frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

OpenBSD: PF queues break the 4 Gbps barrier

https://undeadly.org/cgi?action=article;sid=20260319125859
63•defrost•1h ago

Comments

bell-cot•1h ago
"Values up to 999G are supported, more than enough for interfaces today and the future." - Article

"When we set the upper limit of PC-DOS at 640K, we thought nobody would ever need that much memory." - Bill Gates

WhyNotHugo•1h ago
Honestly, I'm really curious about this number. 10bits is 1024, so why 999G specifically?
elevation•1h ago
Looks like an arbitrary validation cap. By the time we're maxing out the 64-bit underlying representation we probably won't be using Ethernet any more.
palmotea•1h ago
> By the time we're maxing out the 64-bit underlying representation we probably won't be using Ethernet any more.

We will be using Ethernet until the heat death of the universe, if we survive that long.

bell-cot•58m ago
https://en.wikipedia.org/wiki/Ethernet#History (& following sections)

Calling something "Ethernet" amounts to a promise that:

- From far enough up the OSI sandwich*, you can pretend that it's a magically-faster version of old-fashioned Ethernet

- It sticks to broadly accepted standards, so you won't get bitten by cutting-edge or proprietary surprises

*https://en.wikipedia.org/wiki/OSI_model

abound•1h ago
Looking at the patch itself (linked in the article), the description has this:

> We now support configuring bandwidth up to ~1 Tbps (overflow in m2sm at m > 2^40).

So I think that's it, 2^40 is ~1.099 trillion

throw0101d•1h ago
> "Values up to 999G are supported, more than enough for interfaces today and the future." - Article

Especially given that IEEE 802.3dj is working on 1.6T / 1600G, and is expected to publish the final spec in Summer/Autumn 2026:

* https://en.wikipedia.org/wiki/Terabit_Ethernet

Currently these interfaces are only on switches, but there are already NICs at 800G (P1800GO, Thor Ultra, ConnectX-8/9), so if you LACP/LAGG two together your bond is at 1600G.

arsome•1h ago
If you're moving those kind of speeds you're probably not doing packet filtering in software.
throw0101d•43m ago
But you may be using Unix-y software to manage the interfaces and do offload programming:

* https://en.wikipedia.org/wiki/Vector_Packet_Processing

* https://www.youtube.com/watch?v=ptm9h-Lf0gg ("VPP: A 1Tbps+ router with a single IPv4 address")

* https://en.wikipedia.org/wiki/Cumulus_Networks

himata4113•31m ago
I use VPP and handle bonded speeds of 200gbit. Not that far fetched to also do this at 1000gbit.
bitfilped•52m ago
Yes, we're already running 800G networks, so this phrasing seems really silly to me.
rayiner•1h ago
Can pf actually shape at speeds above 4 gbps?
ralferoo•1h ago
In the days when even cheap consumer hardware ships with 2.5G ports, this number seems weirdly low. Does this mean that basically nobody is currently using OpenBSD in the datacentre or anywhere that might be expecting to handle 10G or higher per port, or is it just filtering that's an issue?

I'm not surprised that the issue exists as even 10 years ago these speeds were uncommon outside of the datacentre, I'm just surprised that nobody has felt a pressing enough need to fix this earlier in the previous few years.

citrin_ru•59m ago
AFAIK performance is not a priority for OpenBSD project - security is (and other related qualities like code which is easy to understand and maintain). FreeBSD (at least when I followed it several years ago) had better performance both for ipfw and its own PF fork (not fully compatible with OpenBSD one).
traceroute66•29m ago
> AFAIK performance is not a priority for OpenBSD project - security is

TBF that was the case historically, but they have absolutely been putting in an effort into performance in their more recent releases.

Lots of stuff that used to be simply horrific on OpenBSD, such as multi-peer BGP full-table refreshes is SIGNIFICANTLY better in the last couple of years.

Clearly still not as good as FreeBSD, but compared to what it was...

toast0•51m ago
> Does this mean that basically nobody is currently using OpenBSD in the datacentre or anywhere that might be expecting to handle 10G or higher per port, or is it just filtering that's an issue?

This looks like it only affects bandwidth limiting. I suspect it's pretty niche to use OpenBSD as a traffic shaper at 10G+, and if you did, I'd imagine most of the queue limits would tend toward significantly less than 4G.

IcePic•49m ago
One thing could also be that by the time you have 10GE uplinks, shaping is not as important.

When we had 512kbit links, prioritizing VOIP would be a thing, and for asymmetric links like 128/512kbit it was prudent to prioritize small packets (ssh) and tcp ACKs on the outgoing link or the downloads would suffer, but when you have 5-10-25GE, not being able to stick an ACK packet in the queue is perhaps not the main issue.

traceroute66•34m ago
> Does this mean that basically nobody is currently using OpenBSD in the datacentre or anywhere

Half the problem is lack of proper drivers. I love OpenBSD but all the fibre stuff is just a bit half-baked.

For a long time OpenBSD didn't even have DOM (light-level monitoring etc.) exposed in its 1g fibre drivers. Stuff like that automatically kills off OpenBSD as a choice for datacentres where DOM stats are a non-negotiable hard requirement as they are so critical to troubleshooting.

OpenBSD finally introduced DOM stats for SFP somewhere around 2020–2021, but it doesn't always work, it depends if you have the right magic combination of SFP and card manufacturer. Whilst on FreeBSD it Just Works (TM).

And then overall, for higher speed optics, FreeBSD simply remains lightyears ahead (forgive the pun !). For example, Decisio make nice little router boxes with 10g SFP+ on them, FreeBSD has the drivers out-of-the-box, OpenBSD doesn't. And that's only an SFP+ example, its basically rolling-tumbleweed in a desert territory if you start venturing up to QSFP etc. ...

Someone•32m ago
The article is about allowing bandwidth restrictions in bytes/second that are larger than 2³²-1, not about how fast pf can filter packets.

I guess few people with faster ports felt the need to limit bandwidth for a service to something that’s that large.

FTA:

“OpenBSD's PF packet filter has long supported HFSC traffic shaping with the queue rules in pf.conf(5). However, an internal 32-bit limitation in the HFSC service curve structure (struct hfsc_sc) meant that bandwidth values were silently capped at approximately 4.29 Gbps, ” the maximum value of a u_int ".

With 10G, 25G, and 100G network interfaces now commonplace, OpenBSD devs making huge progress unlocking the kernel for SMP, and adding drivers for cards supporting some of these speeds, this limitation started to get in the way. Configuring bandwidth 10G on a queue would silently wrap around, producing incorrect and unpredictable scheduling behaviour.

A new patch widens the bandwidth fields in the kernel's HFSC scheduler from 32-bit to 64-bit integers, removing this bottleneck entirely.”

MasterSelects: A GPU-First Video Editor Built on WebGPU

https://www.webgpu.com/showcase/masterselects-webgpu-video-editor/
1•FarhadG•40s ago•0 comments

Semiconductor enthusiast builds DIY 'class 100 cleanroom' in his garden shed

https://www.tomshardware.com/tech-industry/semiconductors/ambitious-semiconductor-enthusiast-buil...
1•mywacaday•2m ago•0 comments

DoorDash's New Paid Tasks Turn Couriers into AI and Robot Trainers

https://www.bloomberg.com/news/articles/2026-03-19/doordash-s-new-paid-tasks-turn-couriers-into-a...
1•thm•3m ago•0 comments

OpenClaw and the China-US AI Agent Divide

https://hellochinatech.com/p/openclaw-china-us-agent-divide
2•yarapavan•5m ago•0 comments

Humanoid robot gets surprisingly good at tennis

https://newatlas.com/ai-humanoids/humanoid-robot-tennis-latent-unitree-g1/
2•Brajeshwar•6m ago•0 comments

Reference Code for the Slug Algorithm

https://github.com/EricLengyel/Slug
2•bwidlar•6m ago•0 comments

Special Commitment Center

https://en.wikipedia.org/wiki/Special_Commitment_Center
1•djoldman•7m ago•0 comments

Social media makes people unhappy – World Happiness Report

https://www.dw.com/en/social-media-makes-people-unhappy-world-happiness-report/a-76422753
1•giuliomagnifico•7m ago•0 comments

Millwright: Smarter Tool Selection from Agent Experience

https://minor.gripe/posts/2026-03-13-millwright_smarter_tool_selection_with_adaptive_toolsheds/
2•ai_critic•8m ago•0 comments

Nominal Connect: Shipping Realtime Desktop Software with Rust, Bevy, and Egui

https://nominal.io/blog/nominal-connect-shipping-realtime-desktop-software-with-rust-bevy-and-egui
1•ibobev•9m ago•0 comments

Show HN: Yansu – agent that builds your bespoke app before you ask

https://yansu.app/
1•yubozhao•9m ago•0 comments

Who's really behind the Iran war? [video]

https://www.youtube.com/watch?v=YBGs2XSP58A
1•teleforce•9m ago•0 comments

Moon (2024)

https://ciechanow.ski/moon/
1•darkwater•10m ago•0 comments

Erlang's not about lightweight processes and message passing

https://stevana.github.io/erlangs_not_about_lightweight_processes_and_message_passing.html#correc...
2•Munksgaard•10m ago•0 comments

Show HN: Xtream2M3U – Convert, Edit and Play IPTV Playlist (Browser-Based)

https://www.xtream2m3u.com/
2•onion92•12m ago•0 comments

How I Learned to Think and Work Like a Startup Founder

https://lasans.blog/articles/thoughts/how-i-learned-to-think-and-work-like-a-startup-founder/
1•lasgawe•13m ago•0 comments

Downleveled

https://k2xl.substack.com/p/downleveled
1•k2xl•14m ago•0 comments

Terraform Cloud Drift Detection: How It Works and Setup

https://spacelift.io/blog/terraform-cloud-drift-detection
1•kat-w•15m ago•0 comments

Show HN: Juttu – Open-source Disqus alternative where comments are Bluesky posts

https://github.com/matteomarjanovic/juttu
1•matteomrj•16m ago•0 comments

Opera GX, the browser for gamers, has landed on Linux

https://www.neowin.net/news/opera-gx-the-browser-for-gamers-has-finally-landed-on-linux/
1•bundie•16m ago•0 comments

Show HN: Search dashcam footage by describing what happened

https://github.com/ssrajadh/sentrysearch
1•sohamrj•16m ago•0 comments

Check your websites AI visibility directly from the browser in 1 click

https://chromewebstore.google.com/detail/llm-check-ai-readiness-vi/oalgfkbijifcmkbibinfoppkmhckjjcj
1•sansreal•17m ago•0 comments

Use subagents and custom agents in Codex

https://simonwillison.net/2026/Mar/16/codex-subagents/
1•gmays•19m ago•0 comments

Kash Patel Admits the FBI Is Buying Private Data on Americans

https://gizmodo.com/kash-patel-admits-the-fbi-is-buying-private-data-on-americans-2000735317
3•sebastian_z•19m ago•0 comments

Show HN: Backlit – Lit Web Component SSR for Drupal via Go and WASM, No Node.js

https://bennypowers.dev/posts/drupal-lit-ssr-wasm/
2•bennypowers•20m ago•0 comments

Majority of UK power in the last week was wind

https://grid.iamkate.com/
1•gregorvand•21m ago•2 comments

Dial9: A Flight Recorder for Tokio

https://tokio.rs/blog/2026-03-18-dial9
2•emschwartz•22m ago•0 comments

Instrumenting Rustls with eBPF

https://coroot.com/blog/instrumenting-rust-tls-with-ebpf/
1•DebianDude•24m ago•0 comments

Show HN: Observarium, a simple exception tracking library for Java

https://github.com/hephaistos-io/observarium
1•RicDan•24m ago•0 comments

The case for and against a 5th fundamental force of nature

https://bigthink.com/starts-with-a-bang/5th-fundamental-force/
1•Brajeshwar•24m ago•0 comments