Comment by BoppreH

16 days ago

Worse yet, the Ubuntu security incident page[0] lists "Ubuntu Priority: High" (as opposed to Critical, which would trigger an out-of-schedule patch). There's a "Why this priority?" link that simply says:

> Why is this CVE high priority?

> See ubuntu cvss score

That's it, we don't even get a period on this TBD of a sentence. It's followed by a generic link that explains what "Ubuntu Priority" is, but that page contradicts their choice. Note that it doesn't offer any mitigations either, and even the (horrendously unclear) description is copied from the NIST page, so there's no actual message from Ubuntu.

I already had a bad feeling the first time I saw "Get more security updates through Ubuntu Pro" on `apt upgrade`, this is now the kick I needed to move to a different distro.

[0] https://ubuntu.com/security/CVE-2026-98432#priority-reason

I don't think this is fair. They don't do a good job of pointing you at their rationale which is in their security score here[0]. I ultimately had to have claude translate the kernelese in to something my small brain could understand but this is a bug that could crash your system but is unlikely to expose your confidential system data to attackers or degrade data on the system. Especially given that the exploit can only be done if packets with IP options. IP options are commonly dropped or removed from packets on the open web. You not only need a vulnerable system but a vulnerable request path through the internet which you would only know by trying this attack which is very timing dependent with the SYN,Retransmitted SYN and Reset all happening very close together. A motivated attacker could probably do this on your network but over the internet ... It sounds unlikely.

[0]https://ubuntu.com/security/CVE-2026-98432#impact-score

  • > this is a bug that could crash your system but is unlikely to expose your confidential system data to attackers or degrade data on the system

    Red Hat disagrees[0][1]. They might be parroting Anthropic, and Anthropic might be exagerating the impact (likely!), but that's the current consensus about the vulnerability. If Ubuntu disagrees, they should have said something and not pasted "Cvss 3 Severity Score 9.8 · Critical" at the top of the page.

    > A motivated attacker could probably do this on your network

    I find that's still catastrophic given how centralized cloud environments are.

    [0] https://access.redhat.com/security/cve/cve-2026-98432

    [1] https://access.redhat.com/security/vulnerabilities/RHSB-2026...

    • > I find that's still catastrophic given how centralized cloud environments are.

      It's not like every cloud environment is vulnerable. GCP drops these options for example.

      > If Ubuntu disagrees, they should have said something and not pasted "Cvss 3 Severity Score 9.8 · Critical" at the top of the page.

      Again I'm going not going to argue they did a good job at pointing at their rationale but it looks like 9.8 is from CVSS 3.0 and they did prioritization with 4.0 which is 8.3. They definitely should have put that up front.

      1 reply →

    • > If Ubuntu disagrees, they should have said something

      Exactly. There's room for considered differences of opinion, but that consideration requires discorse and explanation! The lack of communication here is just absolutely baffling. It's like they're deliberately trying to trash their own brand.

      I mean, seriously what's going on? We know they have a process for emergency kernel updates, because of course they do. If they're not going to pull the trigger on this particular one (that everyone else is pulling the trigger on) at least make the case.

  • Greg Kroah-Hartman the long-time Linux kernel maintainer (USB, PCI, driver core, sysfs, and -stable releases) gave a talk on LLM assisted CVE discoveries. In that talk, he downplayed those exposed CVEs from Mythos. What you described above matches his sentiment on downgrade those CVE vuln score.

    See previous discussion and the actual talk on YouTube:

    https://news.ycombinator.com/item?id=49929391

  • > but this is a bug that could crash your system but is unlikely to expose your confidential system data to attackers

    Given people have demonstrated getting root with it, unlikely seems questionable

    > Especially given that the exploit can only be done if packets with IP options. IP options are commonly dropped or removed from packets on the open web.

    It sounds like iphones use mptcp in the wild, which use tcp options, so im not sure how true this is.

    • > It sounds like iphones use mptcp in the wild, which use tcp options, so im not sure how true this is.

      There's so much commentary misunderstanding what ip-options are that I find it difficult to make any judgement call. I do know that I've got routers on 5 continents with hundreds of public IPs and none have triggered an IP packet with an ip-option arriving on the WAN side, or seen on the internal side other than ones I've created with scapy trying to reproduce this. I also can't get these packets to traverse any internet paths I've tried.

    • TCP options are quite different from IP options.

      IP options are annoying in big fast routers because they have to take the slow route via CPU. As there's not a lot of use in accepting packets (externally at least) with IP options, they tend to be dropped totally (but will almost certainly be policed aggressively to the CPU as the CPU is not good for moving many packets).

  • A problem, at least for me at work, is this has to be communicated to the people who make the decision around whether something can continue operating or has to be taken offline until fixed. So far, said people are not particularly interested in the details of CVEs, just if it's critical and has an exploit available. Which is fair I guess, they're absorbing the risk. I would like Canonical to go a bit faster which would make my life easier.

Yeah, their lack of a satisfactory response is the final nail in their coffin. This incident escalates my consulting advice about Ubuntu from “not recommended” to “switch immediately”.

Some of my clients may be surprised by this change, as I once advocated for Ubuntu as a good choice for a development and deployment platform. While that has not been true for quite a while, they have now jumped the shark into eel-infested waters; I cannot take them seriously any longer.

  • Plain Debian was always best option; even RHEL dropped the ball on this CVE and that's distro people pay for support

    • Not always. Notably, if you have compliance requirements like needing FIPS encryption (say because you have government customers), or that requires you to use a commercial vendor, debian won't work. And if you need more up to date system packages, or a reliable release cadence, Ubuntu might be better.

      On desktop, if you need proprietary drivers, Ubuntu is easier to set up.

      But yes, if debian works for you, I wouldn't use Ubuntu.

      13 replies →

    • I planned to install Debian in 2009 and in 2014 but the installer was too hard to use, the desktop in 2009 was ugly (maybe in 2014 too), I could not get a running system that I could use to work. Fast forward to Debian 10, same laptop from 2014, and I got a desktop identical to Ubuntu GNOME.

    • Debian for hosting anything, Fedora for development desktop.

      That's been my rule for years at this point, I think I switched back to debian everywhere after the netplan changeover in ubuntu so somewhere around 17-18 (not against it in principle, just the way they did it in practice).

      No real regrets either.

  • The general advice I've been seeing the past ~8 years is "There is no reason to pick Ubuntu over Debian". I don't know who the opposite advice comes from and why anyone would pick ubuntu for a server.

    • 15-20 years ago the joke was that Ubuntu is a Zulu word that, roughly translated, means "I don't know how to configure Debian". Ubuntu was one of the first distros that allowed one to easily download proprietary WLAN drivers without dealing with the terminal, ndiswrapper, or compiling half-broken OSS implementations; for this reason it caught fire with those who wanted to try Linux but were otherwise unwilling to engage in such shenanigans. I tried it on a laptop around this time and was quite impressed. This reputation for "it just works" has carried them ever since even though other distros now do largely the same thing ( cheap, obscure, or recently-released hardware excepted).

      I took the lazy Mint route with my most recent machine since I wanted to use it for light gaming with minimal fuss. I'm going to redo it with Nix sometime soon because the value prop simply isn't there and I'd like to try a declaratively-configured distro.

      13 replies →

    • I advocated for Ubuntu for years because it was easier to install and had excellent hardware support. For 2026 I have switched back to Debian because Ubuntu has been making my life more difficult. Ubuntu 26 basically forces you onto Wayland even though it breaks some of my use cases, particularly with screen recording. It turns everything into Snaps which is a hassle because it means I need to mirror both the regular repo and the snaps to my closed network instead of just the repo.

      Ubuntu still has the superior installer by a good margin, but it does sometimes like to crash when run offline. Debian's installer at least has the option to go text only. There is also more work to do after the install is finished on Debian, but since that's a script it's not such a big deal.

      The one bit of the Debian installer I dislike the most is how if you aren't online when you do the install you get a completely empty sources.list, and the Debian sources are a lot more complicated than Ubuntu's. Debian's support for third party drivers like nVidia graphics is also more complicated and confusing than Ubuntu's.

      1 reply →

    • Enterprises love paying for support contracts. If you have one, you can demonstrate something has been accounted for. A secure software without support is worse from an audit perspective than a security disaster with proper paperwork. Most of these people never actually involve support they paid for ever.

      6 replies →

    • I deployed Ubuntu on a LOT of servers 10 years ago because at the time, they had a track record of doing good things in, with, and for, the Linux and open source community. I had already been using it on the desktop for a long time by that point and knew it well. Ubuntu was far and away the most popular; it was a very safe and sane choice.

      That changed when they started heavily promoting Snaps as the next big thing and started putting features and updates behind their enterprise offerings. Basically, they stopped being "Debian, but better." And Debian also made lots of incremental improvements in user-friendliness in that time. Basically, Debian is now pretty much what Ubuntu always promised to be.

    • The only recent reason I’ve managed to come up with is it far and away has the best support for Nvidia’s various drivers and utilities for running GPUs and its assorted ecosystem, like Infiniband and Spectrum-X Ethernet.

      If there’s another option I’m unaware of, I’d love to hear it. Also to be fair, I didn’t evaluate Sid for this, only Stable.

    • I've set up a new server literally couple weeks ago on a droplet on Digital Ocean - Ubuntu LTS is the default choice(you can select other distros but Ubuntu is the default). So I guess I picked Ubuntu because I haven't done Linux admin in over a decade and it's presented as the default choice by a big cloud provider.

      1 reply →

    • Certain AMD graphics drivers require specific Ubuntu releases and versions. I guess one might reasonably ask "why would you run AMD graphics drivers?" to which I don't really have a good answer. But that's why I run an Ubuntu box ;p. I'd so much rather be running Alpine.

The more I look at this the angrier I get. This is a CVE that Anthropic claims can get a Remote Code Execution on vulnerable systems, which includes default Ubuntu installs running any sort of Go server like Caddy. Bad enough that I shut down my personal website and self-hosted tools (sorry for any typos, my grammar checker is also down).

And this is how Ubuntu describes the vulnerability to sysadmins:

  In the Linux kernel, the following vulnerability has been resolved:
  ipv4: free inet_opt and ireq_opt after an RCU grace period
  tcp_v4_syn_recv_sock() transfers ownership of ireq->ireq_opt to the
  child socket (newinet->inet_opt) without copying it.
  Another cpu can concurrently process a retransmitted SYN for the same
  request socket, and send a SYNACK from tcp_check_req().
  tcp_v4_send_synack() and inet_csk_route_req() read ireq->ireq_opt
  under rcu_read_lock() only, and ip_build_and_send_pkt() and
  ip_options_build() then read opt->optlen twice.
  Note that the SYNACK timer itself is not an issue: it holds a
  reference on its request socket, and inet_csk_reqsk_queue_drop()
  calls timer_delete_sync() before the child can be freed.
  Since commit 079096f103fa (“tcp/dccp: install syn_recv requests
  into ehash table”), request sockets are processed without holding
  the listener lock, so nothing prevents the child socket from being
  freed while the SYNACK is still being built. TCP child sockets do
  not have SOCK_RCU_FREE, and inet_sock_destruct() frees inet_opt
  with a plain kfree(), leading to a use-after-free in
  ip_options_build().
  A similar issue exists with request socket migration
  (net.ipv4.tcp_migrate_req=1, or a BPF_SK_REUSEPORT_SELECT_OR_MIGRATE
  program). reqsk_timer_handler() clones the request socket with
  inet_reqsk_clone(), so that the old request socket and its clone
  share the same ireq_opt, then reqsk_migrate_reset() clears the
  pointer in the old one. Another cpu holding a reference on the old
  request socket can still be using these options (sending a SYNACK,
  or creating a child in tcp_v4_syn_recv_sock()) when the clone is
  freed, and tcp_v4_reqsk_destructor() also uses a plain kfree().
  Readers already use RCU, and other paths replacing inet_opt
  (do_ip_setsockopt(), cipso_v4_sock_setattr()...) already use
  kfree_rcu(). Use kfree_rcu() in inet_sock_destruct() and
  tcp_v4_reqsk_destructor() as well.
  IPv6 is not affected by the first issue, because tcp_v6_syn_recv_sock()
  duplicates the options. tcp_v6_reqsk_destructor() has the same
  migration issue with ipv6_opt, which is only set by CALIPSO. This
  will be addressed in a separate patch.

How is this in any way, shape, or form acceptable two weeks after the CVE was published?

  • sorry for any typos, my grammar checker is also down

    I assume this is unrelated to the Ubuntu situation, but how does a grammar checker go down? We've been doing offline grammar checking since computers had 64K of RAM.

  • > running any sort of Go server like Caddy

    It's so much worse than this, and there is a staggering amount of misinformation running rampant about this bug.

    MPTCP is not a requirement, it can be exploited merely by sending specially crafted TCP packets with (any) IP options set. No special MPTCP support, web servers or any specific software listeners are even necessary at all to trigger it, just send the packets and boom.

    I think the reason people assume MPTCP is a requirement is because that's what the PoC used... but MPTCP itself also doesn't even rely solely on IP options.

    Ubuntu also does not even list a single mitigation strategy... at least RHEL did.

    And like I said, any normal TCP packet that contains IP (not TCP) options can trigger the bug. I cannot emphasize enough how important and absolutely ridiculous this is.

    Like the post says, this is big enough to get a massive amount of people to completely stop trusting Ubuntu entirely and jump ship.

Moved to Debian few years back and never looked back. We used Ubuntu for a bunch of clients and unfortunately had to let them know they will have to migrate to Debian or similar. It's such a shame because I really used to love Ubuntu until Canonical started turning slowly into your typical everyday corporate.

As far as I know the only working exploit is working on machine with MultiTCP enabled. Redhat also downgraded the issue to "high" and also did not release a patch yet. I would have preferred a patch but I accept that they know better than me what they are doing.

  • Why would they take so long anyway and be behind volunteer run projects like debian? Is it something about corporate shenanigans and testing? because afaik in fedora for example firefox is always behind in security updates compared to debian.

  • RHEL did release a new kernel. Presumably it has the fix (though I guess I didn't actually check)

They're totally asleep at the wheel, it's malpractice. The pitch for RHEL is highly competent maintenance you can't get elsewhere

  • RHEL also dropped a ball on it so yeah the alternative for paying for RHEL license is apparently "just use Debian or hire some Arch nerds"

> I already had a bad feeling the first time I saw "Get more security updates through Ubuntu Pro" on `apt upgrade`

To be clear, Ubuntu Pro doesn't have the fix yet either, in case anyone's confused.

  • This was an old gripe unrelated to today's issue. But your clarification is exactly why it left a bad feeling back then, the message certainly implies that my system is not getting all the security updates it could get.

I believe RHEL/Alma/Rocky has released patches today. I have a notice about they have done so.

  • Alma released a patch ahead of upstream again, just like they did for several other critical vulnerabilities this year.