hello, how long does it typically take for package build servers to switch over to a new release? I updated to 14.3 and I can't load dmabuf.ko, drm.ko, i915kms.ko since they are built for the 14.2 kernel, presumably. (dmesg reports "unsupported file type" and "not avialable or version mismatch")

I have never used ports before, so I thought maybe I could build drm-kmod from source against my 14.3 kernel & solve the problem. I built it and installed it just fine, but I still can't load the modules :/

Hello, I've been running FreeBSD since 14.2 came out and this is my first time going through a minor version upgrade. With the laptop install the freebsd-update install (userland) after the reboot went quickly, but on the server it's taking a long time with apparently no load on the cpu, memory or anything (but the process is progressing)... is this going slow because of the current traffic from it having just released? or because the server has so

much more stuff installed?

a12l: Yeah, I've had the best luck building the port drm-XX-kmod from source

Is this thing on?

No, it's somewhere between on and off.

I hate when it is in the middle like that

Unexpected, in a chroot: bpa.st/GBAQ

mason: you mean a jail?

SponiX: No.

sorry, I'm new to FreeBSD and lost as to why you would want a regular "chroot" when you could be using a jail

I'm not likely to be able to help. So you can safely ignore me if you like lol

SponiX: No worries.

SponiX: FWIW, there are a few things you can do in a chroot that you can't do in a jail.

Not many, but some.

I wish there was a way to write file system hooks. For example, if a file is deleted, changed, or created, execute an action. That action could be something like copy the file to X or delete the file from X

I know you could do something with kqueue/kevent, but I'm looking for something more global

Upgrading from 14.2-RELEASE to 14.3-RELEASE, should I upgrade ZFS storage pools version as well?

Release Notes from 14.3 says, iirc, ZFS was updated.

I'm not sure if that is done in the freebsd-update running process. Or just an option you can do with the zfs upgrade command afterwards

ya i wanna know that too

zpool upgrades are not done by freebsd-update, you have to run zpool upgrade yourself. It is safe to do so if you don't expect to access the pool from older freebsd/zfs versions

zpool status will tell you if a zpool upgrade is needed, it may not be even though zfs was updated

Hey guys, I am going to do a post about trying FreeBSD 15/14 w/e release is available during the time of the post. Idea is to try FreeBSD for work and casual gaming. My use case is browser,email,teamviewer,anydesk,rclone,wireguard

My device is ThinkPad T16 Gen1 AMD

I will do my real work on it, there is a java based spy application too, that keeps track of my work when I am doing the shift

It works in VM but Teamviewer and AnyDesk isn't installing or executing

libminizip.so.1 is missing

What do you think? I am currently on Fedora GNU/Linux Workstation

I would also do casual gaming, I am not sure if Steam works but I will try something w/e that works

DE = Gnome, I will see if Wayland works if not then I will use x11 only. Anyways, I need x11 for the java based spy app to record my stuff in Firefox

EOF :)

for some reason I realize the freebsd community is much less hateful than the others out there

because we embrace forward looking projects like xlibre that just want to write good code, instead of getting hung up on religion and politics

that would be great tbh

ya it's what everyone wants except for the obnoxious controlfreaks

I try my best to avoid a certain "elitism" in software circle these days

ya same. i'd rather code my own thing than join up with ppl that want to force themselves on others

that is becoming increasingly more prevalent as it seems

do COVID vaccines create a "new humanoid race" according to you, demido and scottpedia? theregister.com/2021/06/11/linus_torvalds_vaccine_smackdown

here's one of the angry controlfreaks now. sorry bud, take your religion elsewhere like back to #redhat

LXGHTNXNG: what does that have to do with anything?

don't worry about him scottpedia, i've never seen him in here before no matter how pushy he is

exactly

"angry" projection *loudly draws a checkmark on a hand chalkboard*

you're mentally ill and can't help yourself, i know

"mentally ill" "can't help yourself" *another chalk checkmark*

LXGHTNXNG i just hope you're fully vaxxed and boosted

LXGHTNXNG: yeah, i thought you'd be the kind of person that still writes notes on his hand.

scottpedia: This Enrico is the person you are defending. It suggests they may have trouble with processing complex information. lkml.iu.edu/hypermail/linux/kernel/2106.1/04542.html

LXGHTNXNG why are you so obsessed with controlling the narrative? just use wayland or x11 junk. oh because you're a controlfreak that can't leave others alone

if his code sucks noone will use it, but in fact xlibre is the future and x11 and wayland are the past

Goodness, wherefore all this aggression?

'he says, looking in the mirror'

baseless accusations of aggression *squeaking chalk on a chalkboard again*

baseless... lol the delusion. consistent with your mental illness and controlfreak obsession

LXGHTNXNG: what are you trying to say? is that something to do with the vaccines?

baseless accusations of control-freak, ableism

scottpedia: The LKML post is from Enrico and contains reality-denial of a pandemic ongoing then as now.

Enrico?

notice the idealogue doesn't keep his offtopic noise to #freebsd-social. zero respect for the project or others

Maintainer of XLibre.

ya who cares get over it. ppl think differently, sorry you can't control everyone

the code is all that matters

LXGHTNXNG is angry that the xlibre maintainer doesn't believe in his coc religion

coc religion?

code of conduct

scottpedia: Read the Register article I linked about a page ago, theregister.com/2025/06/10/xlibre_new_xorg_fork

Talking derisively about DEI is a common tactic of racists in the United States, and by cultural osmosis across the Internet.

DIE is racist, you're a closet racist LXGHTNXNG

But as demido pointedly raised, this discussion belongs in #freebsd-social. scottpedia, demido, get on the move.

no amount of delusion changes that reality

you started it, you get moving

Fact check: scottpedia started it at 07:17.07 UTC.

scottpedia said "for some reason I realize the freebsd community is much less hateful than the others out there" which is true, because it shuns ppl like you that obsess over weird shit

you started talking about vaccines and a programmer's personal beliefs that are totally unrelated

you can't spin this psycho

where am I trying to spin any psychopaths? that sounds unpleasant.

you try to spread your religion and politics everywhere like a turd smearing weirdo then get pissed when others mock you for it

noone is saying you can't eat feces, just go do it elsewhere. this is a channel for freebsd talk

demido: You're not even in #freebsd-social.

and

scottpedia: Why did you send the first message here, when it's offtopic?

you think i want to socialize with losers like you LXGHTNXNG?

it's not, he was commenting on freebsd

*squeaking chalk* high school schoolyard insults

you go to #freebsd-social and take your vaccine and whatever shit with you

people like you almost killed freebsd with the goth witch coc a few years back, that had to be retracted. noone needs your poison here

LXGHTNXNG just knock it off

On the authority of you and what army?

Citation needed on the retraction, as an aside.

ah yes, because you don't care about respect or reason, you have to be forced to stop trying to force yourself on others

typical

i wonder what else you've forced on ppl before. remember when i told you last night you might be surprised what i know about you?

More empty threats, I take?

you're imagining a threat because you know you have things you desperately want to keep hidden

K then. Send it here. You won't, because you don't have it.

i'll keep you wondering

More empty threats, then.

more paranoid hysterics

seriously folks enough nobody wants this stuff here

he won't listen to reason so a mod will have to step in

I didn’t know FreeBSD drama is a thing

it's 2 nights ina  row now

it's not, he's a noob

trying to import his linux drama to fbsd

If you're right - and I doubt you are - I have nothing to hide which you don't already have. Say it on the screen, as the old troll army I battled used to say... Don't hide behind "you might be surprised what little I know about you." Cards on the table, son.

did you not just hear ppl speak up and say they don't want this shit here? noone does. drop it or fuck off

You're the one who brought up all this about having info I might want to hide.

Just say it. Pastebin. Whatever. No need to get shirty.

you both need to stop

technically correct is the best kind of correct...

I'll be keeping an eye on you.

I suggest retaining an attorney.

Whatever the case be about all this venom, vitriol and thoughts on vaccines, if you're interested in getting software packaged for FreeBSD, there's a way to do that, which is described at docs.freebsd.org/en/books/porters-handbook/book . If you encounter technical problems with the process of writing a port, don't hesitate to come right back here describing your problem. Someone who

doesn't have shell-shock syndrome from bigoted harassment will be pleased to help you.

stop obsessing

Thought you'd say that.

Further comments about /that/ might be welcome in #freebsd-social .

Anyone using FreeBSD on a recent AMD laptop?

hello, I upgraded to 14.3 and as expected, I can't load kernel modules from drm-kmod due to kernel version mismatch. I thought I could build from ports against my 14.3 kernel, so I did - and I built & installed the package - but still can't load modules due to version mismatch

how long do build servers typically stay one version behind? I will dig to figure out what I can do to build from source in the meantime... any pointers would be appreciated :)

did you also update your source tree in /usr/src?

I pulled up-to-date main branch for ports.git in /usr/ports - is that what you're referring to? (last commit I have is from 9 hours ago)

looking at the Makefile in /usr/src it looks like this may be for FreeBSD userland, kernel, and kernel module building

drm-kmod builds agasinst the source tree in /usr/src

so if you have built it against an old tree it'll still be a version mismatch

METROIDHunter_: Try steps for drm-kmod from freebsdfoundation.org/blog/how-to-u…lock-high-speed-wi-fi-on-freebsd-14

# pkg remove drm-kmod

# git clone git.FreeBSD.org/ports.git /usr/ports

# cd /usr/ports/graphics/drm-kmod

# make install

Once again, please reboot!

hiya: thanks, I will bookmark it and check it out later. I have to leave for work now

[tj]: that may be my problem. I will check that out later & make sure it is up-to-date in /usr/src

thank you both

good luck

I avoid these problems by running main and get different problems

[tj]: main? You mean dev? upcoming 15?

current, trunk, head

14 is production?

Are you using it on a laptop-PC?

I'm using it on all of my machines

I wish to try it full time on bsd-hardware.info/?probe=1fba316204

What are the odds?

Do I need to git clone this freshports.org/archivers/minizip

before doing those commands?

hiya: no, just follow the commands, it will pull all ports

cd /usr/ports/archivers/minizip/

How would this happen unless I do git clone?

Sorry I am new to this

hiya: the `git clone ... /usr/ports` you pasted earlier did this

yes

What is the git clone for freshports.org/archivers/minizip

okay I got it

hiya: it's part of the ports, you have cloned

ridcully: can I register myself as an active user/volunteer for Laptop support update program?

Yes, I found it

fwupd isn't ported yet

'Laptop Support and Usability Project'

had a good time upgrading to 14.3-RELEASE, but gosh ain't that initial $(freebsd-update upgrade) long, got anxious

All explanations of ZFS bookmarks vs snapshots seem to avoid explicitly stating the differences. From my reading, that main difference *seems* to be: with a bookmark, you can't get the version of a file from that point in time. With a snapshot, you can.

cccccbkvnbghdflkgukbtfdebkrivknltffhtrfcdnde

bless you

dvl: are bsdcan tutorials being streamed this year?

jpb: I don't know.

k, thx

tykling: thanks, and my yubikey thanks you too.

you don't see yubisneeses as much as you used to anymore

is there a way to write a make rule that creates "foo.ext" from "foo.ext.in" for any value of "ext"? ".in:" doesn't seem to work, i assume because it wants a rule for that specific suffix

When using 'freebsd-update' to upgrade to 14.3 from 14.2, where are the patched files kept before they are installed?

mns: i think somewhere in /var/db/freebsd-update, although i don't use it myself so take that with a grain of salt

ivy: how do you upgrade from one version to another?

i only run -current, and i upgrade using pkgbase: wiki.freebsd.org/PkgBase

leaving on the edge with -current?

main is pretty stable nowadays, it's not used it used to be when there was a 50/50 chance it wouldn't boot after updating

*living

On the 15.0 release page, the schedule says "15.0 EoL 30 September 2026 15.0-RELEASE no longer supported.", is that some sort of mistake considering that 15.0 hasn't even been released?

siw5ohs0: i don't think so, that sounds about right for a .0 release

Ok, "stable/15 EoL 31 December 2029", so maybe I misunderstand the schedule

15.0 will be EoL'd by 15.1

stable/15 is for every minor release in the 15.x cycle, so presumably that means support for the last 15.x version will end in 2029

while 15.0 support will end in 2026, which is earlier as you'd expect (15.1, 15.2, ... will be supported longer)

Thanks

youtube.com/watch?v=0DdnAgKLdG8 || June 2025 FreeBSD Developer Summit Day 1

vxwarlock: you stalking me?

I can't wait for FreeBSD 15 to release, so I can start tracking 16. I really like even numbers

no

December 2027

FreeBSD 16.0

I only prefer Stable versions.

understandable

-STABLE isn't necessarily stable.

I stick to the security branches.

-RELEASEpX

I am 100% STABLE runner and supporter

I need my OS to be rock solid.

rwp: sorry for the late response and that was my take on arc as well but I ran into strange memory issues with fbsd and samba.

🤔

I don’t think I ever managed to get gnome running using fbsd but I always wished I had.

🤔

(╯°□°)╯

Macer: Gnome2 from ports always (since FreeBSD 6-7 to maybe 9 - not sure) kinda "just worked" for me. Never tried Gnome3.

Gnome 3 is a whole 'nother beast.

anyone know the official irc channel for the devsummit?

mason: Ok. I never got to appreciate this new beast on ready out-of-the-box Linux distros so also never tried setting it up on FreeBSD.

I'm on 15-CURRENT. I was on 14-RELEASE prior but had weird almost random performance problems that I couldn't get tracked down

15 seems to have cured that for me

SponiX: 15-C is is a vague naming since it's 14-R with various changes made over time (so by vague I mean: there's a "different" 15-C every few hours/days), until it slowly freezes with 15-RC*, 15-BETA* and is released as 15-R. I went from 14-C straight to 15-C on laptop and workstation. With boot environments it's safe enough to use as a daily driver, as someone interested in incremental changes

or waiting to test some particular change/feature.

SponiX: "random performance problems" is not very descriptive.

I run CURRENT on everything right now for the purpose of being able to host build boxes for the KDE Plasma team ;)

regis: If I had Plex Media Server going on 14 it would tank my video playback with mpv. And even without that running, having other apps open would often do it also. Talking 10+ seconds to start a video and poor seek etc

If I was still actively trying to troubleshoot that. I would for sure try to be more descriptive

SponiX: amdgpu via LinuxKPI or fancy stuff like this or wasn't it touching the GPU driver? I actually encountered issues related to amdgpu in the opposite situation to yours. Something working flawlessly on 14-C stopped being so perfect after the switch to 15-C. To be more specific: after some random (10-50...) number of clips played one after another in VLC, following ones take 10-30 seconds to

load. I've enabled amdgpu verbose logging, looked for mem leaks, tried restarting Xorg and can't find the culprit. Unloading and loading amdgpu and drm modules wasn't always resulting in expected screen getting back on. Restarting the system always works and the specific slowly loading files are loaded blazingly fast.

Since people are discussing RELEASE, STABLE, CURRENT etc. is CURRENT stable enough to use as a daily system? Not like I'm using it for heavy loads, just as a webserver in a jail, some exploratory coding, IRC, mailserver

on an amd64 system

mns: yeah, it is. you can do so via pkgbase or you can build it yourself, just make sure you build it in release mode rather than debug mode in order to not tank your performance

mns: yes, it is. but you need to be at least somewhat aware of what's gone into -current recently and maybe don't update after a big commit until it's had some testing

mns: to put it into perspective. Netflix runs CURRENT and it does fine for them. And I daily current, so I'm going to say YES you can count on it to do what you need

also, have a dump device configured and know how to debug a kernel panic, just in case

and maybe use boot environments just in case

along with snapshots

mns: If you don't have a need for current though, and want a very good uptime record. RELEASE is honestly best in most situations

as i just mentioned (i think) i use -current everywhere on my physical and virtual servers, and i also used it on my desktop for a while, i don't currently but not because it isn't stable

again, the reason I'm on CURRENT is because the folks building/testing software on my systems need it

I don't have a need for CURRENT per se. When I used to use Debian, I was always using Debian/testing as my daily system. I was doing far more with that than I am now.

mns: Yeah, if you just like being on the edge. CURRENT is the spot

It also has things like higher versions of drm-kmod. Allowing for newer GPU hardware

I always considered RELEASE to be the equivalent of Debian/testing and CURRENT to be Debian/wip.

why would you consider -RELEASE to be Debian testing?

SponiX: I don't have a need for GPU hardware, I use my systems headless and work on trying out infrastructure as code and devops type of things

ivy: STABLE == stable, CURRENT == wip, so RELEASE == testing ... that's sort of the logic I applied. There was no rhyme or reason in particular.

mns: that seems backwards: with freebsd, -RELEASE is "more stable" than -STABLE

mns: -STABLE is the active development branch for a major release (e.g. 14.x), -RELEASE is branched from -STABLE to produce a release

s/wip/unstable/

so FreeBSD -RELEASE = Debian stable, -STABLE = testing, -CURRENT = unstable, perhaps

although that is not a perfect analogy since Debian and FreeBSD releng processes are quite different

lol

yeah they are, but it just made sense in my head when I started using FreeBSD almost a decade ago

mns: there is a portion in the documentation that covers this and gives recommendations

SponiX: ooh that would be good to read

you could perhaps think of -STABLE as equivalent to backports... e.g. 14.1-STABLE contains stuff backported from -CURRENT which will be in 14.2

I think the earlier explanation of STABLE being testing, from which the stable release called RELEASE is derived makes more sense

xkcd standards

mns: Late to answer. Yes - CURRENT is safe as a daily workstation driver. There's basically a bootloader-baked rollback mechanism which allows you to boot previous build if something is wonky. But I'm stressing that I'm talking about workstation use. I'm not Netflix to run prod on CURRENT.

regis: speaking of Netflix running current. are you familiar with their "stabweek" ?

ivy: why another one? don't we have enough of them already?

Macer, Hmm... No idea!

regis: thanks for the input. I'm most likely going to stick with RELEASE or switch to STABLE at some point. Mybe switch from 14.3-RELEASE to 14.0-STABLE

mns: that would be 14.3-STABLE - the version increments after each minor release

most people just call it 14-STABLE although technically that's not what it's called

hm, ucl++.h really could do with more documentation. currently it seems to have no documentation at all

mns: > RELEASE == testing

sounds like Debian-like naming and I don't agree with this classification when it comes to FreeBSD. RELEASE was always "the stable version" for me. I worked for a hosting company once where someone insisted on keeping the FreeBSD version "STABLE" because "it looks nice in PHP info". Absurd and ri

diculous. We were still building latests ports and/or running on latest and not quarterly packages and... due to SECURITY REASONS. Base system and userlan

d having the "STABLE" label slapped meant sh*. You use 3rd party software (ports/pkg) to serve users stuff and in this case "quarterly" is less secure tha

n "latest".

mns: I honestly don't see many reasons to be on "STABLE" apart from some weird compliance agreements with 3rd parties. You usually still have to run 3rd party software and networking software is what is exposed to external attacks. My approach is to track "latest" pkg and "HEAD" ports and be able to install something you're aware is or may be vulnerable. I'm not talking about brainless `pkg

upgrade` but simply not being limited to some quarterly frozen versions.

time to reboot now that I've upgraded to 14.3-RELEASE. BBIAB

welcome back

regis: duly noted. regardless of what the equivalent in Debian is, I've been happy with RELEASE versions and having lived on the "latest" edge, prefer these days to living on the "quarterly" edge. As I have aged, I don't care to live on the edge these days.

And as for STABLE/RELEASE - are you aware of any vulnerabilities that made RELEASE vulnerable when STABLE stayed safe? STABLE gets security patched and RELEASE immediately has the same patches. RELEASE is not less secure than STABLE. For me it's a fancy naming useful for complaince-related purposes.

mns: yep, no time for living on the edge anymore. give me stable.

stable is pretty much living on the edge nowadays, given how much is MFC'd

thedaemon: what have you been up to lately?

mns: Sure. Depends on what software you're using. Some stuff gets security updates often and these are better tracked in "latest" for obvious reasons. But I mentioned a hosting company in this case, not a laptop. Case-by-case approach. I wasn't trying to preach any particular approach. Just mentioned a hosting company running STABLE "because it looks pro" but having to run latest 3rd party ;-)

I know I'm old school... But I really like watching software build

Same

SponiX: upgraded to 14.3 over remote ssh.

thedaemon: Nice

writing some plan9 rc

I might have to tinker with that sometime

right now I'm preparing to test drm-68-kmod on current

brave lad

I know right

I hope that my lockup issues are over with.

DNS lookup?

Oh "lockup"

as in system freeze/crash

drm-kmod system freeze

yeah

I want to say so far so good..

let me check my uptime

1d 6h

But it would go for days sometimes without dying, then sometimes more than 1 time a day lol

well, from my experience any system you are on -- it is best to build the drm-kmod from source

I have been doing that. :)

Want to hear my really weird issues?

the original 14.2 drm-kmod wouldn't even boot

yes

sometimes, but not always: signal-desktop will freeze for 30 seconds to a minute or more when a message rolls in

That has happened on both 14.2-RELEASE and still happens on CURRENT for me

and the other one... I can play videos fine with "mpv". But if I go to just have it open the app without anything specific being played. That has been sometimes but not always crashing my machine all the way to a shutdown/poweroff state

All of my prior performance problems I was having on 14.2-RELEASE seem to be resolved on 15 though

mns: Sorry for adding to the conversation but I've got the best example I can give you to probe my point from security perspective. Hosting companies usually use Exim for email. From ready-to-go DirectAdmin solutions to self-baked - I've always seen Exim and never Postfix+milters. I'm not trying to be disparaging or disrespectful here (the Exim team was always helpful!) but... Exim has a somewhat

long vulns history including multiple RCE. "quarterly" ports/packages just can't cut. In my case even "latest" wasn't something I waited for so we were syncing with Exim devs to immediately fetch the hotfix version, compile it and replace out current Exim with it. Now compare this to using "quarterly" from the security perspective.

But I have to add here: Exim guys always took the time to inspect my use-case and tell me whether some announced security release may be impactful for particular setup/features.

what's signal? chat?

why are you opening mpv without a file, doesn't it just close itself?

I thought that's why I usually open vlc

thedaemon: signal-desktop is Signal chat yes

googling

and mpv without a file specified is just supposed to bring up the blank window

That problem might have existed for a long time before I noticed though. Because I never normally bring up mpv that way, there isn't much of a point to

I found that out by accident when stoned

thedaemon: well, the new drm-68-kmod seems to be working fine. And my "mpv" crash isn't happening YET

cool

lol

Spoke too soon

Burnt!

mpv crash?

Crashed right after I said that

Anybody else able to list go-anubis in their package search?

famous last words

Or are packages still "being built" for 14.3?

Full system crash from mpv

just anubis wavefunction

no go-anubis

14.3

I have not used nor tested nor can recommend it but there are plugins like mpv-file-browser to make mpv work after you invoke it w/o some media-listing argument. I've seen it used but haven't tested myself. vlc just works well for me and I never needed to switch (and the issue I mentioned with vlc happens with mpv as well so I'm pinning it to drm_kmod and/or amdgpu; never mind here)

thedaemon: now the crash did a system reboot, instead of just a poweroff. Is that improvment? LOL

I have "anime" filters in my mpv, so I use it for cartoons and vlc for everything else

SponiX: crash/reboot wouldn't come come from mpv but more likely an underlying module. Do you have some coredump to inspect?

regis: let me see if I have one. I think I have most debugging and dumping disabled. But I can always fix that

it's always DNS.. err drm-kmod

mpv initiates GPU, GPU driver locks up system. imho

thedaemon: Yeah, but then why would it only do it part of the time, and why Never when an actual video is selected and loaded?

that part seems odd

regis: oddly enough the only two .core files I have in my home are firefox and gnome-system-monitor -- and I don't remember either of those two actually crashing

guess they might be crashing hard on the systems way down though

thedaemon: My high-level understanding is that Xorg talks with amdgpu and stuff like mpv interfaces with drm-kmod. But I may be wrong on this.

I can and will be wrong XD

thedaemon: you do a good job of that ;)

XD rude!!

think I might still have you beat though

my "assumptions" often get the better of me

well, I'm gonna reboot with the dumps enabled, and then crash it again

see if I get anything worth while for troubleshooting then

do it

I just upgraded my only desktop FreeBSD system to 14.3-RELEASE; It doesn't attempt to start sddm. I suspect that is related to not finding the mouse "unable to find /dev/psm0" I do see /dev/sysmouse. Was there a name change?

wonder if I need to rebuild my kernel with the nodebug option off as well

mariuss: you most likely just need to pull the most recent /usr/ports and rebuild drm-61-kmod -- and everything will then be fine

this is a common problem, even with little point release upgrades

SponiX: that might well be necessary. It's been a while since I did this.

It's my only machine that uses upstream packages, not those built on my poudriere machine, so everything here is rather manual in this case.

So I just updated from 14.2 to 14.3, and now the system gets stuck at boot. It gets all the way to the "link state changed to UP" for all the network devices, then stays there. Not sure how to proceed. Trying to boot single user mode also gets stuck.

So I need the system sources to rebuild drm-61-kmod. How do I get the correct source tree?

how can I check if external-monitor is connected when running console-mode?

mariuss: have you pulled ports before?

mariuss: git clone -b releng/14.3-RELEASE git.freebsd.org/src.git /usr/src

mariuss: git clone -b 14.3-RELEASE git.freebsd.org/src.git /usr/src

something like that

git -C /usr/src pull

thedaemon: anubis comes up as an SMTP mail processor?

oh wait that's not ports command

I think the the last time I did this may have been before the transition to git. But I did figure it out. Currently waiting for the sources to be cloned.

mariuss needs /usr/ports updated not /usr/src

cd /usr/ports && git pull

wavefunction: yes

I know, but it needs both to be there to build the package

Otherwise: Cannot open /usr/src/sys/conf/kern.opts.mk

oh gotcha

did you try SponiX 's suggestions?

I already was pulling the sys tree when he chimed in. Thanks!

:)

Can't quite get serial console working throughout the boot process. The mobo/bios output is captured. The kernel output is as well. But the prompt for the GELI passphrase is not. I've tried using various flags in /boot.config & /boot/loader.conf, but they only seem to take effect after the passphrase is entered. It looks like zfsbootcfg can provide params to zfsboot for the next boot, but I'm not quite

certain how to set those flags (-D & -S115200)

thedaemon: I don't understand -- anubis (anubis.techaro.lol) is mislabeled in pkg?

wavefunction www/go-anubis

devnull: okay, that lines up with what I was seeing. Since I'm not doing ports, I guess I'll have to sort this one out manually.

wavefunction you can change quartery branch to latest

anyone have luck with tightvnc or rdp remote desktops on freebsd 13?

Okay, I booted into mfsBSD, mounted the pool, disabled PF, rebooted, but that did not help. Any other ideas? I'm completely stumped

freshports.org/www/go-anubis Package not present on quarterly.

zadok: Yep. I use xrdp every day with FSBD.

wavefunction: I am on quarterly I guess :)

I use rdp and vnc on 14, sorry. Don't know about 13.

The bootloader DOES say though "Loader needs to be updated", but I doubt that would help in this case

ek I have the cli renotely, but really don't see a clear path to graphics dektop

zadok: net/xrdp

what os are you trying to connect to?

freebsd

are you running vnc server?

I have tightvnc installed

tightvnc-server ?

Are you running it on the remote machine?

I am using a different vnc server

um. the remote machine is freebsd also and didn't see a client?

I am running tiger

you need a vnc server running on the remote machine

tigervnc is a good client

tigervnc-client

I hear tiger is good. Should I just install it. I really would like to use a client on my iphone to connect wherever

sorry viewer

tigervnc-viewer

ahhh

ok you can do that I guess

I cannot help with iPhone

But you have to setup the vncserver 1st

hahah... No good hacker (whitehat) would use an iphone

zadok: pkg install xrdp; service xrdp enable; service xrdp-sesman enable; edit ${LOCALBASE}/etc/xrdp/startwm.sh and choose what DE you want to run (xfce, mate, gnome, whatever); service xrdp-sesman start; service xrdp start

Use any RDP client to connect to FBSD's xrdp on port 3389

or do that ^ :)

hmmm... I have installed xrdp, but didn't know tp configure it like you mentioned

I use net/remmina as an RDP client in FBSD. Works fine.

zadok: It's a very simple rdp service to setup.

zadok: It's this easy: jeremymorgan.com/tutorials/freebsd/how-to-remote-desktop-in-freebsd

ek: For you, I am sure. I just have never done it till now

You'll need to install your preferred X11 DE, but it's otherwise very quick and simple.

Does FreeBSD have built-in address randomization everywhere now?

ek: I did follow his directions yesterday. I must be using a crappy client trying to get it on my iphone. Now, I will just use my notebook

jmabsd: yes, at least "nearly everywhere", there might be some places which are missing it

ek, and others thank you. So you get bounced if you sit here for awhile?

You're welcome. No idle bouncing here. You can idle as long as you want.

jmabsd: but in general everything is built with ASLR enabled (including PIE executables) by default

frankly I may just use the CLI. irssi is a good program

ek: thnks

Is the Kernel ASLR enabled on ARM?

zadok: I only use the RDP stuff for when I absolutely need it. Like browsing the web or viewing images/videos of some type. Otherwise, pure terminal/console is best for me.

ek: I like that. It is such a hassle otherwise

zadok: For the DE (if you choose to use RDP or VNC or whatever), I'd suggest something light like XFCE4 or Fluxbox. Quick and efficient. Easy to use.

jmabsd: i'm not sure we have KASLR at all right now, perhaps someone else knows

Oh, it was resolv.conf. Oh for god's sake... I tried every trick in the book to tell FreeBSD TO NOT FREAKING TOUCH THAT FILE yet here I am, ended up with not a DHCP-filled resolv.conf, but a blank one

I do use xfce though. I have a 10 year old Dell with both Debian and FeeBSD on it xfce4 is great. KDE and gnome are great too. Please don't hurt me - ;-)

I better go. Others want to be here, I bet. Or they should be. First time here. Thank ek.

ivy: yes, we do

for a few years now

kevans: is there no config(8) option for it? i couldn't find one when i looked

good to know, though

jmabsd: ^^

ivy: no, just a series of tunables in the kern.elf* namespace

drm-61-kmod rebuilt against 14.3-RELEASE, and all is well!

kevans: ah, does kern.elf64.aslr.enable apply to both kernel and userland? the description is a bit vague :-)

maybe I'm misremembering the disputes between ASLR and KASLR

well, i guess it's just changing the load address for each module

Okay, so... *sigh* How on earth do I tell FreeBSD to keep its hands off my resolv.conf?

zilti: echo resolv_enable=NO >>/etc/rc.conf

ivy: thanks, but nope, does not work.

strange, do you have a comment in /etc/resolv.conf saying it was created by resolvconf(8)?

(that's what resolv_enable is supposed to disable)

No, it's completely empty, only containing a newline

Guess I'm gonna try the sledgehammer "chflags schg /etc/resolv.conf"

are you using DHCP?

yes

you may need to manually configure dhclient.conf(5) to not update resolv.conf... although i find it odd that resolv_enable doesn't do this

(i don't use DHCP, so not very familiar with how this works)

zilti, you *MIGHT* find it more satisfying to set: sudo chflags schg /etc/resolv.conf

I guess resolv_enable only enables/disables the completely broken resolvconf

i thought the point of resolvconf was that dhclient would just call resolvconf, but clearly something else is going on

is there a resolvconf.conf?

CrtxReavr: Is this an LLM-generated comment? zilti used the exact same comment you later suggested as "more satisfying" with added `sudo`.

s/comment/command/

ivy: there is, yes, I created it, it has "resolvconf=NO" on the first line, followed by two lines setting name servers. None of the three lines ever worked, though

Anyways, it seems like the sledgehammer method works

regis, nope - my own (possibly) silly comment.

Actually, an old friend and FreeBSD officionado and kernel hacker, talked about how he found that more satisfying than using "official" means.

zilti, you can also: echo resolvconf=NO | sudo tee /etc/resolvconf.conf

Luckily I'm moving to fully self-hosted, which means that I'll soon be able to configure DNS in a way that just tells my servers the DNS servers I want to have

CrtxReavr: get lost with your LLM-generated stuff

zilti, um. . .

I've been using FreeBSD since v3.0. . . so. . . fuck right off.

CrtxReavr: Haha

Also, I generally add 'sudo' to commands to designate that it requires root, knowing that not everyone uses sudo.

... or root.

I do the same thing.

hmm, to add tests to a <bsd.prog.mk> program you have to include <src.opts.mk>, but this doesn't exist in /usr/share/mk... so how do you add tests to an out-of-tree program?

ek, then you must be using an LLM.

zilti: Is editing /etc/dhclient.conf to not pull DNS updates not working for you?

You're too smart for IRC!

Damn those LLM's! Got me again! ;P

dhclient.conf's syntax is. . . tedious.

'Specially when there's dhclient from OpenBSD (which is in base FreeBSD) and ISC-dhclient.

It is. But, specifying the interface and telling what to pull (and excluding dns) should leave that out entirely. At least, that's kinda the design.

ah, you can apparently just not include <src.opts.mk> and it DTRT

ek: that I did not try, because I did not yet figure out how to. So far I assumed that disabling resolvconf would suffice, but it apparently does not; and I did not really have patience left for dhclient. Especially since I use the dual-stack dhclient anyways which I bet has additional quirks

zilti: It very well may. I haven't used DHCP (for v4) in quite a while. But, I remember having the same request as you. Didn't want it touching my resolv.conf.

CrtxReavr: I'm sorry. You just sounded exactly like a bot with that resolvconf hint since I just did write above that I have that line in there.

Though. . . dhclient.conf(5) has some really good syntax examples in the EXMPLES section.

zilti, never saw that.

Anyways, I'm glad that sledgehammer worked at least, and I hope I won't have to deal with needing to disable this again. :)

Some of us have. . . wait for it. . . ACTUAL intelligence and experience.

EXAMPLES*

Key word being "Some".

>_>

Oh! That's right. I ended up using dhclient-enter-hooks to disable DNS updates to resolv.conf.

Don't pick on ivy.

Ouch!

CrtxReavr: huh?

That was more meant towards myself...

ivy, just jerk'n your chain.

ek: Ah, yes, that is something I tried as well. Also did not work with the dual-stack dhclient...

kevans,ivy: Nice. Is it on by default? What about on ARM, is it supported there

very confused about what you're meant to set in Makefile to override the owner/group here (via <bsd.tests.mk>): Var_Parse: ${INSTALL} -C ${${:UtestsFILES}TAG_ARGS} -o ${${:UtestsFILES}OWN_${:UKyuafile}} -g ${${:UtestsFILES}GRP_${:UKyuafile}} -m ${${:UtestsFILES}MODE_${:UKyuafile}} ${.ALLSRC:Ninstalldirs-*} ${${:UtestsFILES}PREFIX_${:UKyuafile}}/${${:UtestsFILES}NAME_${:UKyuafile}} (eval)

thedaemon: just a regular git pull will get CURRENT/HEAD I'm pretty sure

it seems like ${OWN_Kyuafile} should work, but it does not

you need to specify a branch for RELEASE/STABLE pulls

at least initially

ah, ${SHAREOWN}/${SHAREGRP}

Hi does freebsd have a gitlab/ github and which one do you report issues or pull request ?

found it.. Thanks

hm, i hope they aren't about to report an issue on github, you really want bugzilla for that

it is better, just because it has "bug" in the name? lol

SponiX: no, it's better because that's where bugs go. i'm not sure if you can even open bugs on github, but if you can they will be ignored

fair enough

having a central location is important time wise

i really wish this stuff was more documented...

ivy, I have not yet tried to learn the FreeBSD .mk makefile system. It's on the growing task list of things that I know will take a while to understand. I lament that it seems less well documented.