New blog post: The XDP Traffic Generator

Explaining how (as of kernel 5.18) it is possible to use #XDP to implement a high-performance traffic generator, which is also part of the xdp-tools package.

https://blog.tohojo.dk/2023/05/the-xdp-traffic-generator.html

Starting out on my three-day trip by train from Denmark to Portugal. Off to a great start with the Copenhagen-Hamburg train being half an hour late before I even got on it, and also missing a carriage so the remaining ones are packed to the brim. Yay, adventure! 😂 #trains

First blog post (after restarting blogging) is out!

In which I explore the use of kprobes in the middle of a function and reveal that I am not, in fact, very good at reading assembly code... 😅

https://blog.tohojo.dk/2023/04/netfilter-packet-drop-attribution-using-bpf.html

I'm hereby announcing my intent to take up blogging again! 😀

https://blog.tohojo.dk/2023/04/its-not-dead-its-resting.html

Bugbot status update: it's now able to monitor lore lists and start tracking threads as bugs based on an arbitrary query. E.g. you can mention "bugbot engage" in a thread and the entire thread will be converted to a bugzilla bug (if the email of the person issuing this command matches a bugzilla account with "editbugs" group membership). Any subsequent messages in the thread will be automatically added to the bug as new comments. Any comments posted on the bug via bugzilla interface will be sent to original recipients.

Now working on the other direction -- bugs added in bugzilla will be converted to mailing list threads and sent to proper maintainers (based on certain conditions, e.g. a "bugbot" flag needs to be set to "on" and the cf_subsystem custom field needs to match the corresponding MAINTAINERS entry). Should be done tomorrow, at which point I'll be looking for early testers. :)

The recordings from last week's "Understanding latency" webinar are out. Featuring yours truly for a short bit on #BPF on the second day, and lots of great speakers for the rest of the event!

https://www.understandinglatency.com/recordings-2023

I'll be speaking at this webinar on latency on March 6-8:

https://www.understandinglatency.com/

The organisers were even kind enough to make me a fancy graphics to attach to this post! 😃

Released version 1.3.0 of xdp-tools today!

This release includes three new utilities (xdp-bench, xdp-monitor and xdp-trafficgen) as well as frags support for libxdp and refcounting for XSK sockets!

Get it while it's hot off the press! 😃 #linux #XDP #BPF

https://github.com/xdp-project/xdp-tools/releases/tag/v1.3.0

Excellent coverage (as always) from @LWN of the discussions around #BPF API stability in the #Linux kernel

https://lwn.net/SubscriberLink/921088/1946095baf6289a7/

Because it has to be repeated again and again: We need to drastically reduce the number of cars to solve our problems.

Flent is featured in LWN! 😲🤩
https://lwn.net/Articles/920121/

I'm gonna keep posting this until one of you fucking boosts it

FACTS ABOUT MASTODON

If you are curious about leaving the cooked turkey site and going to the elephant site, here are some important tips:

1. It sucks. But then, so does every site.

2. You can still shitpost. Take great glee.

3. Picking your server instance is super important.

Ideally you should start at a large instance, and leave because it's full of white suburban NIMBY reactionaries who joined in November because they were promised 0 uncomfortable experiences in their lives and lash out whenever this turns out not to be the case.

Then you move to a smaller server where suddenly you can't talk to your friends because the admin of your instance is feuding with the admin of their instance. Then you wait a month before you can move again.

In this regard, the feudal structure of Mastodon instances is very like early 2000s message boards, whenever the admin got drunk and deleted the site.

4. You can work around the feudalism by running Mastodon yourself. It's the size of a mastodon and costs a fortune.

You can run Pleroma, which is smaller, and is also favoured by Nazis by unfortunate historical accident. Pleroma is perfectly good software that fulfils a need for something smaller than Mastodon, but also the devs are definitely not Nazis but are the other ten guys at the table.

There was a hilarious moment where the guy behind Spinster was so obnoxious he got kicked out of Pleroma and started his own fork called Soapbox/Rebased. He is now known as Soapbox Terf.

The nice people went to Pleroma fork Akkoma, which Soapbox Terf calls the "tr***y server", a review I understand they were delighted by. Try that.

There's also Misskey, which is a bit weird and Japanese, and supports cat ears right there in the protocol.

5. Any bozo who complains about your posts with assertions about the Fediverse that assume it all runs on the rules of mastodon.social is one of the suburban NIMBYs and invariably joined in November. Block and don't look back.

6. If anyone annoys you about your posting, you can improve their feed for them by blocking them from ever seeing your posts. The blocking tools are marvellous.

7. There are NO QUOTE TWEETS on Mastodon and anyone who wants QUOTE TWEETS is an invader, pollutant and corrupting influence despoiling the suburban vistas of Mastodon who only wants quote tweets so they can wreak EVIL.

So quote-tweeting is well supported in Akkoma and Misskey (and forks thereof), is in the Treehouse fork of Mastodon, and will be coming to more Fediverse software soon.

8. In Mastodon, Eugen Rochko has achieved the creation of something greater than himself. And he will *never forgive it*.

9. The Fediverse interprets Website Boy as damage and routes around him.

10. Mastodon is yet another demonstration that worse is better. So come onto Mastodon, and *be* that worse.
====
EDIT: this post is attracting some very dumb reply guys. Consider *not* posting debate club fatuity.

Are you someone who has a simple fix for something in the Linux kernel, but haven't bothered submitting a patch because you find the email workflow daunting? Can be a bugfix, spelling correction, anything at all.

If so, I want to talk to you! I have tools that simplify this process quite a bit, but I need more feedback from people who aren't long-term Linux maintainers.

Please reach out, either via here or by emailing mricon@kernel.org.

Computer science pioneer and United States Navy rear admiral Grace Hopper was born #OTD in 1906.

As far as I’m aware, she is the only person who has both a supercomputer and a US Navy destroyer named after her.

Image: Computer History Museum

Shared by my Daughter
"I need privacy, not because my actions are questionable, but because your judgement and intentions are"

In response to "if you have nothing to hide you have nothing to fear"

Back in November I did a presentation at the Driving IT conference in Copenhagen, where I tried to make the case that we should all focus more on latency and less on throughput when building systems and applications.

The talk is really channelling Stuart Cheshire's classic rant (from 1996!), from which I also shamelessly stole the title: "It's the Latency, Stupid!"

Anyway, the video is now finally online, if you'd care to have a listen: https://videos.ida.dk/id/1_47l9o3qn

Okay, so let me tell you about my doorbell, from a #networking perspective.

When you push the button by the door, it sends a message over the #zigbee wireless mesh network in my house. It probably goes through a few hops, getting relayed along the way by the various Zigbee light switches and "smart outlets" I have.

Once it makes it to my utility closet, it's received by a Zigbee-to-USB dongle, through a USB hub (a simple tree network) plugged into an SFF PC. From there, it gets fed into zigbee2mqtt, which, as the name implies, publishes it to my local #mqtt broker.

The mqtt broker is in the small #kubernetes cluster of #raspberrypi nodes I run in my utility closet. To get in (via a couple of #ethernet switch hops), it goes through #metallb, which is basically a proxy-ARP type service that advertises the IP address for the mqtt endpoint to the rest of my network, then passes the traffic to the appropriate container via a #linux veth device.

I have #HomeAssistant, running in the same Kubernetes cluster, subscribed to these events. Within Kubernetes, the message goes through the CNI plugin that I use, #flannel. If the message has to pass between hosts, Flannel encapsulates it in VXLAN, so that it can be directed to the correct veth on the destination host.

Because I like #NodeRed for automation tasks more than HomeAssistant, your press of the doorbell takes another hop within the Kubernetes cluster (via a REST call) so that NodeRed can decide whether it's within the time of day I want the doorbell to ring, etc. If we're all good, NodeRed publishes an mqtt message (more VXLANs, veths, etc.)

(Oh and it also sends a notification to my phone, which means another trip through the HomeAssistant container, and leaving my home network involves another soup of acronyms including VLANs, PoE, QoS, PPPoE, NAT or IPv6, DoH, and GPON. And maybe it goes over 5G depending on where my phone is.)

Of course something's got to actually make the "ding dong" sound, and that's another Raspberry Pi that sits on top of my grandmother clock. So to get *there* the message hops through a couple Ethernet switches and my home WiFi, where it gets received by a little custom daemon I wrote that plays the sound via an attached #HiFiBerry board. Oh but wait! We're not quite done with networking, because the sound gets played through PulseAudio, which is done through a UNIX domain socket.

SO ANYWAY, that's why my doorbell rarely works and why you've been standing outside in the snow for five minutes.

Without internet connectivity since yesterday afternoon. Looks like my upstream ISP had a largish equipment failure which took them all night to recover from. This morning I have a physical connection (I get LLDP packets from a device at the other end) but my BGP peer is MIA.

Spent an hour in the helpdesk queue and all they could tell me was "we'll put it in the queue for our technician". Ugh.

Do I know anyone who can provide me with a temporary BGP session over a tunnel (and reannounce my addresses to their upstream)?

Did an interview for a Danish newspaper about #bufferbloat and internet connectivity. The article turned out pretty well, I think! It's in Danish and unfortunately paywalled, but just in case I know anyone with an account, here's the link (article by @LauridsHovgard):

https://ing.dk/artikel/daarlig-koekultur-din-router-flaskehals-din-netforbindelse-263272