How can I set rec src on a device?

for some reason my mic is set to playback

and it is a different device to sound out

things be funky

need rdp (xrdp, xorg, bhye) help termbin.com/p83f

Re bsd.to/VxmW -- looks like mortal user did not initialize the root's environment, much what rwp wrote about the leakage

Title: dpaste/VxmW (Plain Text)

... re fish history, also looks like HOME was either set to "/", or was just unset, after being root

meena: i'm not allowed to run my own mail

skered, what were you looking at in nut_upsmon?

s2r: Just trying to understand cases of upsd going down and why upsmon wants to shutdown.

Which I can trigger by restarting upsd.

skered, But does it shutdown the server? I don't think upsmon kills anything if it losses connectivity

skered it shutsdown when the battery levels reach predetermined low values.

I think it's MINSUPPLIES that's causing it.

But what are you observing?

restart upsd, shutdown starts.

skered wow.

I had many issues with my upss fortunarely detected disconnections to the ups but never initiated the shutdown.

can you paste your upsmon.conf ?

have you tried with upsmon -DDD to debug?

The difference here is upsmon talking to upsd not disconnection to the UPS.

I know that removing power from the UPS works and keep everything up.

I have an idea what was going during my tests.

I don't think the state was saved correctly or I restarted upsd during a calibartion.

Or it's not changing the state when I run a test via test.battery.start.quick or test.battery.start.deep.

Title: gist:047175203cf44a4a31f510ce59dbc93b · GitHub

But that's uspmon.conf

skered check mine. bsd.to/MT8f less options

Title: dpaste/MT8f (Plain Text)

No deadtime, no minsupplies

It should be defaulting to 30. I don't know aout minsupplies.

er well maybe 15. I thought I saw 30 in source.

And MINSUPPLIES should be defaulting to 1.

But I take it if you restart upsd while upsmon is running it doesn't trigger a shutdown?

s2r: How are you connected too? USB? serial?

skered, I've got to UPSs sonnected via USB

skered I'm 95% sure I tried it and it doesn't shutdown.

Don't want to try at 23:50 here before a working day.

:)

People work the last week of the year? Like yuck.

need rdp (xrdp, xorg, bhye) help termbin.com/p83f

why is usbhidaction not doing anything for me?

Title: HOWTO: Using multimedia keys from uhid(4) devices | The FreeBSD Forums

I am TRYING to follow that ^^^

OK, let's take it step by step. What does the grep in step 1 show you?

polyex: Is i3 in your PATH?

o/

or bash even?

Change that shabang to not use /usr/bin/env.

s2r: You know what I'm sure it's was that I run a deep test and forgot about it looking back on upslog logs.

It seems that upsmon doesn't send an even for a deep test.

vs quick where you see the USB flip and and admin off/sleep (and maybe a CAL) then not admin off/sleep.

V_PauAmma_V, it shows me that it is likely Apple keyboard on uhid1

and step 2 confirms it is indeed uhid1

because it hows up page=Comsumer usage=Play/pause etc.

OK. What do you have in the action file you created at step 3?

polyex: "UserWindowManager=.config/xrdp/startwm.sh", you likely want absolute path here? using $HOME or whatever is allowed in that .ini file

V_PauAmma_V, $ cat usbaction.conf

Consumer:Volume_Increment 1 0 mixer -f $1 vol +$2

Consumer:Mute 1 0 mixer -f $1 mute

And both keys show up in step 2?

yes and more, here are some

Input rid=0 pos=6 size=1 count=1 page=Consumer usage=Volume_Increment, logical range 0..1

Input rid=0 pos=4 size=1 count=1 page=Consumer usage=Mute, logical range 0..1

OK. So far so good. Now at step 4, what command do you enter?

yuripv: It should be relative ot the user logging in home dir

# usbhidaction -d -f /dev/uhid1 -c usbaction.conf /dev/mixer3 2

^[[24~^[[24~^[[21~^[[21~^[[21~

the funny input is me trying the MM keys, but getting 24's and 21's

gibberish

the keys are still operating as F8 and F12, instead of mute and volume increment, respectively

even though I am holding down the fn (function key) which is meant to activate them

Just in case that's the problem, try giving the full path to your usbaction.conf file.

it is in the pwd, but ok

no diff

Hmm. "^[[24~^[[24~^[[21~^[[21~^[[21~" smells to me like your fn key is not doing what it should do.

yeah

I do not know what to do, I mean, it works ok in Linux so it should not be a fault of the keyboard

but yes, the fn key

it's an apple A 1243

fwiw

What happens if you *don't* press the fn key?

same as if I do

it just acts as a F10 key or F12 key etc.

and if I do use fn key, same

AumShivaya: is it a thinkpad?

no a desktop

with a usb wired keyboard

so no 'fn lock' key

nope

I'm wondering: is the fn key handled in keyboard firmware as a state change key, or is it passed to the computer as a keypress?

my prototype is working again :3

I do not know

but

xev does not register it if I press the fn key

Is that xev under FreeBSD, or under Linux?

need rdp (xrdp, xorg, bhye) help pls termbin.com/p83f

freebsd

should I check with linux and see?

It would require a reboot a bit later

I have the feeling it is passed as a keypress in linux, but I would have to check :/

I was thinking along those lines too. But no need to run to check. I'm about to go to bed.

polyex, If you don't an answer, try freebsd-questions@ mailing list (lists.freebsd.org/subscription/freebsd-questions )

Title: FreeBSD Mailing lists: subscription for freebsd-questions

ty

s/If you don't an answer/If you don't get an answer/ # fudge

polyex: You might want to look at scrollback?

skered ok

yuripv man page for sesman.ini states "Path of the startup script relative to the user's home directory."

so no absolute not needed

polyex: Go back some more.

i3 in my PATH hmm...

And bash?

Just use /bin/sh

And try again

To see if the init. touch is executing.

PATH has /usr/local/bin in it and which i3 says /usr/local/bin/i3 so ya it must be in PATH

just use /bin/sh?

How are you checking that?

what?

Your PATH might be different at the time of executing that script

hmm

in my ~/.login_conf i have setenv which should build up the env there. any way to test?

lemme add the sesman.log output when i try to connect to the example code

Change the shabang of the script to use /bin/sh

of startwm.sh?

Yeah

k sec

didn't work

Ok I think you need to get the debug logs of sesman up then

almost gottem, sec

Are you sure you're using Xorg session type too?

no

what is that and what should it be?

You have xauth installed?

it seems to be when i type which xauth

Is your user in the video group?

no, only polyex, wheel, and operator

which works just fine in my other bhyve that's working. i'll double check that rn tho

You have debug logs turned on in that sesman.ini?

ok ya no video group for other guest

dunno...

no just info

i'll change to debug rn

Is xrdp-chansrv running after Xorg?

Xorg, chansrv, userwm.

Should be the order of things

ok...

And you have xorgxrdp installed too?

~/..xorgxrdp....log and ~/.xsession-erros might help as wlel

.xorgxrdp....log

xorgxrdp is installed ya. no log at /var/log/xorgxrdp.log

that has xrdp-sesman.log for DEBUG level

problem starting X i guess?

but Xorg shows up in the top for polyex when i'm connecting

And your touch isn't creating the file?

ok and nothing matching "chan" shows up in top when connecting

nope

Is this slow/old hardware?

The two vms are speced the same?

no it's a bhyve guest vm running on nice gear

on the same host

ya

ok

same 13.2 everywhere

You don't have a xorg log in your home dir?

.xorgxrdp.10.log

nope

.xsession-error?

What about

there's .local/share/xrdp/xrdp-chansrv.10.log but 0 size

~/.local/share/xrdp

nothing else in it

no .xsession-error

/var/log/Xorg.anything?

nope

Can you write to /tmp?

and if this isn't an example of why configuration management is worthwhile, i don't know what is

can i write to tmp as my user or?

as the user that's logging to into xrdp

Check for all those .X11 dirs.

/tmp/.X11-unix/X10 is what Xorg should be creatign

OMG

dude /tmp is like 755 not 777

ok, so I booted in to linux, and back...and it does not seem to treat the fn key (xev utility) as a keypress

aaand, that was it

V_PauAmma_V,

skered, tyvm my bro

also I am getting this error for Xorg: > Warning: Key <KO7> not found in evdev+aliases(qwerty) keycodes

> Not treating <KP7> as an overlay key

is it ok if paths in my ~/.login_conf have ~ in them? or should they be absolute?

I would use "$HOME" instead of "~". And set HOME once to the directory

... before any use

set HOME where tho?

Whatever use home directory happens to be ... I don't know , /home/polyex 🤷🏽‍♂️

s/use/&r/

sorry i still don't get it. in .login_conf i use $HOME instead of /home/polyex but i still don't understand where i'm supposed to set HOME

"once to the directory"

Context is same as when you started for setting HOME variable: <polyex> is it ok if paths in my ~/.login_conf have ~ in them? or should they be absolute?

... that is in "~/.login_conf"

i set HOME in .login_conf, like above the me:\ line?

or :HOME=/home/polyex:\ in the me:\ block, before :setenv line?

My purpose of setting HOME, instead of relying on it being set after login, to avoid tedious work of chasing down things not working when using configuration files across multiple software on FreeBSD & CentOS-derivatives

ya ok so you're saying, set HOME right at the core, in ~/.login_conf, then use it safely everywhere else?

polyex, Sorry, ignore me please regarding "~/login_conf"; have not relied or knowingly used it

s,/,&.,

where does HOME get set by on freebsd?

well, one location is the /etc/*passed file

"/etc/*.passwd" that is

polyex: $HOME is set by whatever logs the user in, usually login(1), su(1) or sshd(8)

V_PauAmma_V: it does seem like a timing issue, but i just rebooted a few times (for unrelated reasons) and the problem hasn't come back. i noticed my userland was slightly out of date compared to the kernel (1400502 vs 1400501), perhaps that was the reason

unixwitch ok so it should be pretty reliable whether interactive login or service login then as long as env is invoked for the user. so then what format do you use for paths in ~/.login_conf? absolute /home/unixwitch/foo/bar, $HOME/foo/bar, or?

polyex: i don't have a .login_conf, but if i did, i'd probably use ~/foo/bar following the example in /etc/login.conf

login.conf(5) says this works for "capabilities that set environment variables" which is probably the only place you'd want to use it anyway

specifically only :setenv entries, or :path too?

it must be both, because the default class uses it in :path=

true

alright, done deal. tyvm

well, i haven't tested it with setenv, so perhaps it only works for path, but that would contradict the manpage

UNIX Witches get magic from daemons

what's a witch?

so switching from python 3.9 to 3.10 requires recompiling devel/electron25, time to find something else to do for four hours...

just four?

meena: i think it took about 3:30 last time. no special hardware, Ryzen 7 5800X3D

i only have it installed for signal-desktop, hopefully one of the native IM clients might add signal support at some point

unixwitch: it's very frowned upon by the signal developers. not quite as much as by discord, but still

i'm mildly impressed that this is causing no issues at all with audio playback in firefox: 9:59AM up 2:42, 3 users, load averages: 24.66, 20.58, 15.06

used to be (years ago) doing anything at all on the cpu would make anything latency-sensitive very sad

ya that's really impressive

why load so high?

24.66 never even seen that

polyex: unixwitch is building electron

and smooth audio playback that's some nice scheduler upgrades?

polyex: poudriere... in this case i used 'poudriere bulk -J8' which runs 8 jobs in parallel, but sometimes it ends up building a few packages which i've set ALLOW_MAKE_JOBS_PACKAGES for, which means each of those gets 8 jobs by itself, so you end up with a lot of jobs

it's not really ideal because running that many compilers in parallel is not efficient, but it's quicker than not using -J at all, and sadly setting ALLOW_MAKE_JOBS still breaks some packages

polyex: not sure exactly why it works, but i'm using the new(ish) mac_priority(4) which allows jackd to run with realtime priority. so it's firefox > pulseaudio > jackd > snd_uaudio, i guess pulseaudio has a fairly large buffer, and jackd's realtime priority means it's okay with its smaller buffer

damn that's so cool

how do you apply the mac priority policy to jackd?

load mac_priority (either as a module, or compile it into the kernel), put yourself in the realtime group (gid 47) and then start jackd with -R

jackd_args="-R -doss -r48000 -p671 -n2 -w24 -e -P /dev/dsp0.0 -C /dev/dsp0.1"

it seems to get a bit confused by my 20-channel snd_uaudio device, the default -p1024 prints a warning saying to use a multiple of 671, but any value other than 671 causes it not to start at all... which seems like a bug but i haven't gotten around to investigating it yet

polyex: if you're interested in jackd on freebsd, i found this page very helpful: github.com/0EVSG/freebsd_jack_notes

Title: GitHub - 0EVSG/freebsd_jack_notes: Notes and user guides for jack audio server on FreeBSD

wow ty

unixwitch: how much Software do you have to recompile to work with Jack?

meena: so far only pulseaudio, because the default options don't include the jack sink. then everything else uses PA. (that's obviously not good for serious audio work, but it's fine for desktop applications)

i'm intending to try ardour, which should use jack directly, but didn't have time yet

why would pulseaudio *not* include jack support out of the box?

i don't know. it adds a dependency on audio/jack but that's a pretty small package

given the amount of people who would benefit from this, it seems a small sacrifice to make to the freaks who are like, as little deps as possible

OTOH, most of those people might be recompiling everything to not use pulseaudio because it was once touched by Poettering

i noticed a few things (including firefox) also have native jack support, but i actually prefer to route that stuff through PA because it's a bit more configurable, like forcing it to use only channel 1 for capture instead of thinking i have a stereo mic input. unless you can configure that directly in jack, maybe i should look at the manual...

for jack-pulseaudio tunnelling (or jack-pipewire), jackdbus works better than jackd (both link to libjackserver)

recently jackd got libdbus dependency too, so it could work if jackd emits jack-server-started and jack-server-stopped signals over dbus, so that pulseaudio or pipewire can create the tunnel dynamically

oh, electron25 build finished in 50 minutes, thanks ccache!

i'm jackdbus and ladish upstream and i noticed that ladish (that requires jackdbus) got recently added to freebsd ports

i dont have freebsd installed, but the design is same for linux and freebsd

recently (in freebsd-15)

unixwitch: in jack it is up to the driver what hardware ports to present. whether mic or guitar is attached to the audio io port is not something jack knows

but (oss) drivers can present mono devices i guess

nedko: right. my concern is that i connect firefox directly to jack, and it sees a 20-channel capture device, it's not going to know what to do with that. using PA's jack sink with channels=1 forces all the clients to just use a mono input, which is what i want for desktop apps

unixwitch: i understand, thanks for the info. ladish has "rooms" concept to limit visibility of nodes, but proper support in jack is needed for restricting the view of a program to a virtual hw with ports linked to particular hardware

what ports category would you put a Python Chromecast library in (PyChromecast)? devel/ because it's a library, not an end-user application?

or net/ because it implements a network protocol

nedko: hmm... ladish is like a persistent version of qjackctl?

unixwitch: i think qjackctl can also persist connections, ladish aims to do it automagically

that sounds interesting, i might have a look

ladish requires no session management protocol for the basic functionality (restoring connections)

even when jack client name is random

unixwitch: in case you've not found it already: lists.freebsd.org/archives/dev-comm…ports-all/2023-December/092042.html

Title: git: d92825bc02a5 - main - audio/ladish: New port: Session management system for JACK

good morning

heya

i have an application that requires electron 19 package since upgrading to freebsd 14 that package is no longer avaiable only election 25 my question is is there a way to compile an older version of election on freebsd ?

bsdbandit: if it does compile, you can just grab the old stuff (Makefile, package source, etc.) and make a package with it

bsdbandit: you could perhaps check out an older version of the ports tree (like a quarterly release) that still includes it and build a package

it looks like the port was removed on Fri Jun 30 2023, so 2023Q2 for example should still have it

however, it was removed because it's EOL upstream, so it probably has unfixed security issues

welp i submitted my first port, let's find out how many things i did wrong

hello everyone, merry christmas

can someone help me set up a torrent server on my freebsd 14.0 machine with a webui?

i've tried rtorrent, qttorrent and ctorrent but the furthest i've got is a cli with rtorrent that says it can't read the ,rtorrent.rc file even though i did chmod 777 on it

zc4r: have you considered naming it .rtorrent.rc rather than ,rtorrent.rc?

zc4r: i'd used both rtorrent (no webui) and transmission (with webui) from ports with no issues on freebsd. how are you starting rtorrent, and what's the full error / other output?

(actually, rtorrent wasn't quite no issues because if you run it on an nfs mount, it fails to write data because freebsd mmap() is broken on nfs and won't be fixed... but it should be fine on a local zfs/ufs filesystem)

@merry aren't those identical names?

no, one has a comma

(you typed a comma)

oh thats a typo, it's called .rtorrent.rc with a .

i just used sudo pkg install rtorrent and sudo pkg install transmission

do you think it would be wise to install from ports?

pkg is ports, it's just binary packages built from the ports tree. either is fine

you can always use -o import=‹file› if specify a different location if you have your rc in a non-standard location

ok, you have set up transmission with web-ui with no issues. i would like to be able to do that

it should install an rc.d script in /usr/local/etc/rc.d, just enable the service, start it, and then visit the web ui on the appropriate port, which i can't remember offhand because i don't use it anymore

you may need the 'daemon' flavour, i can't remember if the base transmission package includes that... this changed recently for the 4.0 update

('pkg search transmission', if there's a package with a name like transmission-daemon, make sure that's installed)

i'll start with "pkg install transmission transmission-remote-gui-qt5 transmission-daemon transmision-qt-4"

you don't need the qt packages on the server side. you might want them on the client

also transmission-remote-gui doesn't need to be installed on the server

oh

lemme undo that and try it again

that's an unrelated third-party package that's not related to transmission itself, but again you could install it on the client if you want to use it

so "sudo pkg install transmission transmission-daemon" gives Number of packages to be installed: 27

that's probably fine. i don't know what 'transmission' includes though, maybe just install transmission-daemon and it should pull in everything that's required

sudo service transmission start

transmission already running? (pid=71310).

ok, so you should be able to visit the web ui on whatever port it's running on... you can configure that in the settings.json which again i can't remember the location of (sorry) but find /usr/local -name settings.json should locate it

(if you want to edit settings.json, stop transmission, edit it, then start it again, or it'll be overwritten)

(i think it's something like /usr/local/etc/transmission/home/settings.json)

that's where the online guides seem to fail. the settings.json wasn't where it was supposed to be. i'll search for it

ok it's in /usr/local/etc/transmission/home/settings.json - what should i edit in there to be able to access the webui on my windows machine on the same network?

the internet tells me the default port is 9091, so try 1.2.3.4:9091 where 1.2.3.4 is the IP address of the system running transmission-daemon

if you want to change the port, the setting is called "rpc-port"

i get the same as before when i tried this: 403: Forbidden

ah, because it only allows localhost by default

try setting rpc-whitelist to something like "127.0.0.1,192.168.*.*" - but change 192.168.*.* to whatever your local network is

i figured as much but i tried changing a bunch of settings to no avail. i'll try that but it looks like something i tried

make sure you stop the daemon before editing and restart it after

also, github.com/transmission/transmissio…docs/Editing-Configuration-Files.md has a full list of config options

Title: transmission/docs/Editing-Configuration-Files.md at main · transmission/transmission · GitHub

to test, you could set rpc-whitelist-enabled to false and see if that works, but that's rather unsecure

zc4r: you may actually need a url like 1.2.3.4/transmission if it doesn't redirect by default (rpc-url setting)

"rpc-whitelist": "127.0.0.1,192.168.*.*,192.168.1.0/24,192.168.1.161,::1",

still gives me forbidden 403

does it prompt you for a password or just give the error immediately?

error

no user/pass

have a look at the logfile, which should be somewhere in /var/log

i see no transmission log file, but under /var/log/daemon.log i found something suspicious: Dec 14 16:33:20 myrkur ngircd[78733]: ngIRCd 26.1-CHARCONV+IPv6+IRCPLUS+PAM+SSL+SYSLOG+TCPWRAP+ZLIB-amd64/portbld/freebsd14.0 starting ... even though i don't remember setting up ngircd which is an irc client afaik

that's an irc server. no idea about that

oh nvm i remember trying to set up an irc server

although, try 'grep transmission /var/log/daemon.log'

couldn't connect to it, but nevermind that, do you think this line might be at fault: "rpc-whitelist": "127.0.0.1,192.168.*.*,192.168.1.0/24,192.168.1.161,::1", ?

that should be fine as long as the client is in 192.168/16

Dec 26 13:07:57 myrkur transmission-daemon[97935]: port-forwarding.cc:203 Mapped private port 51413 to public port 51413 (port-forwarding.cc:203)

Dec 26 13:10:29 myrkur transmission-daemon[97935]: rpc-server.cc:706 Couldn't bind to 0.0.0.0:9091 after 10 attempts, giving up (rpc-server.cc:706)

those are the newest two

that's odd, did you end up with two copies of transmission running somehow? ps auxw | grep transmission

Dec 26 13:05:55 myrkur transmission-daemon[97935]: tr-lpd.cc:259 Couldn't initialize LPD: Address already in use (48) (tr-lpd.cc:259)

another one

that could be the case

that sounds like you have two running

'service transmission stop' then look at ps and see if there's another one, if so kill it (or them) and start it again

i tried it twice, first stopped transmission, second gave this error: sudo service transmission stop

transmission not running? (check /var/run/transmission/daemon.pid).

what does ps say?

and if i do sudo ps i get 8 instances of getty whatever that is, and su -, -su (csh), su - z, sudo ps, sudo ps and ps

ps auxw | grep transmission

(getty is the daemon that displays the login prompt on the system console)

z 37129 0.0 0.1 43580 15796 - Ss 11:10 0:01.84 transmission-daemon

z 48317 0.0 0.1 43580 15808 - Ss 11:11 0:01.80 transmission-daemon -a 192.168.1.161

z 81963 0.0 0.0 12796 2456 0 S+ 13:18 0:00.00 grep transmission

Gen Z is killing the getty market.

zc4r: kill both of those, and run ps again to make sure they quit

in the hotdog flavoured water

TommyC: getty is trendy again with Wayland since lots of compositors don't support display managers :-)

heh, I haven't been following Wayland tbh

i've been using it for a week or so, i actually quite like it. it reminds me of old Unix systems where the kernel managed the framebuffer (which is basically what DRM is...)

i killed those two, should i try restarting transmission?

zc4r: if you confirmed they're not running anymore, yes, start it again (with 'service') and try connecting to the web ui again

orate

403 forbidden

i used to have an SVR3 system which had a non-X windowing system that ran on top of the system console... if you opened the terminal emulator, it just used the normal text console to display the output (except in a window) but you could display graphics too. it was actually pretty neat

zc4r: check the logfile again

this one seems fishy: Dec 26 13:21:25 myrkur transmission-daemon[8052]: tr-lpd.cc:259 Couldn't initialize LPD: Address already in use (48) (tr-lpd.cc:259)

okay, stop transmission, run 'netstat -an | grep LISTEN' and paste the output somewhere like bsd.to

Title: dpaste

i just get a "> "

no nvm

Title: dpaste/Mhdv (Plain Text)

there's sockstat for that type of investigation, sockstat -4l

you still have transmission running. do 'ps auxw|grep transmission' again, what does it say?

ps auxw|grep transmission

z 44844 0.0 0.0 12796 2456 0 S+ 13:26 0:00.00 grep transmission

(and there's also pgrep :D)

forgot to sudo, the daemon program was still rnning

ok, kill all copies of it, make sure they're actually dead (by running ps again), then start the service again and try the web ui

we're in business boiz

403 forbidden was defeated

now, how do i make it start on boot? put transmission_enable="YES" in rc.d ?

so, i strongly suggest enabling authentication for the web ui, you can do that with the rpc-authentication-required and rpc-password settings (see github.com/transmission/transmissio…docs/Editing-Configuration-Files.md for info)

Title: transmission/docs/Editing-Configuration-Files.md at main · transmission/transmission · GitHub

zc4r: in /etc/rc.conf, yes, or do 'sysrc transmission_enable=YES' which is an easier way to do it

but if 'service transmission start' worked, you already have that enabled

otherwise you'd have to use 'service transmission onestart'

so could you be more specific in the authentication proceedings?

yes such was the case

set rpc-authentication-required to true, set rpc-password to a password, then start transmission and it'll prompt you for the password. remember to stop it before editing settings.json or it'll overwrite it

(the user is always 'transmission', you can only configure the password)

those settings are in the settings.json file right?

yes

i tried that but the user/pass doesn't work

the password changes to {30105f05e08592cd2f9e51efe5bc3951d49c7903ir3gkz when i re=open settings.json

that's normal, it encrypts the password when it starts

not sure why it doesn't work though, maybe #transmission has an idea... you can turn rpc-authentication-required back to false in the mean time

thanks for the help

win 15

<33

damn cookie monster ate my keystroke

Does anyone the current status of Intel GPU supports in OpenCL in FreeBSD? The question seems easy to google but IMHO it is not

know*

unbound weird error. says fatal error could not read config file /unbound.conf but i'm not trying to use that path

if i comment out the include line in my unbound.conf, it works fine

include line is just include: conf.d/*.conf

ya it's choking on a symlink

any way to make unbound be able to follow symlinks and not choke?

presumably it's a symlink to something outside of the chroot?

ya

ohhh

chroot :)

hmm no

i commented the chroot line out, still get that error

"/unbound.conf"

also i get that msg when i service unbound reload, but NOT when i service unbound restart

kevans

AumShivaya, I'm out of ideas, then. I can try tracing the dataflow from the uhid driver to usbhidcommand, but my C reading skills are rusty, so no guarantee I'll find anything useful.

Oh, is the Xorg key error for Linux, FreeBSD, or both?

V_PauAmma_V any idea why unbound is saying "fatal error could not read config file /unbound.conf" ?

do you have an "/unbound.conf"?

no i have /usr/local/etc/unbound/unbound.conf, and that has an include: conf.d/*.conf

and there's 1 file in there

s2r: Turns out it's a bug in nut. networkupstools/nut #2168

Title: is_ups_critical: seems that UPS [xxx] is OB+LB now, but it is also calibrating : but it's not (longer) true (from 2.8.0) · Issue #2168 · networkupstools/nut · GitHub

2168 – A few common protocols missing from /etc/protocols bugs.freebsd.org/bugzilla/show_bug.cgi?id=2168

it's a symlink into /var

we can fget it much earlier

s2r: If any CAL state is set that isn't cleared. So it thinks the UPS is still calibrating when it loses comms with upsd. Should be fixed in 2.8.2

skered any clue why unbound is saying "fatal error could not read config file /unbound.conf" ? i have /usr/local/etc/unbound/unbound.conf, and that has an include: conf.d/*.conf, and the 1 file in there is a symlink into /var/...

is it configured to chroot?

it was but i commented that out lol

you might have, but it looks like it's still chrooting

i changed the include to be an absolute path to the /var file directly, so no symlink, still fails

dunno how it could be chrooting i commented that out and restarted unbound

ohh

gotta pass ""

ok now it works

uh huh

how bad is it to disable unbound's chroot?

so i can include this file from /var

how much do you trust unbound to not have a root exploit?

or is there any other way so i can keep chroot enabled? basically i have a script that puts a new file out with configs for unbound. the alternative i can think of is to put the file with the script directly into unbound's conf.d dir

eh

yeah, that's what i would do

somehow that means giving my script user permissions to write into unbound's conf.d dir?

how would you do that right?

what script are you doing to update these? not something like ansible, i am guessing

yes, i would have the script place the file from outside the chroot

FreeBSD 14.0-P4 USB is still fragile, has been reported, still no fix. Rebooting a server will sometimes hang if a keyboard is connected, pulling the usb keyboard will complete the boot process.

i haven't noticed anything about that during pre-release

Is there a bug ticket on this USB keyboard fragility that we might read and reference?

Bugzilla Bug_ID 273984

I'ts still ongoing, new mouse/keyboard didnt solve the problem.

that one is a port update by its maintainer and it is already committed and the report closed, you got the wrong id

Title: 273894 – FreeBSD 14.0-REL-p1 USB Connect / Disconnect Issue

Sorry .....

Need to fix my dyslexia, lol

hm, that one doesn't seem to be a hang, as xorg can be started and then it seems to be fine? "The error appears to go away after starting xorg / Mate Desktop."

On the desktop pc it works fine after Xorg has started. I was looking at dbus as a probable fix, andding dbus to the server (no xorg) didn't help.

I had to shut things down due to some wild weather, powering up the server through idrac failed to boot FreeBSD, the countdown timer had stalled. Pulling the keyboard from the server allowed it to continue boot process.

V_PauAmma_V, freebsd Xorg error

BraveheartBSD: I've been having constant USB dropouts, especially with my usb sound card, but I guessing that's more to do with my AM4 Ryzen hardware instead.

Interesting. That sounds like it registers like a keypress in FreeBSD, even if one software fails to recognize.

V_PauAmma_V, yes, and with a server in a remote location, it could make things interesting.

BraveheartBSD, sorry, that was to AumShivaya.