However they also claim the mitigation "reduces exposure to this vulnerability but cannot fully prevent it", without elaborating on what that means or how it would be possible... I've also not seen anyone else make this claim.
That's what I call Steve Jobs style PR / Marketing you deflect the problem by making another problem to make it seem like you're fully aware. "You're holding the phone wrong"
I've seen another poster successfully updating their Kernel from Ubuntu's "proposed" branch which already has the fix. God knows why they don't merge it into stable.
They have over a hundred kernel configurations (multiplied by aarch64 vs amd64), many of which going back years, to do testing on. That probably doesn't help. People comparing them to open source projects like Arch and nixOS which support maybe one or two versions per platform are making a false comparison.
So far they're two days later than IBM/Red Hat. Not great, for sure, but they seem to be that far behind compared to the competition just yet.
If they push a backport that kills a million set top boxes because of changes in the network stack, they're not much better off than if they work a couple more days on testing.
the boundary does elaborate, its a network namespaace scopep statement.
and the rule only filter traffic to the host netns, so, pod to pod traffic is untoched, and forwarded to pods filtering depends on the cluster network config.
FreeBSD is quietly prevalent in commercial and government settings. Juniper and NetApp are just two renowned examples. There have been and are many others.
Separate from any of those, a good chunk of America's more sensitive production database clusters, other workloads, and some network applications are scheduled by Nomad or Kubernetes running directly on a kernel that is not Linux.
The most popular kernel is not necessarily appropriate for all use cases or types of environments where high-value targets reside.
Directly downstream of FreeBSD, HardenedBSD achieved its long history of invulnerability to what were exploitable vulnerabilities in upstream through its advancements in the operating system layer.
Okay... can you name a specific thing? I'm confused. Nomad and Kubernetes both support Windows workers well. Kubernetes, afaik, doesn't support something like OpenBSD in a common flavor like RKE.
Kubernetes' core pieces of userspace software, not with any implied support from the Kubernetes project, also run on illumos, FreeBSD (and HardenedBSD), and OpenBSD.
In some advanced settings where tolerances for potential impacts resulting from exploited software vulnerabilities are intensely low, Kubernetes (userspace software) on HardenedBSD is choice for mission.
Like with the other two systems, those running Kubernetes userspace on them are self-sufficient and are not among the audience of external more general Kubernetes distro support channels.
If you happen to like RKE, check out https://Talos.dev which is the best Kubernetes on Linux distro (for most Kubernetes on Linux use cases).
_
@nijave
Nomad and Kubernetes schedule containers, which are rarely not Linux containers, yeah.
Both can schedule processes in FreeBSD's BSD Jails.
If someone wants to start an open Kubernetes on FreeBSD/HardenedBSD project I would encourage that. Personally I would like to see an open Kubernetes on illumos project.
As far as I know where Kubernetes on HardenedBSD runs is not interested in open projects.
For those not aware: Freenode.net is not Freenode as we all knew and loved it from the early 2000s to 2021. A hostile take-over of the Freenode LLC resulted in the loss of the domain name and some corporate assets. And the hostile takeover person, Andrew Lee, deleted the Freenode database. It no longer exists.
"Freenode" in 2026 is an entity actively hostile to open source software and while what they say here might be true, they are only saying it because it aligns with their biases. They are not a real news source.
My mitigation for this lack of patches has been using iptables.
iptables -A INPUT -p tcp -m u32 --u32 "0>>24&0xF=6:15" -j DROP
Yeah and within 3 weeks it was back to the same old, like freenode had always been libera <3 And no more bullshit and hassle which Lee brought.
I was afraid the community would get splintered out but it didn't. The quick decisive action in establishing libera really helped that. Only a few channels left to go somewhere else (like alpine linux)
I really loved this example of a community massively saying Nope and just avoiding a hostile takeover altogether. Unfortunately it doesn't usually happen so well.
Unfortinatly no. A lot of channels moved to OFTC or other networks instead (and some quit IRC entirely) so Libera isn't the one stop for open source IRC that Freenode was.
Yeah, I thought we stopped posting freenode links here on HN a while ago. Kinda surprised about this one (especially since it is covered in detail elsewhere as well)
Freenode is the worst possible source for discussing open source issues considering the harm its new owners caused to open source communities.
But lately, the Freenode fiasco has been on the back of my mind from time to time. If the same thing happened today, a good chunk of the comments would be cheering for Freenode. Almost every single article critical of the takeover would be flagged too.
> And the hostile takeover person, Andrew Lee, deleted the Freenode database.
But only after it was clear that most channels were leaving anyway. In a way, that was the only good thing he did in this mess: reduce fragmentation that we would have gotten as some stuck to Freenode.
It's rather uninteresting to note, once again, that Ubuntu has decided to do this. I would like to know the (official, if any - unofficial/speculative otherwise) reason WHY they have not.
It does not seem that complicated to patch, and kernel upgrades are a dime a dozen these days. So what is making them delay a fix for so long?
How do we feel now about having pointless services listening for no good reason on open ports?
And do we feel about running no firewall?
For example Linux works totally fine without running dhcpcd and without running avahi. But on many systems where people think they run "no service", these two shall have ports open on your network and happily listening.
Sure, it's typically "only" on your LAN: but how long before we see exploits affecting even private LANs?
> about having pointless services listening for no good reason
I think you'll find that for most people, plugging the laptop in and having the network immediately available and automatically seeing their local printer is a pretty good reason.
> A zero-day (also known as a 0-day) is a vulnerability or security hole in a computer system unknown to its developers or anyone capable of mitigating it.
Apparently the meaning has evolved over the years.
I have not seen an exploit in the wild. I have at least 20 different servers unpatched (obviously) but the risk is blown way out of proportion. There are a number of mitigation options available.
I know something. Every fucking server with public facing infrastructure is not getting hacked left and right at the moment.
So you make of that what you will, but in my experience it means the exploitation comes hand in hand with some far more esoteric chain that is not easy to put in practice.
What was the timeline for the release of WannaCry following the exposure of ETERNALBLUE and DOUBLEPULSAR...?
I accept that one would expect substantially shorter timelines in the age of AI, but given that WannaCry is suspected of being state-funded and even then: 1) it took a month to appear and 2) ETERNALBLUE and DOUBLEPULSAR were essentially already finished, weaponized malware off the shelf.
All of this to say that just because you don't see mass exploitation right this second doesn't mean there isn't a wormable exploit that needs patching. WannaCry didn't exist until it did, and suddenly everyone who failed to patch beforehand looked foolish.
How would you see one? Have you deployed your own army of honeypots that are actively logging and looking for corrupt IP options packets? If you're not already looking for it, I think it's very likely that most people who get hacked will have no way of knowing how it happened, so we can't consider those incidents as being a result of this particular bug.
Are you actually monitoring for this specifically or just trying to use a sample size of 1 group of servers to prove something?
Jesus H Christ! the bug is a narrow use-after-free race that needs SYNs carrying ipv4 options to make it to the box. I do not need an army of honeypots to tell you it most likely doesnt to have a way into a classic 443 machine... come on now.
> Multipath TCP is not a required condition for this flaw.
I contemplated not even responding to you because of how unbelievably abusive and rude I have watched you be on IRC to me and many others for literally decades, and was a big reason why I quit, but I decided this issue was more important for everyone else's sake that I should still speak up.
The recent wave of vulnerabilities has actually prompted a lot of people to set up honeypots, monitoring, etc. and the rate of actual exploits has been notably very low for the general public.
Most of the October CVEs we've seen were situations that require an extremely complex chain of attacks - amazing if you want to break into a bank, or fuck up some grocery freezers, not so useful for scanning random public boxes and doing "hacks of opportunity"
Apologies for the lack of links - it's a scattering of random people I've read talking about this, not any coordinated study. But if you read between the lines on all of the threads about these CVEs, you'll notice a lack of "I was hit by this" stories - which is pretty normal for more complex vulnerabilities.
The reason these sat unnoticed for so long is precisely because it's difficult to actually exploit them.
If anyone has anecdotes to the contrary, I'm definitely interested in hearing it, of course!
> I have not seen an exploit in the wild...but the risk is blown way out of proportion
That's a very naive view on it (now) and I'll try my best to explain why.
The bad actors can discover multiple 0-days on the fly to gain access to an employer/company infrastructure with valuable data/information, or in the case of Binance, crypto. They could chain multiple such attacks going from one-company/provider to another, to another, to eventually reach their target.
To anyone interested, this is good reading material on the topic:
(1) study the Binance hack in detail
(2) OpenAI huggingface hack - focus on 0-day chaining
You have not knowingly seen an exploit in the wild. Competent attackers can be right inside your OS, and many other devices containg within the same network as a result, without you knowing it; persistenly so if they want to.
> The following nftables table drops incoming IPv4 TCP packets that contain IP options, which reduces exposure to this vulnerability but cannot fully prevent it.
I suppose locally-running malicious programs could send packets to the loopback address or one of its own IPs or something.
It could be from a multi-part attack where someone combines an application-level RCE like php2shell, then sends the evil loopback packets to elevate to root.
Every consumer router-wirelessAP I've seen runs linux.
I never see security upgrades for the kernels they use. Not one. Not ever.
Is this something that should be a cause for concern? Even if you don't run a consumer-grade tplink, netgear, asus, whatever router yourself.
Update:
The context which I should have said is that I think
LLMs are giving script kids nation state level tools and expertise. This known bug is always less interesting than the next twenty.
This is one of the biggest benefits of using OpenWRT. They update the entire stack like every other month or so you’re never on an ancient kernel or utils.
The downside is it warns me if I update things will probably break and to take a backup, I've always been wary of updating my openwrt router because of that.
Is it a cause of concern? Yes. Will you get hacked? Most likely not.
Most consumer routers run ancient stacks with all kinds of weird proprietary vendor patches that are tied to a specific kernel version. Looking at you Qualcomm!
However they generally also drop all traffic by default so from the WAN side they look like a black hole. I guess that's "safe" enough for most consumers.
For fun, take a look at the open PRs in the OpenWRT repos. Those people are doing some amazing work but it's also frustrating to see them struggle with things that _could_ be easy if the vendor made some slightly better choices.
I would strongly prefer something like OpenWRT though. Call me paranoid but after the last few weeks I'm increasingly less comfortable with connecting my devices to a blackbox.
I know the context here is consumer routers, but for those reading who might be running a server, or a homelab, or anything that does accept connections from outside, I want to make it clear that "most likely not" turns into "likely trivial if targeted" for servers.
Can you give examples of kernel bugs that can be targeted on an otherwise locked down system? Let's say something running just LAMP and all other ports closed off with a software firewall. Even the one in this article seems very tricky to actually exploit.
There are plenty of out of support systems still in the wild and while I'm not arguing they are completely secure I was under the impression that kernel vulnerabilities generally make a minor exploit in something else worse (chained priv escalation for example) and generally are difficult to exploit on their own. If I'm mistaken I'd love to know
Any TCP packet (with IP options) you can send to a Linux kernel that doesn't get DROP'd by netfilter makes you technically vulnerable.
That includes the one port your LAMP stack uses.
Or, the attacker might be originating from the other side of a device, like within a LAN. Malicious employees/house guests/what-have-you could target devices from inside your network where most devices (not just servers) don't have as much protection and often have other open ports accessible, like file sharing/SMB, SSH, SMTP, RDP, databases etc.
AFAIK all the backdoors which have been found in some models are only accessible from the LAN side, and if you have an attacker inside the network, it's already too late.
> if you have an attacker inside the network, it's already too late.
Disagree, I think it's still a good idea to try to mitigate e.g. root escalations via this bug from devices within your network as well, as they could become compromised at any moment via other means and then promote themselves to root.
> if you have an attacker inside the network, it's already too late.
It shouldn't be though. Just because my smart lightbulb has an RCE and is isolated on its own VLAN with the router doesn't mean it should be able to attack the router and escalate to access my devices on other VLANs. It should only be able to do damage limited within its own vlan.
Any wireless AP is going to run Linux because BSD has had a tough time with anything better than 802.11n
Any router/switch is going to be running an ancient Linux kernel, which isn't patched. EdgeOS for example, won't be patched. The old Brocade/Ruckus switches I got here, won't have this patched. I'm holding my breath for my Unifi APs.
Google Wifi used to be good about that. I even got a notification for when my AP was being EOL'd and would no longer receive automatic firmware updates. I don't know if Google is still on the ball there. The Nest acquisition was confusing. Have the Google Wifi and Nest Wifi teams been merged?
But FWIW I run my AP in bridge mode. Wifi drivers and chip firmware seem to be riddled with exploits. I don't want my router & firewall on the same system as the AP.
According to its logs, my bargain basement Huawei 4G router from 2019 got an update last week.
I hope that most of these devices actually do update themselves automatically, and people just aren't noticing. (Whether the updates are any good is another matter!)
MIPS and Arm have no security issues ever so never need patching which is an x86 only thing? Or maybe this comment is no more relevant than pointing out not all Arm patches apply to x86?
LLMs are giving script kids nation state level tools and expertise.
I'm on Comcast in USA, I BYOD for the router and it's a Netgear C3500 (I think). It gets new firmware pushed to it from Comcast. Can't even load your own manually.
> BYOD = bring your own device? And yet your isp pushes updates to it?
Perhaps better characterized as "buy your own device". Cable ISPs will only support a limited set of modems on their network, and they're fully managed by the ISP, both in terms of software and configuration.
Correct, and they are limited in features. I'm going to swap out this combo-modem+WiFi and get a dedicated -modem and then have a different device as my "edge" router/NAT
Mosty probably: not. Traffic passes throught the ASIC, usually it is not processed on CPU or custom code. Unlike typical web servers, that reduces the attack surface as there aren't software parsers (a common place where vulnerabilities hide). Additionally the OS may have additional protection measures, such as signing each binary, signing the booting image and limiting which syscalls can be done by any binary.
The bug in question is not mitigated by any userspace actions (like signing binaries or boot images) or even by limiting syscalls, and all it takes is a listening socket on the device from either direction (inside or outside the network) to get root since router-destined packets are always processed by the CPU.
> Additionally the OS may have additional protection measures, such as signing each binary, signing the booting image and limiting which syscalls can be done by any binary.
I'd love for that to be true. Unfortunately, a lot of consumer routers are fairly straightforward Linux systems with no particular attempts at hardening.
All traffic? In particular, I'd expect new connection establishment to go through a software path, looking up port forwarding and firewall rules etc. and creating hardware forwarding table entries if applicable.
i admit they don't enumerate their patches well, but the major residential router vendors do regular updates , usually indicated with "security fixes" .
unfortunately most lack SSH access or a real ability to test the kernel patch version.
You may want to try to look at Gl.inet routers. I'm not sure yet what is their modus operandi, but at least Flint 2 has both their official firmware (with Mediatek's SDK, I guess) and a vanilla Openwrt version -- on their company page.
Their features (Glinet's mesh) are not compatible but it does seem better than just drop support totally.
There also seems to be some involvement in the upbringing of openwrt support -- so it's not a hands-off approach, but on the other hand I'm only looking at Flint 2 firmware/OS development.
Additionally, their Flint 3e uses Qualcomm so the route it takes may be different -- though I hope not!!!
Now, if we are talking about my ISP router/modem....... I'd rather not think about it.
PS: Gl.inet also has a router with fiber optic. Not sure how that would work if I wanted to throw my ISP's router/modem out the window.
Can't tell if Ubuntu is smart for deferring this patch until a bigger periodic release because there's no evidence it's been exploited and its exploitability is quite limited or if Ubuntu is dumb because there's a lot of publicity around this CVE and it comes off as failing to read the room.
I was previously Ubuntu-by-default for my random Linux boxen, but going to upgrade things and finding that Ubuntu didn't have a patch has shaken my faith in it. I probably won't choose it for new servers in the future.
Ubuntu was always the benefits of Debian without the inconvenient restrictions anyways, it wouldn't be surprising if their contradictive philosophy caused them to collapse by inability to eat the cake while having it simultaneously.
I've switched to Debian in professional servers around the time I learned how to use it without a Graphical Desktop Environments, that's roughly the stage at which a growing devleoper should drop the Ubuntu training wheels. Don't carry that into the workplace whenever possible.
I find it odd that no public discussions are highlighting that iptables or nftables can drop this before the kernel CVE is exploitable by a U32 drop rule in raw PREROUTING...
sudo iptables -t raw -I PREROUTING -p tcp -m u32 ! --u32 "0>>24&0x0F=5" -j DROP
I think they were saying if one uses such a rule in general (as is normal) but looking into it some more it seems it may actually be the default out of the box as well as long as you've enabled UFW:
> On installation, ufw is disabled with a default incoming policy of deny, a default forward policy of deny, and a default outgoing policy of allow, with stateful tracking for NEW connections for incoming and forwarded connections.
"default deny incoming" would apply to the filter INPUT chain, and there are several other chains that run before that one (raw, conntrack, mangle, nat), so it's better to use an explicit iptables or nft rule targeting the earliest "raw PREROUTING" chain.
Do note that eBPF can still bypass netfilter entirely, as can containers, or any process using a separate network namespace.
The reason I ask is that I have devices out there like that today and I need to know whether I should be freaking out or be completely calm.
I don't get this vulnerability, it's been out there for weeks, completely public, yet it has stayed extremely opaque. Usually, the day a high profile vulnerability is made public, there's all kinds of guidance out there explaining exactly who it affects how, and there are day 0 patches. This one I still don't know whether is an RCE or DoS in practice! I don't know whether a computer blocking connections via a "normal firewall" protects me or not! It's incredibly frustrating.
> This one I still don't know whether is an RCE or DoS in practice!
It is. It depends on dumping (potentially) a lot of traffic at you to "luck" into the cpu core scheduler but obviously computers can increase their luck by doing something a lot. The RCE is real, the risk is probably lower if you're not high profile simply because it requires some about of CPU spend on the attackers side, not something that you just fire off at anything that will respond to a ping or `GET /config.php`.
> I don't know whether a computer blocking connections via a "normal firewall" protects me or not! It's incredibly frustrating.
It doesn't. Apply the above rules until your kernel is updated.
Hard disagree... the packet could be part of an existing TCP stream and get processed by nf_conntrack before it hits the filter.INPUT chain. Or a local application could be sending packets to the loopback interface that bypasses the input chain in various ways.
A custom netfilter rule may already exist that accepts the packet before your mitigation does. And nftables allows custom chains that run at a higher priority than filter.INPUT and could be accepted before it touches it.
This is why vendors are suggesting to use the raw.PREROUTING chain instead.
And again, any processes that are part of a separate network namespace will not have the same netfilter rules applied that one has added within the main namespace.
Local processes / users can also easily bypass netfilter entirely by way of creating their own network namespaces via unshare/firejail/etc regardless of whether user namespaces are enabled.
I respect that there is a quick fix I can do myself with iptables. However, I don't want to have to do that. I run Ubuntu with automatic security patches enabled, I trust the Ubuntu team to issue quick fixes and my server to be patched most of the time while I am asleep. I don't want to have to manually edit my server configuration to patch security holes and have to remember all these little knobs when I get EOL and have to rebuild my server.
Which is the price Canonical chooses to charge, like they could charge money for this and they don't because they'd rather eat the loss for personal users for the advertising capability.
Except in this case they have not issued any fix, for several weeks, to an extremely critical issue that most other distros have already patched for some time now.
> The following nftables table drops incoming IPv4 TCP packets that contain IP options, which reduces exposure to this vulnerability but cannot fully prevent it.
I know things like eBPF can technically bypass netfilter if someone tries hard enough, but I'm not sure if they're referring to any other ways to exploit this that a netfilter rule can't catch.
Firejail by default also allows unprivileged users to completely bypass any netfilter rules from the main network namespace... so if the user has that installed (even if they don't use it), it's something to keep in mind.
A system running multiple network namespaces might not have the same mitigations applied inside all of them, so be sure to check all of those, except there's no great way of even finding all the different namespaces running on a Linux system for some reason (please prove me wrong if something exists! lsns is not enough).
But yes I agree it seems strange that most posts/articles/etc. are not prominently giving out mitigation strategies.
It's one of the major selling points of Linux that Linux admins can do this even if their OS vendor won't, but the whole point of having a distro is to stop you from needing to roll your own solutions, or to have to maintain constant vigiliance to be even remotely secure.
Just patched it everywhere on my servers last night. It’s getting increasingly more difficult to run things exposed to the internet, I get so many “sorry we’ve been hacked your data was stolen” e-mails recently it’s crazy, we must have armies of people combing through the web with AI attack tools.
I have been running pure Debian for years on my router/home server. Ubuntu is a superset of Debian that mostly adds stuff that is irrelevant for a networking device. Just stick to base distributions unless you are a very good reason not to.
Ubuntu made more sense when you installed end user apps from the apt repos and wanted updates more than every 2 years. These days with Flatpak you are better off on Debian.
tl;dr: What is a good vm/container tool (and OS to live inisde that) for a 16GB RAM m1 pro mac that doesn't guzzle up the resources and is safe/easy enough.
-----
I am thinking of setting up a VM/container on my 16GB M1 Pro mac. The purpose mainly is to move harnesses (and the world of NPM et al) to it. Currently I am using few harnesses – Pi, OpenCode, Claude, Codex, and agy (mostly opencode, cluade code, and agy) – kinda controlled by sandbox-exec. It has wroked sufficiently well, as much as sandbox-exec can faciliate that. But what makes me unconfomrtables are things like:
1. I still have to give these harnesses write access to folders I'd otehrwise not be comfortable (esp. at the time of upgrade etc)
2. The face at which updates arrive – and then you accept that and that could be unsafe, you don't accept that and that could be unsafe
3. The asntronomical number of packages/dependencies that get updated everyday (sometimes in one hour) and it's needed for these harnesses. That makes me very nervous.
What tool + OS should I pick for my container environment? (I am comfortable with cli but have enver used VM before other than just installing and playing with it for a few minutes)
My main goals are:
1. less RAM usage (I don't have a lot left after normal OS and few programs I'd like to have running; often I also try some local models – sometimes just for STT; but just the 3-6B range) - so would appreciate if RAM overhead by the cotnainer OS and tool isn't a lot.
2. If possible use available disk space flexibily (without having to reserve in advnace; or I should be able to change/increase the image/container's disk usage capacity)
3. relatively safe and easy to use
4. Not having to create another mac user and then do this logout -> login (it's not comfortable, besides it increases RAM usage)
At this point, just the failure to patch this single vulnerability is enough to make me write Ubuntu off as a viable distribution. Volunteer-run distros can fix this, why not a well-resourced company? It's not as if they haven't had the time to fix it, or that porting the fix to their current kernels would be particularly difficult.
This is the final nail in the coffin. I already order my new vservers with debian. I'll keep the existing ones on ubuntu for now, but will not consider ubuntu for anything server side in the future. My reasoning was "Debian is volunteers, they probably won't release fixes as fast as a commercial entity" - obviously the reasoning is backwards.
But I also must say, I am not giving a single dollar to ubuntu nor did I ever do so, so it is also perfectly valid for Ubuntu to not give a damn about me.
[OP] majorchord | 23 hours ago
However they also claim the mitigation "reduces exposure to this vulnerability but cannot fully prevent it", without elaborating on what that means or how it would be possible... I've also not seen anyone else make this claim.
giancarlostoro | 21 hours ago
That's what I call Steve Jobs style PR / Marketing you deflect the problem by making another problem to make it seem like you're fully aware. "You're holding the phone wrong"
worthless-trash | 16 hours ago
rf15 | 20 hours ago
jeroenhd | 10 hours ago
So far they're two days later than IBM/Red Hat. Not great, for sure, but they seem to be that far behind compared to the competition just yet.
If they push a backport that kills a million set top boxes because of changes in the network stack, they're not much better off than if they work a couple more days on testing.
hanumseous | 7 hours ago
gnabgib | 23 hours ago
jon128 | 22 hours ago
I am hoping to see an indication that https://grsecurity.net/ patches preemptively mitigated CVE-2026-98432 but I'm feeling less optimistic.
This vulnerability has been a stark reminder to why in some places things are running on HardenedBSD instead of Linux.
InsidiousWombat | 20 hours ago
jon128 | 20 hours ago
Separate from any of those, a good chunk of America's more sensitive production database clusters, other workloads, and some network applications are scheduled by Nomad or Kubernetes running directly on a kernel that is not Linux.
The most popular kernel is not necessarily appropriate for all use cases or types of environments where high-value targets reside.
Directly downstream of FreeBSD, HardenedBSD achieved its long history of invulnerability to what were exploitable vulnerabilities in upstream through its advancements in the operating system layer.
_> but security thru obscurity
Arrogance is stupidity.
doctorpangloss | 20 hours ago
jon128 | 20 hours ago
doctorpangloss | 20 hours ago
jon128 | 19 hours ago
In some advanced settings where tolerances for potential impacts resulting from exploited software vulnerabilities are intensely low, Kubernetes (userspace software) on HardenedBSD is choice for mission.
Like with the other two systems, those running Kubernetes userspace on them are self-sufficient and are not among the audience of external more general Kubernetes distro support channels.
If you happen to like RKE, check out https://Talos.dev which is the best Kubernetes on Linux distro (for most Kubernetes on Linux use cases).
_
@nijave
Nomad and Kubernetes schedule containers, which are rarely not Linux containers, yeah.
Both can schedule processes in FreeBSD's BSD Jails.
If someone wants to start an open Kubernetes on FreeBSD/HardenedBSD project I would encourage that. Personally I would like to see an open Kubernetes on illumos project.
As far as I know where Kubernetes on HardenedBSD runs is not interested in open projects.
doctorpangloss | a minute ago
nijave | 20 hours ago
Maybe you're thinking of jails? I'm not sure if there's any mainstream/popular distributed orchestration systems for those.
superkuh | 22 hours ago
"Freenode" in 2026 is an entity actively hostile to open source software and while what they say here might be true, they are only saying it because it aligns with their biases. They are not a real news source.
My mitigation for this lack of patches has been using iptables.
maeln | 21 hours ago
wolvoleo | 14 hours ago
I was afraid the community would get splintered out but it didn't. The quick decisive action in establishing libera really helped that. Only a few channels left to go somewhere else (like alpine linux)
I really loved this example of a community massively saying Nope and just avoiding a hostile takeover altogether. Unfortunately it doesn't usually happen so well.
account42 | 5 hours ago
dark-star | 19 hours ago
soraminazuki | 10 hours ago
But lately, the Freenode fiasco has been on the back of my mind from time to time. If the same thing happened today, a good chunk of the comments would be cheering for Freenode. Almost every single article critical of the takeover would be flagged too.
account42 | 5 hours ago
But only after it was clear that most channels were leaving anyway. In a way, that was the only good thing he did in this mess: reduce fragmentation that we would have gotten as some stuck to Freenode.
c0l0 | 21 hours ago
It does not seem that complicated to patch, and kernel upgrades are a dime a dozen these days. So what is making them delay a fix for so long?
wolvoleo | 14 hours ago
Which is a bad choice considering the impact imo.
The problem is as organisations grow, the more governance types get involved and slow things down with arbitrary rules.
cm2187 | 9 hours ago
philipallstar | 8 hours ago
lambdaone | 7 hours ago
dataengineer56 | 6 hours ago
cm2187 | 5 hours ago
bibelo | 7 hours ago
So indeed, there's probably some reason
PcChip | 4 hours ago
wolvoleo | 2 hours ago
TacticalCoder | 21 hours ago
And do we feel about running no firewall?
For example Linux works totally fine without running dhcpcd and without running avahi. But on many systems where people think they run "no service", these two shall have ports open on your network and happily listening.
Sure, it's typically "only" on your LAN: but how long before we see exploits affecting even private LANs?
Tighten your stuff up guys.
viraptor | 20 hours ago
I think you'll find that for most people, plugging the laptop in and having the network immediately available and automatically seeing their local printer is a pretty good reason.
InsidiousWombat | 20 hours ago
how you refresh dhcp lease if there is no service taking care of ? Why you think DHCP is pointless ?
smackeyacky | 20 hours ago
nijave | 20 hours ago
halJordan | 20 hours ago
warkdarrior | 12 hours ago
gwbas1c | 19 hours ago
warpspin | 20 hours ago
[OP] majorchord | 18 hours ago
> A zero-day (also known as a 0-day) is a vulnerability or security hole in a computer system unknown to its developers or anyone capable of mitigating it.
Apparently the meaning has evolved over the years.
[1] https://en.wikipedia.org/wiki/Zero-day_vulnerability
TutleCpt | 19 hours ago
snapplebobapple | 19 hours ago
pelagicAustral | 19 hours ago
secabeen | 19 hours ago
SP3269 | 19 hours ago
pelagicAustral | 19 hours ago
So you make of that what you will, but in my experience it means the exploitation comes hand in hand with some far more esoteric chain that is not easy to put in practice.
KAMSPioneer | 8 hours ago
I accept that one would expect substantially shorter timelines in the age of AI, but given that WannaCry is suspected of being state-funded and even then: 1) it took a month to appear and 2) ETERNALBLUE and DOUBLEPULSAR were essentially already finished, weaponized malware off the shelf.
All of this to say that just because you don't see mass exploitation right this second doesn't mean there isn't a wormable exploit that needs patching. WannaCry didn't exist until it did, and suddenly everyone who failed to patch beforehand looked foolish.
worthless-trash | 16 hours ago
[OP] majorchord | 18 hours ago
How would you see one? Have you deployed your own army of honeypots that are actively logging and looking for corrupt IP options packets? If you're not already looking for it, I think it's very likely that most people who get hacked will have no way of knowing how it happened, so we can't consider those incidents as being a result of this particular bug.
Are you actually monitoring for this specifically or just trying to use a sample size of 1 group of servers to prove something?
pelagicAustral | 17 hours ago
ranger_danger | 16 hours ago
evilDagmar | 16 hours ago
yxhuvud | 11 hours ago
GordonS | 9 hours ago
ranger_danger | 47 minutes ago
https://access.redhat.com/security/vulnerabilities/RHSB-2026...
> Multipath TCP is not a required condition for this flaw.
I contemplated not even responding to you because of how unbelievably abusive and rude I have watched you be on IRC to me and many others for literally decades, and was a big reason why I quit, but I decided this issue was more important for everyone else's sake that I should still speak up.
handoflixue | 16 hours ago
Most of the October CVEs we've seen were situations that require an extremely complex chain of attacks - amazing if you want to break into a bank, or fuck up some grocery freezers, not so useful for scanning random public boxes and doing "hacks of opportunity"
Apologies for the lack of links - it's a scattering of random people I've read talking about this, not any coordinated study. But if you read between the lines on all of the threads about these CVEs, you'll notice a lack of "I was hit by this" stories - which is pretty normal for more complex vulnerabilities.
The reason these sat unnoticed for so long is precisely because it's difficult to actually exploit them.
If anyone has anecdotes to the contrary, I'm definitely interested in hearing it, of course!
dboreham | 15 hours ago
sheepscreek | 14 hours ago
That's a very naive view on it (now) and I'll try my best to explain why.
The bad actors can discover multiple 0-days on the fly to gain access to an employer/company infrastructure with valuable data/information, or in the case of Binance, crypto. They could chain multiple such attacks going from one-company/provider to another, to another, to eventually reach their target.
To anyone interested, this is good reading material on the topic:
(1) study the Binance hack in detail
(2) OpenAI huggingface hack - focus on 0-day chaining
lambdaone | 7 hours ago
ay | 19 hours ago
Why does everything has to come with so much drama these days.
cpburns2009 | 17 hours ago
[OP] majorchord | 17 hours ago
> The following nftables table drops incoming IPv4 TCP packets that contain IP options, which reduces exposure to this vulnerability but cannot fully prevent it.
I suppose locally-running malicious programs could send packets to the loopback address or one of its own IPs or something.
[1] https://access.redhat.com/security/vulnerabilities/RHSB-2026...
userbinator | 14 hours ago
That's Game Over already.
ranger_danger | 13 hours ago
perching_aix | 12 hours ago
ranger_danger | 13 hours ago
harry8 | 18 hours ago
Every consumer router-wirelessAP I've seen runs linux.
I never see security upgrades for the kernels they use. Not one. Not ever.
Is this something that should be a cause for concern? Even if you don't run a consumer-grade tplink, netgear, asus, whatever router yourself.
Update: The context which I should have said is that I think LLMs are giving script kids nation state level tools and expertise. This known bug is always less interesting than the next twenty.
markasoftware | 18 hours ago
snapplebobapple | 18 hours ago
wps | 18 hours ago
davrosthedalek | 18 hours ago
za_creature | 15 hours ago
Saris | 5 hours ago
fn-mote | 4 hours ago
I haven’t had problems upgrading in at least 4 years. There’s a screen where you click a button to auto-upgrade.
Just get it installed.
LaurensBER | 18 hours ago
Most consumer routers run ancient stacks with all kinds of weird proprietary vendor patches that are tied to a specific kernel version. Looking at you Qualcomm!
However they generally also drop all traffic by default so from the WAN side they look like a black hole. I guess that's "safe" enough for most consumers.
For fun, take a look at the open PRs in the OpenWRT repos. Those people are doing some amazing work but it's also frustrating to see them struggle with things that _could_ be easy if the vendor made some slightly better choices.
I would strongly prefer something like OpenWRT though. Call me paranoid but after the last few weeks I'm increasingly less comfortable with connecting my devices to a blackbox.
[OP] majorchord | 18 hours ago
dwedge | 18 hours ago
There are plenty of out of support systems still in the wild and while I'm not arguing they are completely secure I was under the impression that kernel vulnerabilities generally make a minor exploit in something else worse (chained priv escalation for example) and generally are difficult to exploit on their own. If I'm mistaken I'd love to know
[OP] majorchord | 17 hours ago
That includes the one port your LAMP stack uses.
Or, the attacker might be originating from the other side of a device, like within a LAN. Malicious employees/house guests/what-have-you could target devices from inside your network where most devices (not just servers) don't have as much protection and often have other open ports accessible, like file sharing/SMB, SSH, SMTP, RDP, databases etc.
cjbprime | 17 hours ago
harry8 | 17 hours ago
How about forwarding a port for wireguard or ssh?
userbinator | 14 hours ago
AFAIK all the backdoors which have been found in some models are only accessible from the LAN side, and if you have an attacker inside the network, it's already too late.
ranger_danger | 13 hours ago
Disagree, I think it's still a good idea to try to mitigate e.g. root escalations via this bug from devices within your network as well, as they could become compromised at any moment via other means and then promote themselves to root.
TheDong | 8 hours ago
It shouldn't be though. Just because my smart lightbulb has an RCE and is isolated on its own VLAN with the router doesn't mean it should be able to attack the router and escalate to access my devices on other VLANs. It should only be able to do damage limited within its own vlan.
j45 | 16 hours ago
I'm thinking of what to do myself.
The ghost of android releasing 6 month old patches every month is going to come home to roost.
iAMkenough | 18 hours ago
qw3rty01 | 18 hours ago
harry8 | 18 hours ago
That would be a bold claim and one I've never heard before. Hard not to be suspicious of it.
This security flaw is unlikely to be the only security flaw that ever happens to a kernel used by routers.
The point: Consumer routers do not seem to get updates. Ever.
Is that a worry or what?
harry8 | 17 hours ago
But what do I know about it beyond that?
wolvoleo | 13 hours ago
Some more commercial ones (pfsense is commercial) use it too like juniper junos.
harry8 | 11 hours ago
Fnoord | 9 hours ago
Any router/switch is going to be running an ancient Linux kernel, which isn't patched. EdgeOS for example, won't be patched. The old Brocade/Ruckus switches I got here, won't have this patched. I'm holding my breath for my Unifi APs.
ahoka | 9 hours ago
nullpoint420 | 14 hours ago
wolvoleo | 13 hours ago
Of course it does. But it was not affected in this particular case.
wahern | 18 hours ago
But FWIW I run my AP in bridge mode. Wifi drivers and chip firmware seem to be riddled with exploits. I don't want my router & firewall on the same system as the AP.
toofy | 17 hours ago
at least with an open source os there are more paths forward.
deepsun | 17 hours ago
Granted they are pro-sumer devices, not strictly consumer.
cpburns2009 | 17 hours ago
harry8 | 17 hours ago
roryirvine | 7 hours ago
I hope that most of these devices actually do update themselves automatically, and people just aren't noticing. (Whether the updates are any good is another matter!)
angry_octet | 4 hours ago
iberator | 17 hours ago
harry8 | 17 hours ago
MIPS and Arm have no security issues ever so never need patching which is an x86 only thing? Or maybe this comment is no more relevant than pointing out not all Arm patches apply to x86?
LLMs are giving script kids nation state level tools and expertise.
edoceo | 15 hours ago
harry8 | 14 hours ago
Defintely seems potentially better than no updates at all.
How do you know if you have a wide open CVE they're ignoring, like it's claimed ubuntu are doing?
duskwuff | 11 hours ago
Perhaps better characterized as "buy your own device". Cable ISPs will only support a limited set of modems on their network, and they're fully managed by the ISP, both in terms of software and configuration.
edoceo | 11 hours ago
kevin_thibedeau | 17 hours ago
estebarb | 17 hours ago
[OP] majorchord | 13 hours ago
duskwuff | 12 hours ago
I'd love for that to be true. Unfortunately, a lot of consumer routers are fairly straightforward Linux systems with no particular attempts at hardening.
lxgr | 8 hours ago
pengaru | 16 hours ago
full disclosure: they're my employer
Chaosvex | 5 hours ago
tonymet | 15 hours ago
unfortunately most lack SSH access or a real ability to test the kernel patch version.
0rzech | 15 hours ago
[1] https://www.turris.com
harry8 | 14 hours ago
10G looks a little less consumer-y. I'd imagine cisco, arista, etc do security updates if you want to go down that path.
0rzech | 12 hours ago
AFAIK, none of them supports WPA Enterprise, which I'd consider necessary for enterprise-grade equipment.
OTOH, they are Open Source and Open Hardware, are still pretty powerful, and are much cheaper than enterprise HW.
PS. The proper link to their site in English is https://www.turris.com/en/ .
lambdaone | 7 hours ago
user2722 | 6 hours ago
Their features (Glinet's mesh) are not compatible but it does seem better than just drop support totally.
There also seems to be some involvement in the upbringing of openwrt support -- so it's not a hands-off approach, but on the other hand I'm only looking at Flint 2 firmware/OS development.
Additionally, their Flint 3e uses Qualcomm so the route it takes may be different -- though I hope not!!!
Now, if we are talking about my ISP router/modem....... I'd rather not think about it.
PS: Gl.inet also has a router with fiber optic. Not sure how that would work if I wanted to throw my ISP's router/modem out the window.
Maken | 5 hours ago
[1] https://mikrohub.io/mikrotik/changelog/long-term/7.24.6
sivakon | 18 hours ago
you can use proposed pocket as per this for latest LTS releases
rogeliodh | 14 hours ago
globular-toast | 11 hours ago
rogeliodh | 4 hours ago
mbac32768 | 18 hours ago
jimrandomh | 18 hours ago
heatpump5n | 9 hours ago
TZubiri | 17 hours ago
Ubuntu was always the benefits of Debian without the inconvenient restrictions anyways, it wouldn't be surprising if their contradictive philosophy caused them to collapse by inability to eat the cake while having it simultaneously.
I've switched to Debian in professional servers around the time I learned how to use it without a Graphical Desktop Environments, that's roughly the stage at which a growing devleoper should drop the Ubuntu training wheels. Don't carry that into the workplace whenever possible.
cpburns2009 | 17 hours ago
jordemort | 16 hours ago
debinguy | 16 hours ago
sudo iptables -t raw -I PREROUTING -p tcp -m u32 ! --u32 "0>>24&0x0F=5" -j DROP
mort96 | 15 hours ago
wolvoleo | 14 hours ago
zamadatix | 13 hours ago
> On installation, ufw is disabled with a default incoming policy of deny, a default forward policy of deny, and a default outgoing policy of allow, with stateful tracking for NEW connections for incoming and forwarded connections.
https://manpages.ubuntu.com/manpages/noble/man8/ufw.8.html
The more specific rule is probably more useful if you do intend to expose a TCP port though.
wolvoleo | 11 hours ago
mort96 | 8 hours ago
zamadatix | 2 hours ago
ranger_danger | 13 hours ago
Do note that eBPF can still bypass netfilter entirely, as can containers, or any process using a separate network namespace.
mort96 | 9 hours ago
I don't get this vulnerability, it's been out there for weeks, completely public, yet it has stayed extremely opaque. Usually, the day a high profile vulnerability is made public, there's all kinds of guidance out there explaining exactly who it affects how, and there are day 0 patches. This one I still don't know whether is an RCE or DoS in practice! I don't know whether a computer blocking connections via a "normal firewall" protects me or not! It's incredibly frustrating.
GCUMstlyHarmls | 6 hours ago
https://almalinux.org/blog/2026-10-30-cve-2026-98432/#tempor...
> This one I still don't know whether is an RCE or DoS in practice!
It is. It depends on dumping (potentially) a lot of traffic at you to "luck" into the cpu core scheduler but obviously computers can increase their luck by doing something a lot. The RCE is real, the risk is probably lower if you're not high profile simply because it requires some about of CPU spend on the attackers side, not something that you just fire off at anything that will respond to a ping or `GET /config.php`.
> I don't know whether a computer blocking connections via a "normal firewall" protects me or not! It's incredibly frustrating.
It doesn't. Apply the above rules until your kernel is updated.
InsidiousWombat | 6 hours ago
ranger_danger | an hour ago
Hard disagree... the packet could be part of an existing TCP stream and get processed by nf_conntrack before it hits the filter.INPUT chain. Or a local application could be sending packets to the loopback interface that bypasses the input chain in various ways.
A custom netfilter rule may already exist that accepts the packet before your mitigation does. And nftables allows custom chains that run at a higher priority than filter.INPUT and could be accepted before it touches it.
This is why vendors are suggesting to use the raw.PREROUTING chain instead.
And what I said before about eBPF still stands... someone has written a mitigation that uses that instead since it runs even before PREROUTING: https://github.com/sevaepsteyn/cve-2026-98432-ebpf
And again, any processes that are part of a separate network namespace will not have the same netfilter rules applied that one has added within the main namespace.
Local processes / users can also easily bypass netfilter entirely by way of creating their own network namespaces via unshare/firejail/etc regardless of whether user namespaces are enabled.
mort96 | 59 minutes ago
Lord_Zero | 14 hours ago
retube | 14 hours ago
How much do you pay for this?
thenewnewguy | 14 hours ago
Which is the price Canonical chooses to charge, like they could charge money for this and they don't because they'd rather eat the loss for personal users for the advertising capability.
whatshisface | 8 hours ago
voakbasda | 2 hours ago
aucisson_masque | 9 hours ago
[OP] majorchord | 13 hours ago
Except in this case they have not issued any fix, for several weeks, to an extremely critical issue that most other distros have already patched for some time now.
For many this now means their trust is broken.
ranger_danger | 14 hours ago
Red Hat also claims that netfilter is not enough: https://access.redhat.com/security/vulnerabilities/RHSB-2026...
> The following nftables table drops incoming IPv4 TCP packets that contain IP options, which reduces exposure to this vulnerability but cannot fully prevent it.
I know things like eBPF can technically bypass netfilter if someone tries hard enough, but I'm not sure if they're referring to any other ways to exploit this that a netfilter rule can't catch.
Firejail by default also allows unprivileged users to completely bypass any netfilter rules from the main network namespace... so if the user has that installed (even if they don't use it), it's something to keep in mind.
A system running multiple network namespaces might not have the same mitigations applied inside all of them, so be sure to check all of those, except there's no great way of even finding all the different namespaces running on a Linux system for some reason (please prove me wrong if something exists! lsns is not enough).
But yes I agree it seems strange that most posts/articles/etc. are not prominently giving out mitigation strategies.
amluto | 12 hours ago
LtdJorge | 8 hours ago
lambdaone | 7 hours ago
throwaway63467 | 6 hours ago
bbrake | 32 minutes ago
drnick1 | 15 hours ago
Gigachad | 12 hours ago
h4kunamata | 15 hours ago
Ubuntu abandoned its way years ago, closing source, centralising features under Canonical umbrella., etc.
Seeing people complaining about Ubuntu tells me more about them than the company itself.
riffic | 14 hours ago
justauserid | 13 hours ago
-----
I am thinking of setting up a VM/container on my 16GB M1 Pro mac. The purpose mainly is to move harnesses (and the world of NPM et al) to it. Currently I am using few harnesses – Pi, OpenCode, Claude, Codex, and agy (mostly opencode, cluade code, and agy) – kinda controlled by sandbox-exec. It has wroked sufficiently well, as much as sandbox-exec can faciliate that. But what makes me unconfomrtables are things like:
1. I still have to give these harnesses write access to folders I'd otehrwise not be comfortable (esp. at the time of upgrade etc)
2. The face at which updates arrive – and then you accept that and that could be unsafe, you don't accept that and that could be unsafe
3. The asntronomical number of packages/dependencies that get updated everyday (sometimes in one hour) and it's needed for these harnesses. That makes me very nervous.
What tool + OS should I pick for my container environment? (I am comfortable with cli but have enver used VM before other than just installing and playing with it for a few minutes)
My main goals are:
1. less RAM usage (I don't have a lot left after normal OS and few programs I'd like to have running; often I also try some local models – sometimes just for STT; but just the 3-6B range) - so would appreciate if RAM overhead by the cotnainer OS and tool isn't a lot.
2. If possible use available disk space flexibily (without having to reserve in advnace; or I should be able to change/increase the image/container's disk usage capacity)
3. relatively safe and easy to use
4. Not having to create another mac user and then do this logout -> login (it's not comfortable, besides it increases RAM usage)
totetsu | 12 hours ago
try orbstack server or multipass if all this hasn't put you off canonical.
justauserid | 11 hours ago
Yes, I did.
> in a completly unrelated discussion
I disagree with this point. One of the reasons I have not been able to start on it is I don't know what OSes are secure for this purpose.
> orbstack server or multipass
I will try this. Thank you.
moi2388 | 11 hours ago
atoav | 12 hours ago
lambdaone | 7 hours ago
littlecranky67 | 7 hours ago
But I also must say, I am not giving a single dollar to ubuntu nor did I ever do so, so it is also perfectly valid for Ubuntu to not give a damn about me.
sylware | 6 hours ago
pjmlp | 5 hours ago
Yes, they should probably by now have shipped it, yet all those complaints are the usual entitled folks that only leech distros.
tapoxi | 5 hours ago
pjmlp | 3 hours ago
Good for Debian, they were not required to.
fl0ki | 19 minutes ago
Debian: And you'll never have to.
sakoula | 5 hours ago
https://ubuntu.com/blog/fixes-available-in-proposed-pocket-f...
Gud | 5 hours ago