the latest drm-kmod from github.com/freebsd, does it correspond to the linux kernel 5.10?

or to newer linux kernels?

Oleg: Considering the name of the port is drm-510-kmod, it's probably fairly likely.

Soni: that hostname!!!

thanks =^-^=

why do you need a hostname here?

anyone got a guide for how to _manually_ set up zfs "zroot" on /mnt during installation in a way that actualyl works? The "guided zfs on root" thing doesn't guide much and insists on using entire disk...

I tried to do it manually, but whatever I do - set mountpoints etc... on boot after install, it _only_ mounts zroot/ROOT/default, and nothing else

kolla yes

kolla 1 min, I think I've blogged about it, lemme check

kolla nah, I did not blog about it, but I know where you can find it. 1 minute :-

kolla here, check this out: cgit.freebsd.org/src/tree/usr.sbin/bsdinstall/scripts/zfsboot#n146

Title: zfsboot « scripts « bsdinstall « usr.sbin - src - FreeBSD source tree

kolla: Also maybe worth reading: wiki.freebsd.org/MasonLoringBliss/ZFSandGELIbyHAND

Title: MasonLoringBliss/ZFSandGELIbyHAND - FreeBSD Wiki

mason: do you know for how long we have to wait until freebsd drm-kmod starts corresponding to version 5.13 of the linux kernel? I ran into the same issue this guy did: forums.gentoo.org/viewtopic-t-11362…stdays-0-postorder-asc-start-0.html .

Title: Gentoo Forums :: View topic - [SOLVED] Intel graphics, no signal once kernel loads i915

Oleg: The efnet #freebsd-xorg folks might know.

I don't unfortunately.

antranigv: thanks!

I really find it awkward that the installer insists on using entire disks and doesn't even let the user decide how much of the disks to use

kolla: The installer gives you a shell to partition if you want.

Can't get much more flexible than that.

yes, but that requires skills beyond normal to get it working

that is why I asked, because no matter how nice and tiny and finished I made the zpool manually, after install, on bootup, it _only_ mount the zroot/ROOT/default, and no other zfs

It's not that bad to do

well, clearly the "auto" thing does some magic I didn't do, despite 20-30 lines of manually defining it all

I set all mountpoints correctly, I zpool set bootfs, I could export and import zroot manually and all looked nice

but on reboot... bah meh, no

an empty /usr doesn't get the system far :)

Did you make sure it was all in place before letting the installer populate it?

Or was everything populated, and just not mounting on boot?

This might be cargo-cult of me, but I think you still need vfs.root.mountfrom=

in loader.conf

yes, everything is populated correctly... I boot from installer USB stick, I can import the zpool and all is there, all filesystems, nicely populated

if I let the installed have a go with entire disks, that is default... I don't get any vfs.root.mountfrom= in loader.conf

it would be ok to let it have all disks if I could shrink the pool afterwards... but I believe that isn't an option

Using GPT partitions is the way to go, yeah.

if installer could create a rather minimal zroot pool, it wouldn't be hard to expand it afterwards though...

yes, have to use GPT, 8 x 4TB disks

odd thing is.. I have this other identical system that I installed 3 years ago or so, and I don't recall having this struggle

kolla: I'd recommend taking notes on how you're doing the partitioning, and then having my guide there to make sure as you go through that you're not missing anything.

wellwell, I will try to follow the last guide there, thanks again :)

mason: yes

Well. Do your own thing but make sure if the guide is doing something that you're either doing an equivalent or that you explicitly don't need to.

Good luck anyway. Hitting the sack here.

thanks, and... good idea to also inclide EFI partitions, my systems are indeed legacy (though 64bit), but who knows, maybe I want to move it over to EFI system later

good night :)

also... really wish the installer iso was prepared for running sshd, which today isn't any "out of the box" experience, with read-only /etc/ssh and /etc/passwd

</rant mode>

do old freebsd cd releases include binary packages or distfiles?

dksnd: on other cd's or dvds there were packages... depends how old

by old we surely mean 4.6?

around 2005-2006

Title: Index of /pub/FreeBSD-Archive/old-releases/i386/ISO-IMAGES/

it's all here

i don't think anything included ports distfiles... so you need to search for that somewhere else

but those are pretty useless without the ports/packages

you have packages

but old source is pita

i bet you can get old kde running

won't put money on that though

what's the use case of that?

a retrocomputing experiment

hmm

i have no idea what packages are in 6.0r cd2

that's supposedly full of them

never used that feature

but it comes handy

i also have too many ports distfiles...

somehow i've stored them

unsure if there is mirror of them

seems way too much data

no problem with base though

every installer has copy

i think you could install *some* things from selected packages included in installer medias

funny how windows or dos might be easier to retro-setup

eh

i could install 4.6 for fun

4.6 because that's where i started at

e?

e?

so I installed freebsd 5.3, but it refuses to boot lol

xd

it's UNIX & GNU/Linux useful things written in a chetsheet style using VIM markers in a text file

if you're interested in helping out, you're more than welcome to

As you're so welcome to use it

Title: Mubarak Alrashidi / NixGems · GitLab

it prints a menu "F2 FreeBSD" and nothing happens then

I can use a livecd to manually load its kernel and provide the root path interactively

does it mean F2 key will boot the corresponding partition?

I used boot0cfg on the entire hdd and disklabel on the freebsd slice

interesting, when I use an external loader to boot to the partition directly, it shows a boot menu and a devil ascii art

it seem to mount the root automatically from that menu

àpparently "packet,noupdate" options did the trick

dksnd: which software has those options?

boot0cfg

sh doesn't like non-ascii characters it seems

hello. i just did a pkg-upgrade on my freebsd 13.1 laptop with kde, and after reboot, i no longer get sddm login screen; just black screen with cursor

using pkg quarterly. was working fine before this update. anyone else facing similar issue? any known fix?

sddm log does not show any (obvious) errors

what does Xorg.log say

i don't see any errors there either. i think Xorg is fine, because i the rendered cursor

i think it's something specific to kde plasma and/or sddm

ldd `which sddm`

returns bunch of suff...anything specifc i should look at?

Missing

doesn't show anything missing

do you have the port cloned ?

ports*

i do; probably need to update, but yes

compile it by hand see if that fixes it

hmm, ok. just sddm or all of plasma?

as long as you dont change the dewfault options you wont have issue with pkg

actually as root service sddm stop

yea, i don't think i ever needed to touch sddm options on this laptop

echo "startplasma" > ~/.xinitrc && startx

see if that works

think I got the start script wrong but start<tab> will give you the correct name

if that fails then yeah recompile the full kde*

and note it will take "a while"

CorvetteZR1 I had a few problems like that

I don't think that the problem is with sddm, I can ssh in and I see plasma etc starting

If I start killing things in reverse order of pid then it will restart services and I get KDE up

This seems intermittent, and I haven't yet started taking notes and recording pstacks

sorry, had to take a call...will try that now cpet

paulf, cool, i'll play with it and see

cpet, looks like it's same behaviour with startplasma and startx...black screen, only cursor

so it's probably not a sddm thing then, looks like plasma specific

looking for mpts / 0mp

I wonder if my ZNC has lost messages

I also tried renaming ~/.config to ~/.config.bak and rebooting which had no problems

but I didn't want to lose all my settings so that was just for testing

actually, when i right click, i get the kde right click menu

so it's just the actual desktop that's not being rendered

it can even draw the kde settings window, but there is nothing in it

kolla: Hey, don't know if you saw the link, but while it's maybe vaguely tedious, you can get ssh going from the install environment: wiki.freebsd.org/MasonLoringBliss/InstallEnvironment

Title: MasonLoringBliss/InstallEnvironment - FreeBSD Wiki

Would have mentioned this last night if I knew you wanted it, but I'd gone to bed.

does freebsd have scrolling in tty? shift+ pageup not working for me...

yeah thats not the key combo

and this keyboard does not appear to have a scroll lock

ctrl alt pgup I think

no, that's not doing it

you will need a scroll lock :)

ah, got it...Fn K

i'm seeing bunch of TypeErrors from main.qml

TypeError: Connect read property ;configuration' of null

something is definitely messed up with plasma

is there a plasma5 meta port?

is plasma5-plasma it?

ah, it sure is :)

ok, i think it's working now

i saw some OpenGL errors when starting with startx. i reinstalled the nvidia-driver-390 port (because i have an old fermi). reboot and now it's loading

thanks for your help cpet and paulf !

I'm interested in dualbooting freebsd alongside openbsd as a gaming platform (steam/wine), what pitfalls can I expect, and should I just go with linux instead?

why don't you try it and find out

because it's a bunch of work and if it sucks anyways it'd be nice to get some opinions in advance

ssm_ well, most FreeBSD people dualboot with Windows/Linux, most OpenBSD people don't dualboot at all, this will be a very hard research :D

I used to build a lot of multi-boot images - often with 3-4 OSes each.

The age of virtualization sorta killed off the need for much of that.

Though. . . dual-booting FreeBSD & OpenBSD should be pretty trivial.

Start with FreeBSD, be sure to install the FreeBSD boot manager (and leave un-partitioned (sliced) space on your first drive for OpenBSD.

wasn't worried about the difficulty of dual booting, just wanted to know how well wine (&| the linuxulator w/steam) works comparatively to linux

When you install OpenBSD, but sure not to let it touch your MBR.

And you shoudl be golden.

I'm using efi, so I don't think there should be any problems

I love FreeBSD, but if I wanted a *nix OS for Steam, I'd go with ArchLinux, since it's what Steam Decks are built with. . . best gurantee of support & compatability.

That said, there's plenty of happy people doing it with RHAT & Debian derivatives.

Steam on Debian is reliable. The actual Linux application you install is still based on some old Ubuntu.

I don't want to know about debian or arch, I want to know how well freebsd supports it

I can't stand linux, if I just wanted the easy solution I wouldn't have even asked in the first place

Was replying to CrtxReavr, but sure.

Tell us how it works out, anyway. I'd be curious to know.

ssm_, whats wrong with slackware ?

cpet: No package dependencies. Imports PulseAudio and elogind.

best parts of slackware!

we do as well

Title: I Do Not Want What I Haven't Got - YouTube

cpet: That's part of why I recommend Debian for the purpose - very happy with neither of those.

That said, I've been curious for years now about how Steam on FreeBSD (sans wine) is progressing.

just recompile x without pulse

cpet: no linux distribution has administration tools I enjoy using; when something goes wrong (which I expect is just inevitable with linux) I won't know how to fix it

what sort of admin tools are required that arent in *nix

most sysadmins create there own

right now im on a openbsd host logged into a freebsd system next to this one

ssm_: FWIW, if you get into an install, realize that you're probably safer building drm-kmod from ports, as the binary packages can be problematic as you move up minor revisions. While this is being looked at and will be address someday, for today you want to build it locally for the installed version.

s/address/&ed/

theres an option that rebuils a portt after each kernel

i do this and prevent issues with drm*

Ah, TIL. What's the option?

PORTS_MODULES

Set this to the list of ports you wish to rebuild every

time the kernel is built.

in make.conf

cpet, that mean they'd be build by buildkernel?

yeap

s/build/built

Cool.

this breaks freebsd-update if you add or remove options

this this is nice if you use current

I'm a luddite who maintains is world & kernel from src.

too each there own

Indeed.

Lots of rope.

pkg install tmux firefox xorg xfce4 hexchat drm-kmod

is all I do now heh

Why tmux?

changing drm-kmod for tigervnc

As opposed to screen, I mean.

its whats on open and got used to it

hmm

I've literally built-up over three decades of screen muscle memory.

on a lame point its not gnu

:)

however ive been on some peoples servers and they have screen doesnt bother me i just like tmux

CrtxReavr: I've heard you can remap tmux now to match your finger memories. Also, it's permissively licensed. As for me, it's free software and I like it, and I've actually needed it for serial consoles in the past, so I stick with it.

Not a big GPL fan, certainly, but I prefer functionality over license zealousy.

^

I actually quite like the GPL, just not how some folks interpret it.

tmux (and dvtm if you want something slimmer) can do lot of useful window layouts, unlike screen

BSD and other licenses are a lore more business friendly.

And if making terminal session reattachable is the only thing you want then dtach is much simpler

tmux works nicely when you have process limitations :P

Eh, having been part of several businesses that build things based on GPL'd code, I think that's more notional than real.

then get mad when they are forced to submit src

again too each there own regarding licenses

I mean. . . my employer flat-out forbids running anything covered under gpl3.

Also tmux's interface is quite nice for scripting access to the terminal states, should you for some reason want that.

I like 'em all. Free sofware, gimme more of it.

can someone explain to me what this answer is? forums.freebsd.org/threads/cron-pam_opie-so-and-freebsd-9-1.36934

Title: cron, pam_opie.so and FreeBSD 9.1 | The FreeBSD Forums

" Thus updating libs and binaries but not restarting the services that was already running..."

Someone know if it's possible to run Mac OS on bhyve technically?

what's bad with reboot?

Just not best practices. . .

yeah what did i do wrong

I use it quite often, init* smells too much SyS-V

(often as in, if I ever have to reboot)

You know how your system has the rc.d/* scripts. .

sure

And when your system boots, they all get run with the 'start' keyword?

reboot is just an alias isn't it?

the script's get invoked if you mean that

Well, the scripts don't get run wtih the stop keyword on reboot, if that's what you mean.

Not unless there's been a recent change.

hmm Im not sure if I remember correctly, I dont often reboot - today I had to shutdown a BSD host and honestly just pushed the power knob (Which uses ACPI)..

I mean, depending on what you're running, that may not be a big deal. . . but that's where it comes down to best practices.

so I thought I remember that the scripts issued with "Stopping XYZ..."

Normally, the shutdown(8) utility is used when the system needs to be

halted or restarted, giving users advance warning of their impending doom

and cleanly terminating specific programs.

That's straight from reboot(8).

hmm okay

So. . . ``shutdown -r now`

so reboot just sends SIGKILL?

ah well SIGTERM and SIGKILL

or ``shutdown -p now`` instead of halt.

Right.

which is not too bad since the programs are supposed to handle this :) but yeah maybe not soo great either

And you may be on a system where there's a shutdown or cleanup procedure for a specific service.

but yeah I simply dont reboot that often, not using FreeBSD as desktop

Like a relational database, or some shit.

ya databases are evil

It's just best practice to avoid use of the halt & reboot commands.

... what

ok so i keep getting errors regarding pam_opie.so not being found. this library file does not exist on my system.

SymbioticFemale, the link you provided... basically restart the services or reboot the machine

i see. that won't solve my issue then since i've restarted since any update.

SymbioticFemale: it was removed from freebsd

hi all. strange observation I don't understand. I have an additional loopback interface configured with some addresses assigned to it (for use in jails). when I try and ping an address assigned to lo1 e.g. 127.0.1.1, the traffic appears in tcpdump over lo0 instead.

that's because 127.0.1.1 is in the 127.0.0.1/8 network of lo0

so to compensate for pam_opie.so being removed, can i just comment out any lines in pam.d/* that reference it?

SymbioticFemale: yes; or rmove thehose lines

thanks!

ah yeah, thanks. for some reason I thought loopback was a /24. what addresses would you typically assign to a second loopback then?

do you really need a second interface for it?

hm. probably not. I guess the thinking was to just keep lo0 as standard as possible, but I suppose I could just assign the additional addresses as alises for that

the address is fine, but i'd skip using lo1 unless you had a great reason to do so.

cool thanks, will just assign them to lo0 - should work the same in terms of setting them as jail IPs

Can anyone tell me how to fix this? "The following files are affected by updates. No changes have been downloaded, however, because the files have been modified locally: /var/db/etcupdate/log". Happened after a freebsd-update fetch/install

rtprio: thanks again. it solved my problem and i can login again :-D

👍

not found anything substantial in the freebsd docs/handbook.. not sure if I overlooked something.

i'd look in the `man etcupdate` but i'd probably unmodify that file.

not sure how I'd "unmodify" it.. is there a stock version to compare it to?

i'd probably move it to the side and try to run it again

cpet: there was also a period when screen would randomly segfault on terminal window resize or just plain attach with a different window size than was when detached, it was absolutely infuriating

and that is how I switched to tmux

though I do have its command key bound to ^A

i did that as well

Remilia, was that in 1993?

tmux didnt handle window sizing properly until recently, and even then there's some special "super-long-to-type" options for it

(for reattaching anyways)

It's been a LONG time since I've observed issues like Remilia is describing with screen.

whereas screens reattach has automatically resized for literally decades

i've never seen that and i've been using screen since the 90s

-shrug-

CrtxReavr: it was from 2007 to around 2012

I've seen screen freak out when switching between certain combinations of terminal types.

I think it still happens according to a friend, I just no longer use screen myself

Probably an envirionmental thing, vs. an issue with screen itself.

thats my guess

CrtxReavr: PuTTY + FreeBSD, nothing else

Putty emulates lots of things.

maybe i've just been lucky in picking the right sized terminals to not cause it to crash on reattach this whole time ;-)

Plus, bytes do occasionally get munged.

nacelle: I am not sure what was wrong with tmux resizing tbh, unless you mean when you have two terminals attached to one session and one of them is smaller

Remilia: that is the problem

A ^L or ``echo ^VAlt-O`` can do wonders.

nacelle: you mean screen would resize the terminal window on the smaller side?

tmux attach -f ignore-size,active-pane

CrtxReavr: I mean, it is most definitely not an issue with screen if it segfaults due to terminal resizingf

it is expected behaviour

a tool like screen *should* crash unexpectedly

Again. . . used screen for over 30 years. . .

I've never encountered that, but you have. . .

irssi handles the resizing without issues

What conclusion are we to draw?

<=== irssi inside of screen, all day, every day.

I can't remember ever seeing screen(1) crash.

nacelle: yeah I ran irssi

and attaching to irssi led to segfault once every 1-3 weeks

CrtxReavr: I am doing irssi in screen and irssi in tmux all day, every day... -sigh-

I've seen it crash, but not on any chronic or systematic basis.

it only started sometime in 2007

Update your shit?

I havent bothered to upgrade the tmux on the box where i'm running screen, where its tmux doesnt support the ignore-size bit

also, I always used the latest versions of ports since I do not use binary packages

I guess ‘update your shit’ means using the git HEAD?

instead of ports

screen? why use that when tmux is so much better and cleaner code?

pertho: religion, I assume

tmux doesnt do all the screen does yet

Or "if it's not broken, don't fix it."

if you're used to what screen does and tmux doesnt do it... you still use screen

tmux is also noticeably lighter on my 1 GHz 6W Jaguar SoC

it does everything I need it to

try harder

nacelle: You mean attaching to serial tty?

no, i never use it for that, eww

tmux can attach to bhyve stuff and that is enough for me

(I am a minicom nerd)

14:43 < pertho> screen? why use that when tmux is so much better and cleaner code?

Did you come up with that on your own?

Then I have no clue what screen-specific feature you are talking about.

I already talked about some of it

CrtxReavr: screen has (or had?) a lot of really ancient, legacy code

tmux doesnt handle resizing well, and only recently (in the past year) took on the correct fix that works like screen -always has-, but you have to enable it with a horrible command line option

its not just like "screen -x" and bam yer in

I've never had issue with the way tmux handles resizes FWIW.

CrtxReavr: I'll go with pertho on this one. screen(1) uses the GNU brace style. It's far more pleasing looking at tmux source.

ccx: I definitely have, to the point of pursuing the fix with the tmux devs.

which is why its in place.

That said, I don't look at it ever so that hasn't been a factor for me. :P

Was it FreeBSD-specific? Because I've haven't encountered it in past decade or so.

no

tmux handles resizing fine; perhaps you should polish your tmuxrc

its been there forever. if you reattach in tmux without any options, the tmux virtual terminal stays the same size as the originating session

yeah I've not had issues with tmux resizing

I still think it is a matter of religion, I am not going to tell anyone to switch to tmux but when people tell me ‘screen is better, tmux is bad’ I can only shrug

it wont resize to your current terminal

setw -g aggressive-resize on in your tmux.conf

nacelle: it resizes for me as long as there are no other sessions with smaller sizes

automatically

I have no clue why

i'm not disconnecting my other sessions

(and it has been doing that since I started using it)

rtprio: I use ignore-size

the correct fix for me has been to use -f ignore-size,active-pane

(active-pain!)

"By default, all windows in a session are constrained to the size of the smallest client connected to that session, even if both clients are looking at different windows. It seems that in this particular case, Screen has the better default where a window is only constrained in size if a smaller client is actively looking at it. This behaviour can be fixed by setting tmux’s aggressive-resize option." hmm

I have mutliple workstations in my house, with multiple floors, and a phone that I journey out into the world with. From all of these I can access the same tmux sessions (through various vpns and whatnot, of course)... I dont disconnect my existing tmuxes, which a lot of paranoid people do (same with screen - how many people use screen -rd vs screen -x ? probably heavily for -rd and far less for -x,

though -x is -far- more convenient for the lazy)

I find tmux default perfectly reasonable there. But that option has been there for ages.

aggressive-resize doesnt take care of it iirc

its been a while since i've dug into this

(which is approximately when ignore-size came around)

So you want to crop the screen on some clients? Weird, but you do you.

(I'm not asking for a solution)

I want the terminal to be wrapped normally on each client

the irssi status bar should be roughly readable, etc.

its not hard

its not weird

It is hard if the application uses cursor-moving escapes and relies on certain terminal size. It will break certain applications if the actual and assumed terminal sizes are different. You may not use those applications but it's certainly not a setting that should be default.

It is a sad fact of the tty design that it hasn't been designed to be easily reflowable, but it's one of the many warts we've been stuck with for past 40 years.

Someone know how I can define an outgoing interface for an ipv6 defaultroute when I've two interfaces in same network?

megaTherion: I don't know if this works but try creating a static route for :: with a next-hop of <router ipv6>%<desired interface>

So like fe80::abcd%igb0

bahamat: thanks, good idea - that worked

damn I totally forgot about the % prefix syntax

ccx: I disagree with not as a default, but i'm basing that on screen having it that way as default for 30+ years - it worked well there imo, but thats just me

as a single user, I just use `tmux -u new -DAs default` and as far as I know, just tmux defaults

"-s default" is kind of redundant, no?

mason: regarding ssh to iso boot, yes that is what I ended up with doing... like I wrote, not entirely an "out of the box" experience :)

kolla: Maybe someday I'll get involved enough to offer up patches.

also... I got my zroot as I want eventually - I really cannot tell what I did different, but it _might_ have been a mislocated "canmount" vs. specified mountpoints

That'd do it, sure.

so now I have 8x4TB, all disks partitioned the same, all bootable, all set with first 50GB zfs partitions in a mirror zpool, and remaining diskspace in a raidz3