it's it's not the all right. I got to be in the terminal I. I gotta know it's a lot of just point and click gooey and there's a lot of them, especially like kde, where and and there's other there's other desktops too that if you can run windows, you can figure out kde pretty fast. Oh, let me click on the corner icon. Oh, they're here. Here's my office suite or here's my web stuff or here, you know, yeah, very familiar yeah, so it's interesting to think about this.

45:36 - Jonathan Bennett (Host)
So, like microsoft is not doing itself any favors with with the overall user experience of windows, right, and they have not been for a long time, um, I think, I think the the windows experience really started to to suffer, we'll say, when they started putting ads in the start menu, when they started essentially downloading games automatically that were that were paid right, so you would, you would see, your, your, your dorky, essentially browser games in some cases were automatically getting downloaded to people's computers. Like that was the beginning of of the end for being able to take windows seriously as like a serious desktop. In my opinion, that was the beginning of the end. It's real hard to take it seriously when they're automatically delivering you adware. And then now you've got things like windows, recall, you've got the push to windows 11, which nobody likes. Uh, windows recall the.

46:35
The big problem with it is it's just, it's creepy, it. It touches people, uh it, it triggers the creepy response, which is not ever a good thing for a company. Um, and then on the other side you say, oh yeah, everybody's going to apple. Well, apple sort of game up gave up on gaming years ago when they said that metal is going to be the only game in town, and they just absolutely refused to support open gl or vulcan and everything has to be metal. It's like who's gonna? Who's gonna put up with that, and so I don't know. It almost seems like we're approaching this point to where, if you want to game on a computer, linux is about to be the place to do it, which is so wild. For those of us that have been watching this for a while, and particularly been trying to game on Linux for years, it's such a wild thing to see.

47:35 - Jeff Massie (Co-host)
Oh yeah, been trying to, particularly been trying to game on linux for years. It's such a wild thing to see. Oh yeah, well, and now some games run better on linux than they do on windows. And I think you know windows has a key in this, not only just because everything they're doing mac, like you said is they're kind of like, oh sure, you can game on mac, just go figure it out, you know, we'll allow it. And that was kind of the response. Windows gives Linux something. It needs a stable gaming API with DirectX 12. It does.

48:03
That's one of the problems we have in Linux is some of that stuff changes so much. You're like, ah, freedom, it's great, we evolve, but you can't write anything without having it changing all the time. And that's why even games, that a lot of times you don't have as many native games when you could, just because they, uh, they can look and say you know what, let's just do it for direct X 12, proton. We'll do it. We'll just make sure we use the proper directx commands so that it works in proton, or we'll just add into proton and so that the software will work. And then they got one platform they develop, for it takes, takes it over and yeah, you know, I we've talked about.

48:48 - Jonathan Bennett (Host)
We've talked about this before, but it is. It's interesting. So it's not a problem for open source games because somebody will get in there and fix it. But when you have a closed sourced game, particularly one where, like it's done, and they don't want to continue making changes, they don't want to, they don't want to push updates to fix things yeah, uh, you, you run into the same problem on windows. By the way, you know there are, there are games for older versions of windows that are almost impossible to run on modern um, on modern windows, right like windows xp games. Trying to get that running on windows 11 is a challenge. Anything older than that, I mean, I think you can just forget it right. So anything that's a 16-bit game just will not run on a 64-bit os inside windows at this point yeah, there's a lot of those older ones that you have to run it on linux.

49:36 - Jeff Massie (Co-host)
If it's going to run, that's where it's it's going to happen. Because they had so many custom libraries and special, you know they didn't really follow the standards as well as games do now, and so there was in linux was more apt to be able to load in those special libraries, in their, in their containers, you know, for like proton yeah, yeah, so it it is.

50:02 - Jonathan Bennett (Host)
It's kind of ironic, but I think it's actually accurate, that the best gaming api we have right now is the windows api running inside of wine, that that is where things actually work and will continue to work, which is wild.

50:17 - Jeff Massie (Co-host)
And I could see. You know as long as they don't come out with a new DirectX which I don't see why they would that you know that could be the new standard for Linux gaming. Is just, here's your API and you can just handle it. You know that that could be the the new standard for linux gaming is just, here's your api and you just handle it you know, if they come out with another direct x, what are they going to call it?

50:37 - Jonathan Bennett (Host)
who knows, it could be 758 or something you know but surely, I mean, they're not going to call it direct x 13, right well?

50:47 - Jeff Massie (Co-host)
well, it would follow their direct x naming convention. But if you got the xbox, then you know they could add whatever random number to it. Or you know we're gonna. We're gonna go up by one, we're gonna multiply it by itself, we're gonna square it, we're gonna, we're gonna take the cube root of pi to that number to come up with the next.

51:08 - Jonathan Bennett (Host)
Who knows direct x 144. There you go, that's the right name for it that's what comes after direct x 12, direct x 144 for the next xbox. There you go. It's brilliant. Be an xbox exclusive. Never come to windows. Well, and uh, eat.

51:25 - Jeff Massie (Co-host)
The oligarchs said you know, microsoft just wants xbox on linux. Well, I could see that, because you know the, the hardware, they're not basically kind of given up on. Sony won that one. But if you think, well, handhelds or other things, they want the gaming, they want the games. That's where the money is. And if it's like, oh, these all run on lin, on Linux, all right, we just get the cash from the games. We don't have to support an operating system. We, you know, makes it easier.

51:57 - Jonathan Bennett (Host)
Yeah, I don't think Microsoft has ever made money on the Xbox consoles. I think it's always been a lost leader to sell the games.

52:07 - Jeff Massie (Co-host)
Usually they do make money towards the end of the the run but, yeah, the first couple years they're losing money on them, but they make it up in the, they make it up on the sales of the games.

52:18 - Jonathan Bennett (Host)
It's like when they you know hey, this phone is free if you sign up for this plan they're losing money on the individual units, but they make it up in bulk yeah, ah.

52:28 - Jeff Massie (Co-host)
Yes, it's just like a service plan where it's like hey, okay, we're gonna give you this hardware for free, but guess what? You have to sign up for this contract. Well, you're paying for it. You're just not paying for it right up front. Yeah keith.

52:41 - Jonathan Bennett (Host)
Keith s 12. Ks 512 says it's going to be direct x 12 b. No, we got to follow the windows naming scheme. It it's going to be DirectX 12, 25h2. Service pack six. Service pack. Yeah yeah, servicing pack. All right, well, I'm glad we don't have anything like this on Linux. Oh wait, there's Wayland.

53:07 - Jeff Massie (Co-host)
And that's a segue. Right there it wrote itself it did. We just walked right into that one. Can it be proud of us?

53:16
So the next story is a Hackaday article and a YouTube video from Brody Robertson. So Brody put out a poll on several different platforms asking why a person was still on X11. Now he did say he wanted real reasons, not theoretical issues which you don't personally use. You know he wanted the actual. Here's what's a showstopper for me personally. And he said there was mostly good responses. There were a few which were off the wall, but 99%, he said, were usable and very clear what the problem was. And one of the first ones was the desktop environment of choice doesn't support it. Now this would be like XFCE Mate or a window manager that no one uses. Now the example he gave on the video was like one person was using Stump WM. I know he said he's like well, he kind of looked it up and it was. I don't even remember the language it was in, but he's like yeah, that's just not going to be one that's probably going to support Wayland because it's such a small project that it's not going to have the resources to build the Wayland support into it. But he does go on to say that's a valid reason. And there are some others which are still in experimental stages, like Cinnamon because it was called out. Somebody said, well, cinnamon supports it, but it's not very good and you know it's still pretty pretty bad support. Well, that's not Wayland's fault, that's Cinnamon because. And it's flagged as experimental, so they're not claiming support is where it needs to be. It's just, you know, people may be not realizing it's they're, they're just too optimistic and don't realize how flaky experimental can be.

55:17
Now, the next reason is because you're limited by your software. Now some users reported very specific, very specific programs they needed because of their job and they couldn't change out to others. Now think, think niche type programs and not mainstream ones which have known replacements. So it's like well, I can't use open office because I have to use which I mean could be. But a lot of the ones they gave were ones I hadn't heard of before, but it was required by their, their work, so, and those programs did not work on Wayland. So very specific use cases. Now they also brought up X11 display forwarding, which is used a lot of times in corporate settings. And you're limited by the software and hardware you're allowed to use, and that could be a reason. Brody goes on to say that you can display forward through Wayland and there's X-Wayland which can also display forward. It's just in this case the Wayland and X-Wayland options were not available. Again, corporate setting, where you know a lot of your software, is very locked down and very controlled. So you you can't just say, oh, just use this. And anybody in a large corporation when you try to say, hey, I want to get this open source software, yeah, good luck. If you do get it, it's going to take you probably a year to go through all the checks and you probably need to know a few people on the inside just to keep it moving.

56:52
Um, other things, brody goes on to mention the graphic tablet support, which does work on whalen, just not a hundred percent like it does on x11. So this, this is brought up as a known issue and, for example, he, he gave it, uh, he showed a posting on the kde website where they're they've got people working on it. Or I mean on the whalen website, they've got people working on it, or I mean on the Wayland website and they've got people working on it. So if you're a professional and need to have your graphics tablet work in a very specific way because you're a graphic artist and how it operates affects your art. That's also a valid reason, but know that there is a plan to have it fully working in the future. So it's, some of this is it's getting worked on, it's just not there yet.

57:36
Now some things were brought up which they blamed wayland. But there's some issue with you know, other desktop environments that are having problems or, like I said before, things like cinnamon wayland, where it's still flagged as experimental. So there there was some things that people listed that you really can't blame wayland, because the, the desktop environment is still building out support. Or they had another bug that people just filed under wayland, which was wasn't true. Uh, an issue which is a problem is playing games when you have more than one monitor. So when you're trying to play a game, a person never knows which monitor the game is going to land on, and brody does call that out. That that's a real problem, needs to be fixed. You know there's. There needs to be a way to set which monitor a program is going to open up on. So it that also he. He talked about that and you know you can do it on other operating systems, but we need to figure out how we do that.

58:39
On wayland, global hotkeys has been called out as a known issue. Now there is a solution, but he mentions it was written kind of in a vacuum and developers were not involved. So their hotkeys are not set up in the application, they need to be set up in a third-party app. And you know Brody goes on to even say with KD and GNOME they support this. But there really is going to need to be an official, supported way to pass global hotkeys. This current solution is considered to be kind of kludged together. Basically it's just kind of a in case non-native English speaking kludged together means it's a terrible fix. It's not a true, lasting, perfect fix, it's just we were able to get it working. So don't touch it because it might fall apart, but it works now.

59:37
The last one I'm going to mention is old hardware which, like, for example, somebody was using an NVIDIA 710. So they're stuck on the NVIDIA 470 driver which is not going to have the full Wayland support. The driver's just too old and there aren't going to be any kind of changes of the old driver. You know old laptops are in this issue with Wayland, which again fall into the old, non-supported drivers. I mean, well, it has supported drivers for the hardware at the time, but they don't support Wayland or they don't support it fully because it's just too old and even, you know, partially supported you'll never have the fullest, full, polished implementation. So take a look at the article in the video, in the show notes, and you know, let us know what you, if you're on still on X11, you know why. You know I left out a ton of reasons and thoughts from the sources. So there there's a lot of reasons that are very valid. So you know, love, love to hear it all. You know, put them on the discord and let us know.

01:00:40 - Jonathan Bennett (Host)
One of the things I wonder is people that have tried Wayland and had problems with it, or even have recently tried it and had problems with it. Are they running a? We talked had problems with it? Are they running a? We talked about this earlier? Are they running a bleeding edge distro like fedora? Or are they trying it on something like ubuntu, where they're running you know two or three versions back and seeing all of the bugs that were there when, when that version was written?

01:01:05
Um, that might make a difference, uh, to some of this. Uh, I I know there is always the whole. You have very interesting hardware and your very interesting hardware does not play well with everything else. Laptops were really bad at this for a while, but they would have you would have a video card that only existed for this one laptop. It was the only place where you could find this exact configuration.

01:01:30
I had one of those. It was a pain, pain. Um, you know you get into things like that well, and, oh, go ahead. One of the things to keep in mind is, no matter whether wayland or is ready or not, uh, x11 is basically unmaintained, and so that's, that's going to. That's going to be a reality check for distros here. Soon it's going to have to get dropped from places like Fedora that have strict rules about not shipping unmaintained software, and I know Ex Libre exists. I still have doubts about how well that's actually being maintained with only one main developer there, but who knows they they may, they may actually make that work, um, and we may see that they could be dedicated and stick with it.

01:02:22 - Jeff Massie (Co-host)
Time. Time will tell, uh. One thing he did talk about, though, is there was some friction in the whalen camp.

01:02:28
It sounds like there's, um, certain certain things that are done on like mac and windows and x11, that sounded like people kind of were pushing back against just because it's on windows. You know, like some of the graphical selections of like where windows open, wow, we can't, you know it's on windows, therefore it's bad, it's like no, it doesn't mean it's bad it wow, we can't. You know it's on windows, therefore it's bad. And it's like no, it doesn't mean it's bad, it's we. Just we should implement it because, you know, even even an operating system we don't like is going to have good ideas in it, and we shouldn't shun those ideas just because of where they came from. If they're good ideas, they're a good idea. Blind squirrel finds a nut every once in a while, exactly, and hopefully this, you know his video gets a little traction and maybe gets the wayland supporter, the developers, to kind of soften their view on certain uh topics of things and it's been an issue with what like that.

01:03:24 - Jonathan Bennett (Host)
That's probably the one thing that has annoyed me the most about wayland is that they have been very purist. We could say um, could say other things, but we'll be nice and go with purist about what they want to land inside the wayland uh protocol and what they think should exist outside of it. Um, and just the, the, the bike shedding and the wool gathering over some of these things. It is like you know this pull request. Some of these pull requests were open for years and just constant bickering and debate back and forth, like, oh no, I think it should work like this. Oh no, no, you shouldn't do that. Nobody should ever do that. Well, our OS wants to do that. Well, your OS is wrong. If you want to do this, like, oh, my goodness, guys, users want this, just get over yourselves and pull it in. And uh, what?

01:04:19
What actually happened is uh, and again, we've talked about this before but valve valve came along and said okay, we're just going to fork wayland, we'll start making our own protocol. They called it the frog protocol. And uh, not very long after that happened, wayland sort of got its act together for a while and it's like, all right, we're gonna we're to drop some of these developers that have really been causing problems. We're going to make some rules about what people are allowed to threaten to do. And you know, if you don't have veto actually have veto power, then you don't get to tell people you're going to veto their ideas. And yeah, it was. It was fun times. We finally got some stuff to land, though. So I don't know, maybe valve will have to step up again.

01:04:57 - Jeff Massie (Co-host)
It's like, let us tell you about our new toad protocol I well, I remember when that happened because they had single people that could veto, that really were just kind of a standard developer and it. Well, you know, and and I appreciate every open source developer but sometimes it's like, boy, I think, dude, some of you live in a bubble and it's like, hey, let's even even an idea you think is crazy. You should at least stop and, rather than immediately reject it, say, okay, what happens if we did this? What is, what is the real reason? What's what makes sense, what you know? Even critically evaluate something you think is crazy, because maybe it's not that crazy, you know, maybe maybe your ideology could shift a little bit and you could open your mind and go. Maybe maybe this would work, maybe this would be something that would be helpful yeah, there's a.

01:05:51 - Jonathan Bennett (Host)
There's a term for that is uh, is it called iron Manning?

01:05:54 - Jeff Massie (Co-host)
Uh, steel Manning, I think uh, opening up your thoughts like that or no, uh, the the opposite of a straw man.

01:06:04 - Jonathan Bennett (Host)
Um, oh yeah, steel Manning, right, and that's it reminds me very much of this idea. Um, where you know, a straw man argument is you take someone's argument and you sort of represent it in its weakest form so that you can easily defeat it. Steel manning is the opposite. You take someone's argument or their idea and, in trying to think about it, you kind of represent it in its strongest possible form and then evaluate it and it gives you a much clearer, like a much more, uh, a much less arbitrary view of what they're actually talking about. It lets you evaluate it a lot, a lot more clearly and rationally.

01:06:43
Um, I never heard of that term before it was introduced to me on a podcast. I heard somebody say it there and me too went and looked it up. Went and looked it up and kind of like the idea um, yeah, so it's. It's a thing with way wayland is literally designed by a, uh, a convention, by, um, a commission, excuse me and and you know there's a reason that that's such a stereotypical trope. I'm like, oh, this was designed by committee. Well, wayland is actually designed by committee. Well, wayland is actually designed by committee. And there are some of those problems that have popped up.

01:07:22 - Jeff Massie (Co-host)
I in my personal life I found that designed by committee usually does not work. It's better to throw something out there here's a framework of something, even if it's not right. Just put something out there that people can then kind of build off of and fix and you know, rather than just like, okay, we got a blank sheet of paper, let's all decide what we want. You know, you kind of need that initial direction to get things going and then when somebody's focused on something, then I don't know. I find it works a lot better in my personal experiences in leading teams and things like that Running projects.

01:08:02 - Jonathan Bennett (Host)
Yeah, makes sense. All right, let's talk about KDE. We've sort of danced around KDE a little bit and we've got the latest weekly update and there's a couple of fun things in here we can chat about, one of which and sort of the highlight for this week is the day night theme switching, and so if you want your machine to automatically go into dark mode once the sun goes down, you can now do that in KDE 6.5. And they've done quite a bit of theming work here. They've even got like the day and night theme versions of wallpapers, which is a lot of fun. There is now the ability to drag things to a panel itself rather than to a task manager widget for 644.

01:08:56
There's a bunch of bug fixes and crash fixes in 6.5.0. The Orca screen reader in 6.5 will now tell you when you have hit the caps lock key and which state it is in, and so that's pretty interesting. You can hibernate from the SDDM, the Simple Display Desktop Manager login screen, which you probably shouldn't be hibernating your computer, but if you really want to, you can do it from there. Now you know lots of little things, little bug fixes, stuff coming, you know. So they are working on 644 as their point update and 6.5.0 is about to come as the new big feature update and lots of fun stuff happening in both of those. Yeah, I think the day-night theming is probably the most interesting there. I think that might drive me crazy but it could be fun to play with.

01:09:55 - Jeff Massie (Co-host)
I think it'd be cool, because just at night it's going to just dim things slightly, so it's not so blinding. Oh, and they do mention still four very high priority plasma bugs same as last week and 23 15-minute bugs, which is down from 28 from last week.

01:10:12 - Jonathan Bennett (Host)
So they're gaining ground there, yeah, yeah eight from last week, so they were gaining ground there. Yeah, yeah, although it's always, you can't just, you can't just go based on the number of bugs, right, correct?

01:10:23 - Jeff Massie (Co-host)
because, because maybe eight got fixed and three got put in and right.

01:10:30 - Jonathan Bennett (Host)
But there's also the issue of like. Just because the bugs are not written up doesn't necessarily mean that they're not there, and so part of the process of fixing a bug is getting it put into a bug report that is in on the list, and so you know, your 15-minute bugs going up in one week may actually mean that you're moving in the right direction, because people finally were able to replicate the bugs and they got them written up and the process of fixing them sometime. Sometimes the process of actually fixing the bug is very simple. It's figuring out how to replicate it. That's the hard part. Often the case Very true. Often the case All right. Do you want to get into some command line tips?

01:11:13 - Jeff Massie (Co-host)
I think we should, but I don't, mine's kind of only sort of command line-ish.

01:11:18 - Jonathan Bennett (Host)
Mine is a very command line tip, so we'll even out for the day, okay.

01:11:24 - Jeff Massie (Co-host)
So this week's command line tip is OpenSnitch. Now, this is a port for Linux of the macOS tool LittleSnitch. Basically, it's a firewall which reports back network requests from applications, so then a user can create rules to block the request if needed. So why do you need this? Well, when an app says it doesn't need internet access and suddenly starts sending out network packets, why would it need to do that? Is it sending telemetry data that you know was unknown? You know it's not supposed to send telemetry, but it is. Is it nefarious software? Did somebody get into something and what you got isn't what you think you have? Or is it just simply maybe checking for an update? I mean, there can be legitimate reasons and there can be bad reasons. You know it requires more investigation, but it's helpful to know first. So the article in the show notes has a link to the GitHub page where they have the prepackaged deb and rpm files. Now, if you're not using those, there's of course source code, and so then you can compile it yourself.

01:12:37
Basically, you install the daemon and then you fire up the client GUI. You know, install that and fire it up and start the service. Then you open the GUI and you can start making rules for your applications. You can enable or deny the traffic or you can even limit it to a specific host. So, yeah, it's got to talk, but it needs to only talk to this specific IP. You can set that up.

01:13:08
I won't go into all the details. The article goes into various things you can do to customize the rules and a lot of details on how to use them. But if you just fire it up, you know just the basic deny or allow are pretty simple to use. You know if you're not getting anything too fancy or crazy. You know and with all the malicious software out there and you know people sometimes being a little rambunctious on their telemetry data you know being a little rambunctious on their telemetry data, opting you in rather than and requiring you to opt out versus hey, maybe I would like the opportunity to say yes, I would give this information out. So take a look at OpenSnitch and take more granular control of your system and give it a try.

01:13:55 - Jonathan Bennett (Host)
Yeah, very cool.

01:13:56
I'm curious and I'm sure this was probably not something that was on your radar when you were looking at this.

01:14:02
But what immediately comes to mind is I wonder how well this does defeating programs that sort of are intending to hide their network activity.

01:14:13
You know, they're sort of the root kit kind of approaches where you you do things to avoid getting on the radar of somebody like so, for example, there was a security story from this week where a new technique was found where malicious programs were bind mounting.

01:14:33
So in the slash proc directory that's proc directory, that's where most of these sort of stats for processes lives, in slash proc and there was this new technique that was found in the wilds malware was doing this where a malicious program was bind mounting another applications slash proc folder over its own, and so whenever these tools would go and access it, they would just they would go by default to the one that was bind mounted and so you know you would go to look up the dat, the stats on this process that you're trying to figure out, and it would go over to something that was completely benign and the malicious process was hiding that way. Um, it's just, that was a fascinating story to read about and sort of understand what was going on. Uh, but also with something like this, I always have to wonder well it would.

01:15:22 - Jeff Massie (Co-host)
It would probably show up in the firewall. It's just what. What would it label it as? And it you could, but you you could maybe catch it, depending on what it tried to hide, as, oh, I'm hiding under the web browser and you're like, but I don't have my web browser open and this thing is pinging, it's like, and maybe start looking at the addresses that it's sending packets out to, or you know, yeah, so that's stuff you can't hide.

01:15:52
Yeah, I think you could probably catch it, but you'd have to be paying close attention and checking out where's this actually going.

01:16:05 - Jonathan Bennett (Host)
Absolutely, absolutely All right. I've got something that's very different. I had a problem show up this week. I was working on someone's Windows desktop. We're trying to get a Windows 10 to upgrade to Windows 11. And it's like the hardware should be able to do it, but the problem was that it got installed as an MBR Windows install instead of a GPT Windows install, and so we had to run. Microsoft actually has an application that will let you do this upgrade, by the way, to switch MBR to GPT, and I thought this machine has important stuff on it.

01:16:41
I really don't want to kill this thing by this Microsoft tool going crazy in the middle of doing this. It's like, okay, let's get a backup of the hard drive. Usually what I go to is DD Rescue to pull a hard drive image off, but it's like, okay, it's a one terabyte drive but they're only using like a hundred gigs, but I want to get something that's a disc image so I can just go directly back over the hard drive if I need to. So I was. I was sort of looking around for so go, what tools are out there that would let me do this? And I ended up with a very different usage for a familiar tool and that is QEMU-IMG. That is the image manipulation tool for QEMU, and one of the things that it can do is it can convert between different image types, and one of those image types that it supports is the raw disk image, which is essentially what you've got with a physical hard drive. That is essentially a raw disk image, just backed by an actual drive, as opposed to a file. So you can use QEMU image to convert from the raw over to something like a QCAL2, which is going to be a sparse image, so it's not going to take up the entirety of the one terabyte. But the other thing that you can do with this is that you could tell it hey, in this QCAL2 image I want you to use Z standard compression. And so I ended up with a much smaller disk image saved on my RAID hard drive and was able to maintain a backup, and so if something had gone wrong I could have flashed it right back over. So I've got the whole string here in the show notes.

01:18:24
It's QEMU image. Convert dash C I don't remember what all of these flags do, unfortunately. Dash C, dash P, dash capital S 512. It's set in the sector size, uh, and then dash o. Compression type equals the standard um, dash f for the incoming file type, we set to raw, and then dash capital o for the outgoing type, which we set to zcal 2. Uh, dash m8, I think that sets how many parallel threads it uses and then you give it the physical location. In my case it was slash dev, slash SDG and then the imageqcal2. So I used that this past week, this past weekend in fact, doing a Windows 10 to Windows 11, the first steps of that upgrade, use this to be able to get a good disk image for a customer. And again, I was very intrigued because it has a very different use than what we normally use QEMU image for. But it works.

01:19:28 - Jeff Massie (Co-host)
That is cool. That's pretty hardcore command line there.

01:19:33 - Jonathan Bennett (Host)
I didn't entirely come up with it myself. It was based off of a. Was it a Reddit thread or a? I forget the name of the other place, stack Overflow, like either Reddit or Stack Overflow, one of those two, but yeah, I was.

01:19:47 - Jeff Massie (Co-host)
I was pretty impressed when I found that and adapted it to my own needs, so yeah, it's interesting how some of those commands like I did not know you could do this. Yeah, it's interesting how some of those commands like I did not know you could do this Yep, or use it this way, absolutely.

01:19:59 - Jonathan Bennett (Host)
All right, jeff, you had something fun come up this past week too, didn't you I?

01:20:05 - Jeff Massie (Co-host)
did. So Rob has been gone for the you know, like last week and this week because he's on vacation. Him and the family are going to the coast. Well, he actually came by where I live it well, in the metropolis, metropolitan area. So he, I drove about 15 miles to meet him and we met in Nampa, idaho, at a little bar and I, I got my, I got my money, so everybody that donated coffee in my name, I got it and we actually had a couple beers and we talked for probably a couple hours. It was a good time. We had to keep it kind of a little bit short just because he had to get up early the next morning. They were still driving west to hit the coast.

01:20:55
But yeah, it was great meeting in person and we had, we had a lot of fun and it was uh, a nice, uh nice time and I bought a guinness with that first, uh, coffee money. So if you're wondering whether it was kind of coffeeish, you know guinness, same color. Yeah, yeah, I, I buy it. It's a stout, yep, extra stout, but uh, good stuff, very cool, but but yeah, it was great actually being able to meet in person. So now, now we got to figure out getting down to jonathan and ken yeah, ken and I've done a meetup.

01:21:29 - Jonathan Bennett (Host)
Uh, because he's not that far away from me, I got sent up to oklahoma city for a work call and I'm like, hey, let's do lunch sometime. So ken and I, we grabbed lunch. This has been probably a year ago now, but uh, we did, we hung out together.

01:21:39 - Jeff Massie (Co-host)
But you guys you or Rob ever get down this way, we'll definitely have to do it, but we we talked about sometime maybe meeting up down there, cause between where he lives and I live, it's actually you'd think it's closer for him, but it's really almost equal distance, I mean. I mean within you know, I think it was like 50 miles or 100 miles or something, you know it was. It was pretty close where um same distance, come down and maybe we get a, get a show where, uh, we're all, we're all in the all in the basement with jonathan we can do it.

01:22:14
We can do it, I've got the stuff to do that oh, we'll have to tell you, and then you can start panicking about the sound stage and how I'm gonna have all everybody in here and not cross talk and echoes and cross talk and echo is the difficulty.

01:22:27 - Jonathan Bennett (Host)
Yeah, yeah, might do it. Might do it real old style and just gather around a single microphone. Oh, that would be cool, that'd be fun, all right, do you have anything else you want to plug?

01:22:39 - Jeff Massie (Co-host)
That was kind of my, that was kind of my plug. I was I was not going to do a poem, cause I just want to talk about, you know, getting the getting the coffee money and meeting up and it was. It was a great time. We both had a blast no-transcript.

01:23:18 - Jonathan Bennett (Host)
And then, if you also want, you can check out Club Twit. That is definitely the other thing that we want to plug and let folks know about. It's not much more than the price of a cup of coffee or two per month and you get ad-free access to all the shows and behind the scenes sneak peeks, all kinds of good stuff. So if you're not part of Club Twit, you should definitely check it out. All right, Thank you everybody for being here. We appreciate everyone that's here, those that get us live and on the download, and we will see you next week on the Untitled Linux Show.