Laoise Nic Giolla Uidhir

IPv6 Pro Tip: to make network configuration examples easier to understand and more applicable to your own network, use the IPv6 documentation prefix (2001:db8::/32) for your local network, and deploy NAT66 for Internet access.
#ipv6

2026-08-12 02:43:52 UTC

Grafana has helpfully invented a new unit: "mp/s", which is apparently millipackets per second.
i might've thought it would be easier to display this as "0.6 p/s", but let's see if millipackets catch on.
https://media.bsd.cafe/bsdmmedia01/media_attachments/files/117/080/230/471/360/586/original/93fe2a02cee72727.png

2026-08-11 02:05:48 UTC

- reply

how can you even afford 64kB of memory at current prices

2026-08-05 18:01:05 UTC

it's a shame they couldn't use an AI to condense this into a single email, but perhaps they can't afford enough tokens for that.
https://media.bsd.cafe/bsdmmedia01/media_attachments/files/117/044/205/333/598/798/original/4ea822c95a8d5be3.png

2026-07-23 14:08:29 UTC

i wonder why we build mandoc as a bootstrap tool. is this for catpages? (i don't think those exist anymore...)
#freebsd

2026-07-01 14:04:37 UTC

- reply

it already has a check to see if pkgbase is in use, in which case it will abort. but you may still need to use it again, e.g. if you have non-pkgbase jails -- which is why we'll have to keep shipping it in 16.0 even though we won't provide updates for 16.0 via freebsd-update.

2026-07-01 13:58:56 UTC

- reply

i think this is unlikely to happen since i don't see any real benefit, but i could entertain arguments in favour

2026-06-17 17:13:56 UTC

- reply

do you happen to know whether that works in clients? or do they need special support or something? (i’m wondering about Ivory specifically)

2026-06-17 14:41:00 UTC

@npub1ef5…gcpc if i move to the Snac instance, will i be able to include basic formatting (Markdown or w/e) in my posts? not being able to do this on Mastodon is super annoying.

2026-06-04 19:39:12 UTC

- reply

do we know when this is coming to a stable release?

2026-05-06 20:09:01 UTC

i see people fleeing GitHub for Codeberg, which is understandable, but Codeberg is only one good DDoS away from being down for a week.
the fix for this isn't to move from one provider to another, it's for people to host their own projects, or ask a friend to host it. Git has a built-in email workflow that works great, you do not need "pull requests" to accept contributions.

2026-05-06 16:28:49 UTC

- reply

could you please show the output of
% pkg repos
and
% pkg query -e '%o ~ base/*' '%n %v'

2026-04-22 20:49:46 UTC

- reply

hm, i'm not 100% sure about this... don't most MMO raids have mechanics which are both difficult to figure out, and difficult to do?

2026-04-16 16:11:12 UTC

- reply

this sounds like it'll work about as well as anycast 6to4 did, tbh. (i.e., not very.)

2026-04-09 12:32:17 UTC

- reply

although, i just noticed it's only displaying 164 threads, hm.

2026-04-09 12:30:38 UTC

- reply

the way btop(1) visualises this is actually quite nice... i usually prefer htop(1) but it seems to give up entirely here and not display per-CPU stats at all.
https://media.bsd.cafe/bsdmmedia01/media_attachments/files/116/374/754/390/923/808/original/2ba8e0a4bf6c864d.png

2026-04-09 12:12:54 UTC

what type of Grafana graph would you use to visualise CPU usage on a system with 176 CPU threads? i usually like heatmap, but i don't think that's going to work unlses the graph is three miles tall.

2026-03-27 21:28:27 UTC

- reply

this includes anyone in the UK who voted for Labour. this is your fault. fuck y'all.

2026-03-27 21:25:52 UTC

as of yesterday, i cannot use my own phone, which i own and paid for in cash, to visit websites on the Internet or to send and receive text messages.
unless i "verify my age" with Apple, i am considered a "teen" and do not own my own device.
anyone who voted for this nonsense: fuck you.

2026-03-25 09:24:04 UTC

@npub1sv0…rq9x any news on a cheaper data plan for SIP2SIM?

2025-12-25 21:40:49 UTC

a user is arguing with me that setting securelevel on FreeBSD prevents an attacker gaining persistence, which... is not even remotely true, but i'm sure we still ship documentation that suggests it is.
i don't think we should remove securelevel, since the basic functionality it provides is fine, but perhaps it needs to be split into a few different sysctls. like one for enforcing file flags, one for device access, one for memory access... the firewall stuff can probably just be ditched.
is there a reason we didn't already do that? it seems like a fairly obvious thing to do.

2025-12-25 21:38:36 UTC

- reply

"With securelevel, they are locked out of the kernel.”
except that /boot/kernel/kernel isn't schg, so they can just replace the kernel itself and wait for you to reboot. but then, if they can replace /boot/kernel/kernel, they already have root... so what is securelevel protecting against?
if the answer is persistence - did you modify *every* executable that root can run (including everything installed from ports/packages) to ensure it can't be replaced? did you make atrun and crontab read-only (thus breaking at(1) for everyone on the system)?
if you're put the required effort into building a system which is actually secure enough for schg/securelevel to make a difference, perhaps it can marginally increases security, but your original toot makes it sound like simply setting securelevel increases security, which it doesn’t. and i think it's a shame we trick users into thinking it does by calling it 'securelevel'. (it should be called immutable_level, or something, or more usefully, split into multiple sysctls.)

2025-12-25 21:22:05 UTC

- reply

yes, i know what it does — i’m just saying it doesn’t actually increase security.
for example, did you ensure every executable on the system that root might execute is schg? if not, a few random executables being schg doesn’t help.
firewall modification is already restricted to root. what sort of attack are you protecting against where someone has root but needs to modify the firewall rules? (and are you sure they can’t just edit some file in /etc and wait for you to reboot?)
the problem with securelevel is it does a bunch of things that *look* like they should increase security, but in practice, aren’t useful against any real world attacks.

2025-12-25 21:10:48 UTC

- reply

securelevel doesn’t really do anything to increase security. it’s a relic of early 1990s security practice and should probably just be removed from the system, or at least renamed.

2025-10-06 11:20:50 UTC

i previously didn't dislike Lennart Poettering, but it's funny to see him claim that freedesktop.org is the sole arbiter of where Unix applications are allowed to create files.

2025-09-28 00:46:14 UTC

does anyone have experience with Patreon? is there a reason to avoid them? do they refuse payouts or close your account for absurd reasons (like PayPal)?

2025-05-16 18:31:19 UTC

- reply

you shouldn't usually need net.link.bridge.inherit_mac here... was there a particular problem you ran into that required that?
(it's not wrong if you need it, i'm just curious)

2025-04-30 07:14:39 UTC

merge freebsd traceroute6 into traceroute, take 4: failed.
i think to get this done, i need to initially propose a change to do it in a bad way, then make it not bad in followup commits, instead of just committing the not-bad version to begin with.
which i find a little frustrating, but let it never be said that i won't commit bad code in the name of pragmatism.
#freebsd #ipv6

2025-03-15 12:40:49 UTC

- reply

that seems very warm for something running OPNsense ... assuming you're not running 10Gb/s traffic constantly or something like that.
did they forget to set up CPU frequency scaling?