Untitled Linux Show 161 Transcript
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)
Hey, this week we're talking about AMD video drivers and ray tracing. We give you the scoop on the Linux kernel version 6.10 release. And then there's the SUSE, opensuse, kerfluffle, there's more Rust in Linux and, oh yeah, about half of the world's computers showed up beside. We talk about that too. You don't want to miss it, so stay tuned.
00:22
Podcasts you love, from people you trust this is this is the untitled linux show, episode 161, recorded saturday, july 20th. A mouthful of laptops. Hey folks, it is saturday and you know what that means. It is time for the Untitled Linux Show, where we let our geek out and we talk about Linux and open source. And we're going to be talking about a lot today, even some Windows stuff. You may have noticed something really interesting that happened. What was that? Friday, sort of the Windows computers around the world blew up. We're going to tell you why and what's going on with that, and we're going to revel a little bit in the fact that our Linux machines didn't go down and so kind of a victory lap for us. It's not just me. I've got Mr Jeff Massey with me today. Jeff, how are you doing?
01:13 - Jeff Massie (Host)
I'm doing good, excited for the show.
01:16 - Jonathan Bennett (Host)
Yeah, yeah, it's going to be fun. Now it's just the two of us, so, which it's fine, we can do this. It's just it's a little bit different of an energy, a back of the back and forth. It's a little different.
01:24 - Jeff Massie (Host)
So well, not not necessarily, because, you know, in today's ending notes, my little plug is going to reveal a little something where you and I did something earlier in the week together. It's true, that's true.
01:39 - Jonathan Bennett (Host)
All right, well, let's, let's dive into some stories, and you've got some ray tracing to talk about, don't you?
01:46 - Jeff Massie (Host)
I do. I kind of got a little bit of a graphics corner going on here. So it's two different stories. So the first one is last week we talked about a new ray tracing benchmarking. There's a paid version with more features for the true professionals. You know some of the automation and deeper analysis and things like that, if you really want to get into it and willing to give up a little bit of money there.
02:25
Nvidia, and while the results are what we've always expected, with team green taking the win and Vulcan ray tracing. Michael Arable over at Pharonix mentioned that the next week he was going to take a look at running the open source AMD driver versus the AMD official driver. Well, this is the next week and the time has come to see the results of the benchmarking. And just so everybody's clear, the community driver is also known as the mesa rad v driver, radv, and the official driver is the amd vlk driver. So that's that's how I'm going to uh reference them through the rest of the story here. So michael mentioned that the amd drivers. You know that both of them had come a long way in the Vulcan ray tracing field and that's why he wanted to revisit this comparison, you know.
03:08
Basically see what ground has been made up and you know, honestly, the results are pretty straightforward when ray tracing was turned off, the RADV driver so that's the MESA community driver took the lead, but when ray tracing was turned on, the AMD VLK driver took the trophy. So it's looking like there's still improvements that need to be made. Michael mentions that he's looking forward to more improvements in the Mesa RADV driver, just as there's still stuff coming down the pipe for it. The version he was testing, so in case anybody wanted to know or duplicate it, he was in the 610 kernel and using the MESA 24.2-development release. So stay tuned for more on the driver development and more benchmarking as things develop with that. So he's kind of mentioned that there's still a lot of work going into the Mesa driver to work on ray tracing, on Vulcan. So we'll keep you abreast as more things land in future releases.
04:15 - Jonathan Bennett (Host)
So real quick, I want to say, and I'll let you go on, but I remember when that ray tracing performance was terrible and completely unplayable on that amd driver and the, the, the strides they have made and say the last year, I've just been really incredible, and so I, I don't imagine it's going to be too much longer, say another year and we'll hit parody between those two.
04:39 - Jeff Massie (Host)
I would, I would imagine oh, I, I think it's coming, and michael didn't come out and say it, but it sounds like he's kind of hinting that that's what's coming in in future. Uh, kernel releases, yeah, yeah, I would say so, all right continue.
04:52
And well, and something to note, and we we talked about it last week, so I'm just going to only touch on it ray tracing isn't necessarily the end all be all for everybody. So, you know, make sure you evaluate, or evaluate what's important in games, because some games don't do much with ray tracing. You know slight puddles, you know some puddles look better, uh, sometimes a little bit in shadows, but you know, in other games it makes a difference. But take it, take it with a grain of salt, indeed, indeed, take it. Take it with a grain of salt, indeed, indeed. But second story, and we're speaking of open source here, nvidia has finally done what they said they were going to do with the in the 560 driver and they said they were going to fully adopt open source in the kernel. Now that that last few words is important in the kernel. Now, this actually that last few words is important in the kernel. Now, this actually started two years ago and they've been working on replacing their proprietary driver with the open-source one. A few months ago we even talked about how it was possible to switch over to the open-source drivers, but they were not the default. You had to have some settings to actually turn it on. Well, now they've done enough testing on them to say that, going forward, the open source ones will be the default. Now, while this is good news, it does come with some buts, and some of them might be big buts for people. So Maxwell, pascal and Volta architectures are not yet compatible with the open source GPU kernel modules, so NVIDIA recommends using the NVIDIA proprietary driver. For those older architectures. The open source and proprietary kernel modules are known to work with the Turing, ampere, ada, lovelaceace and hopper, but the newest architecture. So grace hopper now it's different than hopper. Grace hopper is different. And blackwell will only work with the open source driver. So you know, depending on your architecture you're running, the age of your graphics card is going to determine what you can run.
07:04
And next is a quote from NVIDIA about this. They said two years on, we've achieved equivalent or better application performance with our open source GPU kernel modules and added substantial new capabilities. We are now at the point where transitioning fully to the open source GPU kernel modules is the right move and we're making that change in the upcoming R560 driver release. Now, one thing I do want to mention. Here's another part of that, but this is the kernel code. So things like NVIDIA's user space library, such as OpenGL, vulkan, opencl, cuda those drivers remain closed source and right now, at least as far as I've heard, there's no plans to open source them. So it's only going to affect what is compiled into the kernel. So if people are thinking, oh, now we can see how they're running CUDA and things like that, that's not going to happen, or at least not in the near future. Take a look at the articles in the show notes for all the GPU driver. Goodness, your heart can stand.
08:13 - Jonathan Bennett (Host)
We've got some other interesting stuff here to take a look at, and I'm going to take it next and talk about something else actually going on with the NVIDIA drivers, and that is in Fedora. There is a bit of a change coming. The NVIDIA proprietary driver stack was causing problems in Fedora when you had a UAFI secure boot turned on, and that was the idea was, when you, when you install the nvidia driver, you have to install the driver module and secure boot was seeing that it kind of freaking out, going this is not a properly assigned module and then cutting things off, not letting you boot. And so apparently, when you went through the infidora, you would go through the GNOME software application and you could install it there. In some cases it could result in an unbootable system, which you know is no bueno. That's not what we want.
09:16
They've looked at this and they figured out a solution, but it's one of those solutions that really nobody likes. They're going to do it because if they don't, it's going to really break things, but nobody really likes it. So what they're doing is if you go through this and you install the NVIDIA driver this is for Fedora 41, by the way so if you go through GNOME software and they're looking at adding this to KDE Discover as well. But in GNOME software, if you go through andome software and they're looking at adding this to kde discover as well but in gnome software, if you go through and you install the proprietary nvidia drivers, it's going to give you a prompt to sign the driver with a local key and then, because the the I believe the way it works is you know Fedora has the key, the assigned key, and then they can let you sign your own driver module and you know they've got the cryptography worked out to where this actually works. So what they're doing it's enrolling MOK is the term, so you enroll the Ks and then you can sign the driver and then when you reboot, it actually works rather than failing to boot, and I just I love the.
10:36
So they had FESCO vote about it. It passed 62. And there's this note on it. It's like, while nobody is enthusiastic about this, the consensus among those who vote in favor is that waiting will just make things worse, as will lag behind other distros and the chance of swaying NVIDIA is very minimal. So it's kind of not the most exciting, excited way to address that. But you know, at the same time. Hopefully that'll fix things for some Fedora NVIDIA users, and then, you know, there's some hope for the future. As Jeff just mentioned, we do have some open source drivers that are getting better and better and maybe one of these days this will actually land in the upstream kernel and then we won't have to do any of this with NVIDIA. So we'll see.
11:27 - Jeff Massie (Host)
Well, hopefully the audio is a little better oh, absolutely I did a little resetting there, um, and I would say it's not totally true that you can't change nvidia's mind, because kde did it so with with their EGL streams? Okay, yeah that's true, although it took years.
11:49 - Jonathan Bennett (Host)
It took a really long time. Yeah, it took a very long time and that wasn't necessarily them, I mean. I suppose that's what happened, but it's just more like you're not going to force your ideas on us. It doesn't bother us at all to not have, you know, the support for the way you want to do it and we don't care. So either get on board or don't have support for it. And like seven years went by and NVIDIA finally went okay, fine, we give up, we'll do it your way.
12:22 - Jeff Massie (Host)
Yeah, that is. That is true. I mean you're, we'll do it your way. Yeah, that is true. I mean it's pushing a rope to try to change NVIDIA's mind. For the most part, yes.
12:32 - Jonathan Bennett (Host)
Yes, all right.
12:36 - Jeff Massie (Host)
But they're making the right move.
12:37 - Jonathan Bennett (Host)
Yeah, Now what do you think? Are we eventually going to get to the point to where NVIDIA's kernel driver is upstreamed and we just don't have to mess with this anymore?
12:51 - Jeff Massie (Host)
I think so. I think that's where it's going to be, because and I've I don't know how exactly, but I'm sure it's tying into ai somewhere and they're they're they're thinking, okay, we've got to do this to tie into certain markets. That's going to make us, you know, more approachable in these certain markets, because we know there's some AI companies that are really pushing for open source and NVIDIA is not going to lock themselves out of a piece of that cash pie.
13:16 - Jonathan Bennett (Host)
Yeah Well, so we know we've gotten a little insight in this over the last few weeks, because I believe it was NVIDIA that came out and said that they have some customers that have asked for this. Honestly, I think it's just big businesses have also found that it is a pain to install out-of-tree modules. I think that's what it boils down to. It's just so much of a barrier to entry, especially when you have 1,000 or 2,000 or 3,000 servers, and sometimes scripting this sort of thing is difficult and you don't want to hose 1,000 servers or more at a time. We'll talk about that here in just a minute.
13:58
Yeah, you don't want to do that on Linux. That happens enough in other places. So that's sort of the driving thing that's pushing it, and so I Well, and free support.
14:09 - Jeff Massie (Host)
I mean, yeah, NVIDIA is still doing a lot of stuff, but they're probably going to get commits and things like that to say, hey, here's how you can make this better and fix it. And they're thinking, oh, we get free developers too.
14:21 - Jonathan Bennett (Host)
Some of it. Yes, although one of the other things to keep in mind amd has done this too to some extent. Um, they have pushed all of the interesting things out of the kernel driver and into the closed source firmware. Amd did it when they open source theirs to some extent, and nvidia is definitely doing it in this open source um endeavor, and and so there's a limitation on how much the open source community can fix, because they've moved so much of their code into that closed source firmware which you know. I mean that's fine, that's an approach. There's nothing necessarily we would love to be able to see inside the firmware, but there's nothing inherently wrong with doing it that way. But you kind of just turn your kernel driver into a remote procedure call engine and that's all that's going on here. So that does limit the realm of things that the open source community can come along and fix for them. Something to keep in mind.
15:20 - Jeff Massie (Host)
Very true, that's a good point, yeah.
15:23 - Jonathan Bennett (Host)
All right. Well, something rather interesting just happened this weekend. We have a new kernel, and I think Jeff has the rundown on what's new in kernel 6.10.
15:34 - Jeff Massie (Host)
I do, or at least some pieces of it, because it's big. So the 6.10 kernel is officially out, though Linus Torvalds wasn't exactly thrilled with the last few days. He had this to say about the release it was also. He said it also wasn't noisy enough to warrant an extra rc. So here we are with the latest linux curl kernel, ready for work. Because you know, there was some noise, but he felt, yeah, it's enough.
16:04
This release has a significant performance fix for Intel Core hybrid systems with buggy firmware. On systems with an Intel Core i5-13500H processor and using the earliest eligible virtual deadline, first scheduler, eevdf. That's a mouthful. Users saw as much as a 50% performance hit. Well, that's gotten fixed. Amd has been included with better Rock M and AMD KFD, support for smaller Ryzen APUs and new additions for the upcoming Zen 5 architecture, which we see that quite often, where they put code in the kernel before the actual hardware hits. So when the hardware shows up it's ready to go. Rust support has been updated to Rust 1.78.0. And the ARM architecture receives support for Clang CFI, which are Control Flow Integrity. Support for clang cfi, which are control flow integrity and plae privileged access never support. And there's a new nt sync subsystem for providing windows nt synchronization primitives for linux and wine gaming ah yeah, that finally landed, that's's interesting.
17:23
Yes, that should be nice. This latest kernel also introduces MSEAL. Now this protects a given virtual memory range against modifications such as changes to their permission bits, and it also implements Trusted Platform Module, or TPM as it's also known, bus encryption and integrity protection. There is also a new boot option called init underscore M locked underscore on underscore free. That will zero any pages locked into RAM when freed. Now we also have implementation of other hardware which has been included. Of other hardware which has been included, some interesting hardware support added includes the RADXA or R-A-D-X-A ROC 3C development board. Intel Arrow Lake H processors. Lenovo ThinkBook 13X Gen 4, lenovo ThinkBook 16p, Gen 5s and.
18:35
Lenovo ThinkBook X13 laptops, asus ROG 2024 laptops, laptops asus rogue radic carry pro controller and machine ike g5 pro game controllers. So lots of laptops and game controllers and development boards. It's pretty cool, yeah, and it's also when it's a mouthful, sometimes it kind of trips me up a little. Uh, intel got more preparation for upcoming XE2 graphics and the upcoming XE2 architectures supposed to significantly improve computational capabilities by providing up to 67 tops and offering increased ray tracing units compared to the XE-LPG on Meteor Lake, lpg on Meteor Lake. So we should see a big improvement from Intel, or that's what the rumors say. We'll wait for the benchmarking before we have final judgment. There is the new Panther Graphics Direct Rendering Manager DRM driver. This supports newer ARM Mali graphics processors and this development is particularly significant for the next generation of devices built around ARM-based architectures and it will improve their graphics performance and compatibility.
19:53
While there's a lot to love in this release, you need to take a look at the article in the show notes to see even more items I didn't cover. I mean, if you ever have a chance, look at at least one of the kernel releases. There are a ton of different things that go in and there's going to be some. Unless you're a kernel programmer, probably you're going to say I don't even know what that is. I know there is for me, but for those that want to get it, there is a link in the article to the kernelorg site where you can see the download and get the download for the latest kernel.
20:30
Now, ideally, you want to hit your distribution, but if you like to compile yourself and go at it because your distribution is too slow, that's the place to go. If you are going to get it, enjoy it fast though, as support is going to end in just a few months, november 14th of this year to be exact. The 611 merge window is now open and the expected 611 release should be around mid or late september. So you'll have 611 going for a bit before 610 is, uh, sunsetted. You know that now that mid to late september is assuming normal kernel release timing is adhered to, you know, things could go sideways or things could go really well. Time time will tell.
21:13 - Jonathan Bennett (Host)
But happy compiling, yes, um and it's interesting, there's already some some nifty things stacking up for 6.11, uh like. Uh I I've seen one story about in a particular case there's like a 20% speed increase for EXT4 and some other crazy stuff like that. So things march along and continue to get better and better. There was one thing in particular that I saw in here oh, so the Windows NT synchronization primitives. That is fascinating.
21:46
Um, it's not clear whether you'll see a whole lot of performance improvement if you're running proton, because proton already does a version of this in user space, and this is one thread that I was reading.
21:59
This is what they were suggesting. Proton already has a version of this in user space, but this adds it as a system call and then that will be available. Ideally that will get merged into upstream wine and so it's really going to make the biggest performance impact on upstream wine, on vanilla wine, although you know there's always the possibility that doing it as a kernel system call is going to be much faster than emulating it in user space. So, regardless, that'll be interesting to see. And I'm also curious like is is proton going to just, is it already in there and proton is going to do some sort of test like, am I running on a new enough kernel to have this, so could you? Could you go and compile, you know? Could you get 610 and proton automagically is better, or do you have to wait for the Proton update to go with it? I'm not sure.
22:50 - Jeff Massie (Host)
Yeah, I don't know, but I could see where they would want to. You know, implement this just from. Oh, this is one less thing we have to worry about in the code stack. It's taken care of in the kernel. Let them handle it.
23:04 - Jonathan Bennett (Host)
Yep, absolutely we can worry about other things.
23:07
Yep, absolutely okay. So, uh, speaking of kernel, this lines up rather well. It's almost like we did that on purpose. Um, there is something very interesting happening in the kernel, and so something jeff talked about just a second ago was the rust update, and because Rust is so new, they're kind of marching along. You know, you have an updated kernel, you have an updated Rust version, and those just kind of march along together. They are pushing to and hopefully this will happen very soon that instead of having to have those two linked together like that, there will just be a as of this kernel. The minimum Rust version is this, and then there'll be better compatibility between the two, where you don't have to have an exactly matched version, you have the same thing. Actually, that happens with C and GCC. That's actually one of the reasons that the Linux from scratch project was developed, because it was so difficult to get all those things matched up. You had to have the right version of gcc installed, you had to have the right kernel installed, and because your kernel wouldn't necessarily build with too old or too new of a version of gcc, you would have just have problems. Um, and so we're. We're seeing that again. So and and I will add on top of that, that is less so now because gcc and the kernel they've been doing this for so long They've kind of synchronized and made all of these things kind of default. It's all been worked out now for so long, so that'll come along with Rust as we're looking at things changing in the kernel. With Rust there was a really interesting Linux kernel mailing list and that is they are looking to provide a Rust safety standard for the kernel.
24:55
Now what are we talking about? Well, remember, one of the big advantages of Rust is that it's memory safe. So at compile time we'll check and make sure that your code, no matter which way you know, no matter where the inputs are, no matter which way you turn it, it's not going to crash because of memory access errors. You can't accidentally do a, you know, a null kernel dereference or writing past the end of a buffer. Like those things are just because of the way Rust is put together. Those are impossible. Like those things are just because of the way Rust is put together. Those are impossible. That doesn't necessarily work in kernel space, because when you're writing device drivers, literally the way that device drivers work is that you're reading and writing to memory locations. That's not real, it's not RAM, it's not system RAM, but you're actually reading and writing to other hardware devices. That's on your system's memory address. That's the way kernel drivers work and that just does not work with Rust's safety guarantees.
25:55
So the way this is handled is that Rust has the unsafe keyword and you can say, hey, this portion of my code is unsafe, I'm going to do these things, and you just have to be able to do that to write kernel drivers. The issue with that is that once you go into Rust's unsafe land, you have exactly the same memory problems that you do in C. You can write to places you shouldn't write to. You can read back uninitialized memory. It's what it says on the tin. It's unsafe.
26:29
So it's been known for a long time that to do this and to get any of the advantages out of Rust that you would want, you want to carefully control what portions of your driver and your code are marked unsafe. So that is what this is starting to do. They're writing out a document of. Here's the way the kernel is going to handle this. Here is the way that you are to do unsafe and Rust code in the kernel.
26:57
And now this has not been adopted yet. It is an RFC, but you know there's, I think, five patches here and quite a bit of conversation going back and forth about. You know, some of it is simple formatting things, some of it is, you know, improvements that can be made. But it looks like this has been fairly well received and so I would imagine maybe even for 611 we will see this get, uh, get pulled into the kernel. And again, this is not code, this is just a document, a, a um as a code standard, and this is the standard for how we do safe rust in the kernel and we've been talking about a lot.
27:40 - Jeff Massie (Host)
you know that rust in the kernel. There's a lot of foundation and prep work that has to be done before it's so much more than just putting code in. There's a lot of like this standards, and how are we going to do it? How are we going to tie in what's the proper way, you know, and even cleanup to make it happen?
28:05 - Jonathan Bennett (Host)
Yeah, and not even just that, but also development happening on Rust side, like things had to change in Rust For this to even be possible. So it's been a lot of work to make it happen and I don't think there's actually any Rust Like everything that's gone into the kernel so far. I think it's just prep work. I don't think there's any Like actual Rust code being run in the kernel yet.
28:35 - Jeff Massie (Host)
Unless I've missed where somebody wrote a module that got included. I don't think so.
28:48 - Jonathan Bennett (Host)
I'm trying to think, because so much of it we've talked about has been just this kind of prep work and design work or getting existing code ready to accept rust. Yeah, yeah, because. And so what they're having to do is they're essentially having to write um adapter code, you could say, because the kernel has been in c, so all of it uses these, these c data structures, and so you have to write adapter code to be able to access those C-style data structures in Rust. And so there's been a lot of that that's been added, but I don't think anybody has added a new module. There's no new drivers that are actually Rust code, but soon it'll happen soon. I know people have been working on it. There's been a lot of example code that's been written, but I don't think anything new has been done in Rust yet.
29:29 - Jeff Massie (Host)
But it's coming Well, and if somebody knows different, let us know. On the Discord server.
29:32 - Jonathan Bennett (Host)
Oh yeah, absolutely, but we're talking in kernel not.
29:36
Linux. Indeed, you can run Rust core utils if you want to. In fact, if you go check out this week's Floss Weekly, that's exactly what that was about. We talked with Sylvester Ledger about the Rust core utils and that was a super interesting discussion. One of the things that surprised me so much about that is the reason they're doing the Rust core utils like. The whole point of that is not anything to do with security. Like in the 50 years, no, in the last 20 years. I believe this is a statement he made. In the last 20 years of the c core utils, there's only been like 13 cves found in all of those programs. It was really impressive how well put together those programs are um, yeah, they originally started in 71 yes, they've been around for a very long time 53 years.
30:23
Yep, yep, yep. All right, well, let's see. Do we want to tell, oh, oh, this story do.
30:32 - Jeff Massie (Host)
Do we want to save that one for last, or should we just hit it now?
30:36 - Jonathan Bennett (Host)
no, go ahead and hit it. This, this is. This is special. I'm a little disappointed by this one. Tell us what susE and OpenSUSE the distro formerly known as OpenSUSE is doing.
30:49 - Jeff Massie (Host)
Well, there's a disturbance in the force SUSE. The company has officially requested that OpenSUSE to stop using its brand name. So now there's a difference between the two Linux distributions. Suse is an enterprise-based solution and they play in the same field as red hat. Open susa is a community driven distribution, but they do get a lot of support from susa. So susa provides a lot, you know, resources and code and things like that. So, and that's so. Here's where the rub comes in.
31:23
Opensuse has said that SUSE's request was calm and respectful, so they weren't being mean about it, they were very, very nice about it. But it could have implications on their symbiotic relationship. One of the big questions is how much help do they give each other? Now the support that SUSE provides OpenSUSE is obvious. The question is how much does the desktop open distribution feed users and companies into the paid support of SUSE? So at the OpenSUSE conference you know there were discussions about governance. Issues that were brought into view had some issues and it's caused the senior managers and budget holders of the corporate SUSE to express concerns about what's going on and they're saying there needs to be changes with how the open source project governs itself, while officially SUSE says they are going to continue to support the community project. They're also at a crossroads. They're also at a crossroads. Opensuse needs to either adapt to the evolving landscape by embracing change and addressing its internal challenges, or it can continue on a path that might lead to its obsolescence. Do with the code that SUSE developers have contributed and the infrastructure provided. The thought is, if SUSE withdrew support, it would mean a reduction in the number of different flavors of OpenSUSE, because they've got a few different flavors, like tumbleweed and stuff like that, depending on what you want, which could also mean that priorities would have to shift to focus on what's deemed most important, basically saying, okay, we've got a bunch of different flavors, we might have to reduce and we've got to refocus everybody onto the most important tasks and some things are just going to have to go by the wayside.
33:39
Now there is a lot of confidence around that OpenSUSE will rebrand and they'll find a new name and logo. I do have a second link in the show notes, which is a link to an OpenSUSE mailing list where there is talk about what should be done and there's talk about people working hard but burning out and leaving the project, not getting replaced quick enough, or those they do bring in, start enthusiastically and then get burnt out because there's a lack of onboarding and mentoring. There are suggestions on there how to refocus the board and teams to make this change happen efficiently and work on the other problems. There's also suggestions that a small team even rebranding a small team could come up with three to five new names and branding ideas so the larger community can pick what they like the best. Basically, they want to stay away from design by committee because that never works. Some even look forward to a new logo with a new color palette that works better with print and digital screens. So not everybody's looking at a rebrand as a bad thing.
34:49
As you can imagine, there's a lot of wandering conversations as people reply and counterpoint arguments. There's even one person who says they believe it's just a proposal and not an official request, though most are taking it as an official request, but I'll let the listener follow the link in the mailing list as it's a very, very lengthy discussion. Looking at the comments around the internet at various places, the speculation of what this means is all over the place. Some say it will continue on just fine, with no issues. People bring up Fedora and Red Hat. They have different names and are doing okay.
35:29
Some think this will kill OpenSUSE. Losing the name will be something it can't come back from. And then there's the third the people who think with the shift in the name and the logo, the two distributions will drift apart and, you know, one or both will fade away. There's also a ton of wordplay on what the new name should be. You know there's some interesting stuff out there, but you know, jonathan, I am sure you have some thoughts on what this. You know what the future is going to hold for SUSE and OpenSUSE.
36:01 - Jonathan Bennett (Host)
It's unfortunate to see there are questions here about what's actually going on, because I don't know that. Let me put it this way there are things at work here that are not being explicitly stated. Right, that's obvious. There is something going on at SUSE that has led them to make this request, and it is not at all obvious what that is.
36:28 - Jeff Massie (Host)
And I did some digging about what the problems were with OpenSUSE and people referenced a lot of stuff but nobody came out and really stated it other than, you know, burnout of people and not getting people replaced quick enough. But there was still other things you could tell were going on, but nobody you know it's, I couldn't find anything.
36:46 - Jonathan Bennett (Host)
You know they're keeping it fairly quiet yeah, uh, one of the other things that they mentioned is, uh, that they are losing contributors, and I I found that to be a fascinating thing to for someone to say, and I'm looking now at one of the uh, one of the threads where someone is is leaving the open suza community, and there's something really fascinating here um, he says, unfortunately in recent years, this is uh, let's see what's the name.
37:16
Jan jan ritzerfeld says um, almost 20 years ago, I subscribed to the project's german mailing list. Unfortunately, in recent years, I've witnessed numerous discussions here ending poorly, while each incident could be attributed to misunderstandings or assumptions. The repeated assumption of bad faith and the subsequent aggressive enforcement of the code of conduct, including personal attacks, harassment and violent language, contradicts the guiding principle of working together in an open, transparent and friendly manner. That's really interesting that someone is leaving because of overambitious code of conduct enforcement, and I don't know that. We've seen a whole lot of that. You know, maybe explicitly stated, although there have been in recent days some code of conduct related problems in different projects. I don't know. It's interesting, it's really interesting to see.
38:21 - Jeff Massie (Host)
I think there was probably some in-person know development channels and maybe irc or discord for open susa that you know most of the public isn't in. You probably would have a good idea what's going on. But yeah, I I searched around some. You know I didn't do hours and hours but I bet I spent half an hour kind of looking like what's going on. Why are so many people saying there's problems but nobody's really coming out and saying much? I I couldn't really get anything other than burnout, but there's much more to the story. I mean.
38:54 - Jonathan Bennett (Host)
Yeah, although burnout is a thing.
38:57 - Jeff Massie (Host)
Oh it is, that's a real issue.
39:00 - Jonathan Bennett (Host)
And I think we all feel that from time to time especially those of us that are working open source projects feel that from time to time especially those of us that are working open source projects.
39:06 - Jeff Massie (Host)
But when they talk about changing structure and changing, you know there's there's something else. It seems like going out rather than just simple burnout.
39:14 - Jonathan Bennett (Host)
I mean not saying that that's not a problem, just that that's one of the problems they're experiencing right, and some of the other ones are behind the scenes and open suza has always sort of had this, um, this kind of odd structure where, like, so they're, they're associated with upstream susa, that like the company susa, um, and they they kind of repackage susa, so it's, it's. You know, it's similar to how red hat and sent os interacted there for a while, um, and so maybe I don't know maybe part of it is SUSE is thinking some of the same things that Red Hat thought, that you know, we have this free and so, in this case, free as in beer version of our enterprise offering and too many people are using that, and so, rather than killing it like what Red Hat did with CentOS, maybe they're just thinking if we can do a name change, then it will make people a little less likely, like there'll be less confusion between the difference between SUSE and OpenSUSE, and then, of course, trying to deal with that. So the impression that I got was this new reality that OpenSUSE is looking at. In trying to deal with that, the board, the people that are in charge of it, are kind of coming to the conclusion that what they're doing is not working to be able to come to terms with this and come to the solution. So we'll see what happens here.
40:44
I saw also some suggestions that maybe it's time to take some of these different distros and try to merge them. You know, do the opposite of a fork. Uh, that's always challenging to do, but every once in a while it happens. So we, we will. We will keep our eye on what happens with SUSE and OpenSUSE.
41:02 - Jeff Massie (Host)
Well, it's interesting. You say, you know, opposite of fork merge, fork merge. Linus Torvalds, now it's been probably over a decade now, or at a decade where he had his speech talking about why Linux fails on the desktop and it has to do with so many distributions using different versions of various programs and libraries. You can't compile one binary and have it run everywhere. Now you can argue, you know, snap and flat pack have kind of taken care of that.
41:33
An app image, you know not everybody's yeah, not now everybody, not everybody's on it, but it definitely kind of leveled the playing field because now you don't have to have oh, I, I need the Debian one, I need the Ubuntu one, I need the Fedora one, I need the OpenSUSE, I need the Arch, I need the. You know, you say, okay, I package this up now in Flatpak. But that was still one of the things where he said you know the average person, I just need to click. If it doesn't go, they kind of start losing interest. And maybe some consolidation and a little better alignment in the Linux universe would be a good thing. Yeah, or standardization.
42:18 - Jonathan Bennett (Host)
Yeah, so to speak to that, there have been standardization efforts, and I think we are better now in the Linux desktop than we ever have been. Oh, yeah, not only because of Snap and Flatpak, although those definitely have helped, but there's even been things like and I'm not going to be able to come up with the titles of them at the moment, but there are standardization efforts about where do you put your binaries, and people have gotten on the same page on that. Um, even wayland itself is. And one of those efforts, right, um, almost everybody is using systemd these days, and so that's pretty much the same. Um, it's, it's a lot better than it used to be.
43:03
And then, of course, you've got uh, you've, you've got proton, which brings all of those windows applications you know most of those windows applications over and make some work on linux. So we're, we're miles ahead of where we used to be when, when torvalds gave that speech. Um, yeah, but as far as, as far as open susa I we just going to have to wait and see what they do, what decisions they make, see what else comes out Like. I imagine more people will, in the days to come, share more stories about what they think has gone wrong, yeah, so hopefully the real reasons are not all locked under NDA Right.
43:45 - Jeff Massie (Host)
Well, and maybe a new branding is going to breathe new life into things. But when people talk about Red Hat and Fedora, I mean Fedora's been tied to Red Hat for a long time and they've kind of got the momentum, so I think it's going to come down to okay. Whatever they call this new, whatever OpenSUSE rebrands itself to, I think they can have the success. But they've got to make it those first couple years just to, when people don't know the rebrand or they are kind of still figuring things out. And if they can do that and survive and SUSE keeps supporting them like they are, I think they'll be okay. It's just the next couple years, I think, are going to be the critical years.
44:33 - Jonathan Bennett (Host)
Yeah, and there is one thing to point out about the difference between Fedora and OpenSUSE Fedora is upstream of Red Hat and OpenSUSE is downstream of SUSE. Now, in both cases, you have people at the corporations are writing some of the checks, like. A lot of people that work at Red Hat are also Fedora contributors, although, from what I understand, a lot of those do that on their own time. But I believe it's the same case with SUSE and OpenSUSE that there are people that are part of OpenSUSE. Excuse me, there are people that are employees of SUSE that are contributors to open susa, um, but open susa is downstream and that's a very different relationship than you know something like fedora has with red hat um, that's a good point and maybe that's a change they should make and say okay, you know what'll be your beta tester.
45:24
Isn't that kind of what Tumbleweed is?
45:26 - Jeff Massie (Host)
Tumbleweed is yeah, because it's the rolling, cutting-edge version of OpenSUSE, right, right.
45:35 - Jonathan Bennett (Host)
And so I imagine that if they don't do this now, they probably eventually will. They'll start snapshotting Tumbleweed as their next version of SUSE, as you know, their next version of susa. Um, so I you know well, it's a developing story and we will see what happens.
45:55 - Jeff Massie (Host)
Yeah, rob is traveling this week and he was. He was disappointed that he wasn't going to get to comment on this because I know he had some thoughts, so there's a decent chance that we'll be talking about this for weeks to come, so Rob will get his shot.
46:06 - Jonathan Bennett (Host)
Yes, all right. Well, let's move on to something that is a little bit more positive, and that is that things are happening at KDE, and so somebody in our chat room before we got started was talking about how annoyed he was that KDE on Wayland kept crashing. Well, there's a blog post for that. So Nate Graham is out with his you know his latest adventures in Linux and KDE the pointed stick blog and so he talks about two main things. One, they are doing a lot of accessibility work in Wayland, and that's things like sticky keys, and you might think well, what? First off, you may wonder what in the world are sticky keys actually for? Well, the idea with sticky keys is you know a lot of us that are listening we have 10 working fingers and we can easily hold down shift and control and hit another button. But there are some people that, for various reasons, do not have that ability to easily hit multiple keys at the same time, and so a sticky key is so that you can hit shift and it holds itself down and then you can go and hit your other buttons, and someone that has a challenge in how they are able to use the keyboard for again, for whatever reason. It gives them the tools they need to be able to continue to use the computer. So in KDE Under Wayland they are working feverishly towards making all those accessibility features work again and they are bringing a lot, very much of it into line to where they have the same features they used to have. There's a bunch of other things that they talk about with improvements, little improvements, some of the bigger improvements Apparently, filelite has gotten a lot faster and prettier to look at. That is the application in KDE to let you see essentially what files and folders are using up all of your disk space. But when you get to bug fixes they talk about I think it was five common crashes that was causing people problems and because they have the automatic crash reporting system, people can, when they have a plasma shell crash, there's a panel that pops up and says would you like to report this? And it's super easy, like you don't have to have an account or anything, you just hit the button and it sends the report. As Nate says, because of this, because they were able to get those automatic crashes, they were able to find the five most common crashes and get all of those fixed. And I believe that's going to land in some of it's going to be plasma 613. And so it's going to look what 6.2 is available and hopefully all of that stuff will get fixed for you and it'll once again be usable.
49:06
So some of the things that I like to watch for in these no further news on nothing more has been done on HDR to watch for in these no further news on nothing more has been done on hdr. They've not convinced uh wayland to finally, you know, make the hdr specification. Some days I wonder if it would be. Maybe kde just needs to officially fork uh wayland, the way wayland works, everybody sort of has their fork of it because the desktop environment is actually implemented. But, man, I wonder some days, like maybe KDE and the guy from Hyperland and one or two other stakeholders could just like fork the Wayland specification and go through and fix all of these things that have been years in the making. We'll see Make Vaxxery, make Vaxxery, make Vaxxery. The BDFL for Wayland. Oh, that would be fun, A little terrifying, but just the chaos. Maybe I'm just the guy that wants to watch the world burn today. Grab them.
50:11 - Jeff Massie (Host)
You must have got a good deal on a bag of marshmallows or something, I guess I guess. Are you running KDE Wayland these days? Grab them. You must have got a good deal on a bag of marshmallows or something, I guess I guess uh, are you?
50:16 - Jonathan Bennett (Host)
are you running kde wayland these days? I don't remember jeff I am not I, so I your main system is a bun to right my main system is kubuntu.
50:27 - Jeff Massie (Host)
Okay, so, so kde. So I'm on 24.04. It is still on Plasma 5. So I'm still running KDE 5. I loaded up KDE or sorry, I loaded up Kubuntu 24.10 a few days ago and checked it. You know the daily build and it's still on Plasma 5. It's still on Plasma 5. Well, I know a person that does some development you know packaging work for KDE or Kubuntu personally and he said it's probably about two weeks before that's going to hit, because they pull from Debian and Debian's working. Right now, to put it in the hard part, they can't just add it because, while, okay, debian does a lot of the work, they go to put it in there. There's some conflicts a lot of times with Ubuntu, because Ubuntu doesn't care about KDE, so some of their programs are old and they don't care because it's maybe something that doesn't matter for GNOME and so they don't update it. But to make the new kde6 work, they've got to update that.
51:37
So there's some background packaging, uh, shenanigans that have to go on yeah, it makes sense yeah so it's a little more complicated than versus something like neon, where, oh, we only care about kde. Everything's aligned and it's you you're fighting.
51:57 - Jonathan Bennett (Host)
You know, mama, canonical a little bit, and yeah, and then on the other, on the other hand, yet again you have somebody like fedora, where it's like none of the packages are old and fedora, almost everything is bleeding edge.
52:09
Uh, so you don't have that problem there as well it's, it's just, it's interesting that ubuntu is still on five, but at the same time that makes sense, because ubuntu is not about being bleeding edge. Ubuntu is about being stable, but debbie and even more so, um. So you know, one one can make the argument that maybe fedora jumped to kde too soon. I would not necessarily agree with that. I think those of us that run Fedora know what we're getting, but there's definitely something to be said for Ubuntu taking a bit slower approach to it.
52:41 - Jeff Massie (Host)
And I will say Fedora was great. You know, I loaded it on a drive and was running it for a while and it was awesome. I didn't jump off at Fedora 40 like I should have and I was still on Rawhide. Well, I booted back in and it had like 3,500 updates or something and something went a little wonky in there. So that's where I was like, oh okay, I'm going to put 2410 in and see if I can run that, and so I'm gonna have to wait a couple weeks and then, once they get kde6, whatever packaged, I'll jump in and run that and see how that works, give it some test driving, and so the?
53:25 - Jonathan Bennett (Host)
uh, I'm looking to see if I can quickly find the fedora 41 schedule. So, um, we branched from rawhide on um august 13th. That's actually coming coming up very soon, so you know, sometime after that. Uh, the beta freeze is august 27th. Get you a beta release right after that. And uh, yeah, jump, jump on fedora again. Maybe we can get you to uh to change your main system over I.
53:59 - Jeff Massie (Host)
I will tell you, running running fedora actually made me consider jumping over I. I seriously thought about it. One of the one of the things was just okay, I have to learn a new package manager, but but it was pretty close. It was pretty close to app. So I'm like this isn't bad. This isn't like some totally new interface. It's kind of like you're talking to somebody from australia or britain where they're speaking english, but you kind of gotta learn the slang a little. It's kind of like what it was, because it's like I know what's going on here. It's just slight differences in certain places. You know it's going to have a Debian version, which Debian? Ubuntu, kubuntu, a lot of the flavors they all are Debian, so you're pretty safe on that. But the other juggernaut is Red Hat. So you're going to have that. A lot of times you got oh, we got a deb package, we've got a rpm or whatever for Red Hat, and then you got TAR where you're going to have to compile or you might have to run their install script.
55:28
Or you have to go to your distributions, their repositories, to specifically pull it in. Yeah, but I liked Fedora when I used it and it was surprisingly close to what I see in Kubuntu. So they're both pretty similar in their adaptation of KDE. I liked Fedora.
55:51 - Jonathan Bennett (Host)
Yeah, there's some neat things coming in Fedora 41 as well. It goes to DNF5. Dnf5 is nice, Much quicker. I've got it kicking around on a single-board computer and been very impressed with that. I will make sure and bug you once the Fedora 41 beta is out and we'll get Jeff to try it again. Maybe we'll make a convert out of it.
56:13 - Jeff Massie (Host)
Maybe and this will be a good testament of just running 2410 Kubuntu versus Fedora 41. What do I like better? Any hiccups, any issues, that kind of thing? Yep.
56:33 - Jonathan Bennett (Host)
All right, so you've got a story here on blender. Are you ready to take it away and tell us what's new in the newest blender?
56:41 - Jeff Massie (Host)
I am all right, I am so. So the blender foundation just released blender 42 or 4.2 lts, so it's a long-term support version and it will get updates and support until July of 2026. So that's pretty good. They're adding support for GPU-accelerated composting for final renderers. The acceleration will be enabled by default for AMD GPUs and it has a rewritten render composter CPU back in which will give a performance boost. There's even a faster undo feature which will make your revisions even faster, so you can back out of things quickly. Now, if you look at the article in the show notes, they go into all the details of how that happens, but suffice to say undo is much.
57:32
Now the eevee render engine has been completely rewritten from scratch. Now. This allows for global illumination, displacement, viewport, motion blur and a ton of other features, and again I'm I'm skipping a bunch of the more inside baseball type or deep, deep features, because I don't know fully understand what all of them are, and no sense hearing me rattle off a long, long list. Um, intel gpu rendering with the cycles engine now supports memory fallback and you know there's a. There's a lot of other stuff that's gone on there now, while there's a ton of things under the hood. There's a lot of other stuff that's gone on there.
58:15
Now, while there's a ton of things under the hood, there's also user interface improvements being updated. There's a new copy driver to selected. Right-click menu option for copying the proprietary drivers to the same property on all selected items. Support for horizontal rules and vertical layouts and vertical rules and horizontal layouts, font shader optimizations, improved square color picker on wide screens, improved status bar, key map display and improved small and large confirmation dialogues. This release also unifies add-ons and themes as extensions which you can download from extensionsblenderorg website, so it makes it even easier than ever to integrate in studio pipelines and things like that. It also adds a new socket type for matrices to make it easier than ever to work with transformations using nodes and makes node tools more interactive.
59:14
So take a look at the link in the show notes. Look at all the improvements in there. There's a lot of deep, deep stuff. You know professional level. If you're a superpower blender user, you're going to want to go through the entire list of what's changed in here. The entire list of what's changed in here. But a lot of gpu and feature improvements that should make life a lot better for people people using this all right, you know I have.
59:43 - Jonathan Bennett (Host)
I have spent very, very little time in blender. Um, I've I've tried it once or twice um to do a uh, a bit of 3D print model manipulation, and I had the hardest time trying to wrap my head around it. But I know there are people that are wizard at it. I guess I just need to dive into it with more time to spare. Watch more YouTube tutorials, because there's a lot. If you want to learn Blender, there are oodles of good YouTube tutorials out there on how to do it. I just I've not been able to take the amount of time that I would like to.
01:00:21 - Jeff Massie (Host)
I played with it some Now, both just using simple shapes and trying different surfaces and lighting and things like that, or taking existing models. Like you can take the BMW render, which is kind of a standard default image that you can or a model that you can play with, and then, oh, I'm going to change it this way and stretch and pull and morph and, you know, do a lot of a lot of fun stuff and but I barely dip my toe into it. I am, you know, know, I'm not even novice level, I'm below that, but but it was kind of fun, you know, just playing and seeing what, what the changes make and what you can do, and you know yep, yep, fun stuff, all right, so we've got.
01:01:10 - Jonathan Bennett (Host)
So we've got one more story that we want to talk about, and it is not exactly a Linux story, although there is a Linux angle here, and I'm trying to find a particular tweet that I saw earlier, and now, of course, I can't find it because I can't remember enough of it. I don't remember the name of the company that made the tweet, so it'll have to go unfound.
01:01:36 - Jeff Massie (Host)
Is the Linux angle? Just laughing at the company.
01:01:39 - Jonathan Bennett (Host)
Well, no, actually, actually, no, it is not, and so that's the tweet I was looking for. That's what it is, and now, of course, I can't remember. I don't remember the name of the company. It was like Helium 8 or something, but of course that's not exactly right. I just I'm not going to be able to remember it, unless, of course, it was Helium 8. We'll see.
01:02:01
Anyway, so you may have noticed this is Saturday, this is Saturday, july 20th. You may have noticed, on Friday, july 19th, that, uh, the world sort of lost its mind and about a million and a half, if not more than that, uh, blue screen of deaths happened on windows machines everywhere, and, uh, I saw stories of airplanes that would not get off the ground because they had windows machines built into them. You had lots and lots of digital signage that were showing blue screens of death. Uh, it was, um, it was sort of, uh, it was, it was something. It I saw, one person, in fact. So troy. Troy hunt is who I first saw say something about this, and I saw it late, late, late thursday night, because I was up late trying to finish my security article for Hackaday, and right before I thought I had it finished, I saw this tweet from Troy Hunt saying you know, I'm getting reports that Windows are BSODing all over the place. I'm like you got to stay up a little later and figure out what's going on and it very quickly became obvious that this was a huge problem. All right, so CloudStrike, one of the big, one of the big it's not fair to call them just an antivirus company, because they do a lot more than that. So the cybersecurity vendor put it that way it was they pushed out a rule update, essentially, and when they did that, computers around the globe pulled that update down and because it was a broken update, it caused those computers to go to the blue screen of death and then start rebooting. It's funny. So earlier today, before the show, I went over to my parents' house to pick something up and my dad's like I assume you heard about all of the computers he points at his laptop, his work laptop, across the room. He's like I can't get into that one yet. I'm waiting for the IT guys at the big business where he works at to come and actually work on it, because they're going to have to do it one at a time. I'm like, yeah, you might be able to reboot it. No joke, you may be able to reboot 15 times, at least 15 times, and it may eventually fix itself, which is hilarious. Um, okay, so what happened? Well, they've got.
01:04:29
They've got something in CrowdStrike called Falcon and it is their antivirus piece. It works as a Windows kernel driver, so it's a file. It's a driver that runs in kernel space, so it has kernel permissions. It's part of the Windows NT kernel. When they pushed this update, they call it a channel file and the configuration files. So I have in the show notes the CrowdStrikecom blog post about this, and so it says the configuration files mentioned above are referred to as channel files and are part of the behavioral protection mechanisms used by the Falcon sensor, of the behavioral protection mechanisms used by the Falcon sensor. And what they did is they pushed out one of these channel files, which again, it's an update to their definitions, what they're looking for.
01:05:22
It had an error in it. I've seen one report that said that the bad channel file was nothing but nulls. Crowdstrike says that that's not the case. I don't know who to believe. Nulls. Crowdstrike says that that's not the case. I don't know who to believe. But anyway, when they tried to load this file, it did a bad pointer dereference. I don't think it was actually a null pointer dereference, but it was just they tried to access memory at a place that didn't make any sense. But because it was in the kernel, the machine just went kaboomy and everything went down and all across the world the airports were shut down, the airplanes stopped running.
01:06:05
I think it was also Troy Hunt that said, basically, what we had yesterday was what we were warned was going to happen with the Y2K bug, except this time it actually happened. We didn't have, you know, a year and a half warning to be able to fix all the things, so it was pretty bad. And then, of course, all of us running Linux machines, we were totally unaffected, and I will mention that, from what I understand, this problem only happened if you were running CrowdStrike, if you were running the CrowdStrike Falcon antivirus solution, which I was not. None of my customers are either. So you know, yesterday was a pretty easy day for me. I did not have to answer those phone calls. I lucked out this time, anyway. So those of us running Linux, a lot of us kind of you know back on our haunches. And we, of course, we did not fall victim to that Except except there is a CrowdStrike for Linux, and apparently CrowdStrike for Linux also uses a kernel module, and that kernel module is going to have exactly the same weakness that if they do something really boneheaded they could cripple a whole bunch of Linux machines just like they did Windows machines.
01:07:23
So I have seen reports that the CrowdStrike for Linux application has two different modes that it can run in. It can run using the eBPF, which that is essentially user land scripting in the kernel, and that is going to be a lot safer. Or it can run using an actual kernel module. And so people are saying if you're running CrowdStrike, go in and make sure it's in eBPF mode, because otherwise you'll have this same problem potentially at some point in the future.
01:07:53
And then the tweet that I was trying to find was yet another antivirus vendor that was saying we knew this could be a problem, which is why when we wrote our Linux antivirus, we did everything with system calls and we didn't write a kernel module at all. And now unfortunately, I can't find who it was that actually did that a kernel module at all. And now, unfortunately I can't find who it was that actually did that. I was searching frantically while Jeff was talking and I've not been able to. So that is what happened on Friday, july 19th, the day that all the computers went blue, or as I called it, the breaking BSOD event, the blue screen of death. In this case. Not burglary style development, um, but interesting times. Did this affect? Did this affect you, jeff? Were you stranded anywhere?
01:08:44 - Jeff Massie (Host)
no, I didn't even really realize it was happening until, um, somebody else had said oh I, my flight is, I can't get out because I'm stuck and because of this and I'm like what, what's going on? You know, I'm kind of surprised, though it affected so many machines, because a lot of big companies or at least, in my experience, the ones I know they will take these updates and they'll load them on a small group of machines and not push it out to the enterprise to catch things like this.
01:09:19 - Jonathan Bennett (Host)
I'm glad you said that. So there was another layer of fail here in CrowdStrike. This is a huge black eye for CrowdStrike, so I saw multiple reports that companies that ran CrowdStrike that had their automatic updates turned off for exactly that reason still pulled the update down and crashed. Crowdstrike was pushing updates even though they were turned off, oh, which I think is why it hit so many people so badly.
01:09:51 - Jeff Massie (Host)
Well, I wasn't Well. Okay, the way I understood it so like is.
01:09:57 - Jonathan Bennett (Host)
If your computer is connected to the internet and is running crowd strike, it does not matter what you have the settings set to. The crowd strike application will pull these down, and as soon as that happened it was dead yes, yeah, I wasn't even sure you could.
01:10:13 - Jeff Massie (Host)
Let's just say, some companies kind of isolate the machines so that you pull updates from internal to the company, because they will download them, set them up. Okay, does this work? Does this not work? Okay, we approve it. Okay, we'll supply the update then and push it out ourselves, versus letting the internet do it yeah.
01:10:40 - Jonathan Bennett (Host)
So if you, if you, if you have a computer that is completely isolated and does not have an update server that it can go through, then, yeah, you should have been safe. But because of the way this update was pushed out by CrowdStrike, it was an automatic apply do it right now. It doesn't matter what your settings are, and I don't know if CrowdStrike has the capability to do a proxied update server like that, but even if it does, from what I understand, it would have pushed to the server and then pushed out to the endpoints, regardless of what your settings were. Okay, and so it. Yeah, it hit. It hit everything all at once and I I felt so sorry for, like troy hunt, he's in, he's in australia, which is why people were aware of this as soon as it happened.
01:11:26
It was like 4 pm in australia when, when all these computers started going down and uh, so that's that's where I saw it. I was some of the first people that had the problem. You had tweets in Japanese, because the people in Japan were awake at the time and their computers went down. And then all of us got up groggily on Friday morning at 2.30 or 3.30 or whatever time in the morning and people's phones started ringing hey, your computers are down. And it was crazy. I saw at least one Reddit thread where the poor lone IT guy at one company was like I have 4,000 machines to go fix. I really want to quit, but my boss says he'll sue me if I do.
01:12:11 - Jeff Massie (Host)
It was not a good day for some people.
01:12:12 - Jonathan Bennett (Host)
It was a real bad day for some people.
01:12:14 - Jeff Massie (Host)
Yeah, and I heard that it was all manual, it just yes.
01:12:21 - Jonathan Bennett (Host)
So the first solution that came out was that you could reboot into safe mode and go rename the CrowdStrike directory and that would fix it. And then you know, things got more precise. Ok, well, here's the actual file that you need to delete or rename, and that was the bad update that got pushed. But I also saw on the Microsoft blog post on this that apparently if you were to just keep rebooting the machine it would eventually fix itself. And I'm guessing what would happen. There is that, like, the CrowdStrike application would run just a little bit and it would try to pull down an update before it pushed its instructions into the kernel, more or less, and so if you just kept doing it, it would eventually pull down the new, the new fixed version, um, but yeah, like rebooting at least 15 times may solve your issue at least yeah, so that's not the upper bound, that's.
01:13:22 - Jeff Massie (Host)
That's the lower bound. Yeah, that's the start of the fix. Yeah, yeah, it was not good. I mean really. I mean really this could be. We always joke about it, you know, testing in production. Well, oops, kind of solving the meme.
01:13:43 - Jonathan Bennett (Host)
Capsule 8. Capsule 8 is the company that I saw. I'll see if I can find the exact tweet. But yeah, capsule 8 is the company, I believe, that has the uh, the crowd strike, or it has an antivirus and they very intentionally did not take the kernel module approach. Um, yeah, this I will, I'll find the exact tweet. And um, put it into the show notes. But, yeah, fascinating, yeah, but luckily for me, my day was. I didn't know anything happened. Put it into the show notes.
01:14:14 - Jeff Massie (Host)
But, yeah, fascinating. Yeah, luckily for me, my day was. I didn't know anything happened until somebody mentioned something. And oh, let me go look at the news. Holy cow, for me it was just a normal day.
01:14:29 - Jonathan Bennett (Host)
I luckily was one of the unaffected. Yup, yup, yup. All right, uh, let's, uh. Unless we have anything more to say about that, let's move on to some command line tips. And uh, jeff, you've got the first one. What is what is shuff? It's, it's, it's almost, it's almost like. Sure enough, said very quickly.
01:14:58 - Jeff Massie (Host)
Or a enough said very quickly, or a lazy way of saying shuffle, oh no, that could be yeah, which. Basically it's a simple command that stands for shuffle, so it will generate random permutations and the simplest use is shuff space file name and it will give you the lines back of the file, like of a text file, in random order, repeating the command again. Just, you know, if you hit up arrow and return, it will change the output every time. Interestingly, if you use the switch dash dash, random dash, source equals and then a file name, and then you have to do a space in the file name again, it will give you the same output. It'll be randomized but it'll be the same one every time. So if you need repeatability, that will do it.
01:15:45
There's options to allow lines to be repeated, which, for those math people out there, you know you, you're, you're not, you're not in a permutation anymore. And if you can specify ranges and numbers of lines, you know you, you can do that. So you can say, well, I only get, only give me six lines, or give me, you know, if there's numbers, you can say, well, I only want one to 10 and maybe there's a hundred in the file, so you can just pick what you want. Um, and it's a pretty simple little command, but you know, if you want a little more randomness in your life, this might be the command line tip for you.
01:16:26 - Jonathan Bennett (Host)
All right, interesting stuff. Um, I I stumbled across, uh, yet another command line tip that has apparently been around for a very long time that I didn't know anything about until now. And it is, it's the well, it's the Mr Command, but actually that stands for my Repos. It's available as a download. You know, it's in your package manager on almost every system. It's been around for a long time.
01:16:50
But it lets you do some neat things like automatically pull updates to get repos.
01:16:57
You can tell it, you know, a couple of times a day, go out and pull whatever's new for all of these different repos. Or you can say you know, I want you to update all of my repos, or I want you to push all of my repos, or make a commit and push all my repos. Now, obviously, some of that, it could be dangerous if you are not doing what you think you're doing, or you know if you don't have a good reason to do so. But, depending on there could be times that you want to do that, you want to automate those things, and so a very useful little tool potentially if you have a bunch of repos, if you're doing interesting things with Git repos, eventually, if you have a bunch of repos if you're doing interesting things with Git repos and that is the MR or the my repos command. And I've got a link in the show notes to my reposbranchablecom, which is their website with some documentation on what it does. So there, you go.
01:17:50 - Jeff Massie (Host)
That is interesting, it is, it's fun. I have not heard that one before, but that's pretty cool.
01:17:55 - Jonathan Bennett (Host)
It is. I'm actually pretty fascinated by some of what they do. I'm going to have to see if I can't get it set up to be able to do that automatic pulling of repos, because that would be nice.
01:18:06 - Jeff Massie (Host)
Yeah, that is pretty cool.
01:18:10 - Jonathan Bennett (Host)
All right, do you have anything?
01:18:12 - Jeff Massie (Host)
you want to plug jeff before we let folks go yeah, so I I've got not only my weekly poem but I've got. You can find me this week on floss weekly, episode 792, where we talk about rust core utils. So that was uh excellent. Uh, it was really fun. I love doing Floss Weekly. Can't do it all the time because of scheduling, but it was a lot of fun. And inside baseball here, when Jonathan said hey, can you do this? I first went I don't think I'm your guy for Rust, no, no, no, it's about the utilities and the core utils.
01:18:51 - Jonathan Bennett (Host)
You know core utils? Oh, okay.
01:18:53 - Jeff Massie (Host)
I know core utils, yeah, yeah. So I'm like, yes, I will do it, I'll be there. So you can find me there. The episode's out 792. And the last thing is my poem. No Keyboard Present. Hit F1 to continue Zen engineering. Have a great week, everybody.
01:19:16 - Jonathan Bennett (Host)
Bravo. All right, thank you, sir, for being here. I sure appreciate it. And then so, as far as stuff that I want to plug, of course we've got my work on Hackaday. We have Floss Weekly there we talked about, and then the security column goes live every Friday morning. Be sure to check that out. Sometimes we do breaking news there, like the CrowdStrike story Got a bit of a blurb, this past story, this past column, and then I'm sure next week we will talk about it in more detail because, boy, it's the big story for the month.
01:19:44
Anyway, I do want to make sure and mention that you should be on Club Twit. You want to help support the network, support the show, support the show. Scan the QR code, jump onto Club Twit and be a part of the club. It's about the cost of a cup of coffee per month. It gets you into the Discord, it gets you the ad-free versions of the shows. It is definitely worth it. All right and hey, that has been the show and we have sure enjoyed it. All right and hey, that has been the show and we have sure enjoyed it. Thank you to everyone that watched live and those that get us on the download, and we will see you next week on the Untitled Linux Show.