can't load acpi_call: no such file or directory

that was from FreeBSD 14 kldload

oh, I had to install it

is loader.conf in /etc/ or /boot?

/boot

...I think you can easily remember that as a nmenomic to "boot loader" with /boot/loader.conf :)

ok

thanks

hmm, was that supposed to be sarcasm?

ascreen: somethingm ust have clobbered the session

I'd be curious what the rest of t_session looks like

my first suspicion would've been something assigning it &null->blah, but the only non-NULL assignment to s_ttyvp dereferences it beforehand even

(which made more sense after I actually looked at the name)

Have a system that was running fine on 13.2 booting on a GPT partition'd disk with EFI and ZFS. Ran into a sudden power-off that left the EFI partition needing fsck_msdosfs. Wouldn't boot. Restored the EFI partition from backup. Now, when trying to boot, the system claims the disk doesn't have a partition table. Hook the disk to any other than the boot controller, and it shows up just fine with the partition table.

What's going on? How do I track this down?

kevans: i would take a thanks as a thanks.

meena: It depends on the user.

Hi! My serial console got stuck at ifconfig tun2 destroy -- neither Ctrl+C or Ctrl+Z helps -- any hints how I can recover from this without hard resetting the server?

Greetings people. I tried to dual-boot Windows 10 with FreeBSD 14.0-RELEASE, I did it. But when I was on Windows 10, after a short time Windows 10 freezed and was not responsive. I thought maybe the sectors are corrupt. But FreeBSD did not freeze, I tried fsck on FreeBSD's UFS partitons, no errors found. Also checked Windows 10's partition with Windows 10's utilites(chkdsh, sfc), even no errors on

Windows 10's partitions.

I tried to install FreeBSD first, and then Windows, so the locations of the partitions would change.

But still Windows 10 was freezing.

The repair man says my HDD is corrupted and needs to be changed.

Is it possible that my HDD is corrupted?

Is it old hdd?

tankf33der: I have been using it for 3 years. And I did not use my laptop much.

It was a 1TB HDD.

Install windows10 only and wait for freezing

tankf33der: I have done it several times. What is the point of it?

To check if hdd is ok

Hmm. Good idea. What if it freezes?

It means HDD is corrupted?

Maybe broken from age

It's age is not high.

I bought the laptop 3 years ago.

Sata?

But if it is corrupted, how does FreeBSD work?

Uses unbroken space

I do not remember, but on FreeBSD the device node was ada0p[1-10].

tankf33der: I reversed the partitions' locations.

I installed FreeBSD at the top and Windows 10 at the bottom.

try to write to disk much

maybe check disk by programs from windows10

I have already done t.

done it*

ask master how he detect corruption

With chkdsk and sfc.

tankf33der: I will ask.

i am sure this is just hardware issue :)

If it was, then why FreeBSD worked normally?

I have to install FreeBSD alone.

And see.

Thank you for your time.

See you.

Welcome

its cold outside 12,4F outside inside i have 66F its getting cold in building

try recompile your port tree?

me?

Yes.

A joke/meme.

huh

k

Recompiling might heat yoru PC

your*

\o/

-rvvv

any other ways to heat my building

any other joke to heat my building

Burn some Ethylene and recompile cromium at the same time?

Or....dont be naked?

Or if you have a magma-chamber near you, you can try to drill a hole in it : newscientist.com/article/mg26134722…mber-could-unleash-unlimited-energy

Title: World's first tunnel to a magma chamber could unleash unlimited energy | New Scientist

But the best thing is probably open the windows a let the cold air out.

lol

meena: will do, but i don't have great steps to reproduce it

kevans: the s_* all seem weird? 0x0.st/H67n.txt and t_sessioncnt is 0

is it possible to have multiple access points in one box set up on freebsd?

i have a router with two ath cards, hostapd works fine

but there seems to be no way to set up multiple cards in hostapd

i want to have both 2.4ghz and 5ghz, thus multiple cards

Hi. Installing kernel using make installkernel I got following error:

kldxref: /boot/kernel/kernel: too many segments

What can I do to avoid troubles?

FreeBSD is generating IP packets with invalid checksum. How do I debug that? I tcpdumped traffic and noticed checksums are wrong but I have no idea what I can do to fix it.

dautor: did you check on the receiving end? i have a feeling that might happen when the NIC is doing hardware checksums

Yeah, that's on the receiving side. I haven't tcpdumped sending side, though.

How do I disable hardware checksum offloading? I think it be that the kernel expects the hardware to do it, but in my case, the packets are not routed to the hardware actually.

because they are sent to a ng_bridge and destined to another netgraph node

ifconfig <intf> -rxcsum -txcsum should work (although i've never tried that)

oh, you also need -rxcsum6 -txcsum6 as those are separate flags, and -vlanhwcsum as well. and maybe -vlanhwtso and -tso ... jeez modern nics do everything in hardware

i have no idea how this interacts with hw/bridges

er, ng/bridges

unixwitch: It worked! Thanks a lot! <3

dautor: i think you should file a bug about that, it should never be required to disable hw csums

I will file it now. I'm not sure what should even happen... I would actually expect the kernel to automatically disable offloading when lower hook is not connected to upper.

i know if_bridge(4) (not the ng version) has some weird interactions with tso/lso and forcibly disables those on the underlying interfaces. perhaps ng needs something similar for csums

Can I jump from 12.4 to 13.2 using freebsd-update?

ferz: yes

unixwitch: thank you

unixwitch: I see ng_checksum exists... It seems like that could be the solution. I'm going to try and make it work with that first.

ferz: dont forgett to update your bootloader after upgrade BUT before "zpool upgrade" and "zfs upgrade"

Nixkernal: thank you

Is there any guide for dummies to update /etc/default/ files using freebsd-update?

For example: is local_startup var dropped from rc.conf? I don't see it anymore in new 13.2 rc.conf file while I'm using editor to fix it.

ferz: you should never edit files in /etc/default. for /etc/default/rc.conf, you put your local settings in /etc/rc.conf

unixwitch: yes, I've never touched them in /etc/default/rc.conf but the freebsd-update is finding differences between old /etc/default/rc.conf and 13.2 version.

I would like to keep only the latest version but it's not so easy to understand what to drop and what to keep: paste.scsys.co.uk/19441

Title: magnet_web paste from Someone at 217.168.150.38...

ferz, freebsd-update is walking you through a merge cycle so that you can update the /etc copy from the reference copy. If there are differences then just update them. That's the expected path.

The old mergemaster was more clear to understand: you can keep old, use new or merge file.

If things have gotten screwed up then I would manually update /etc/default/ files from a pristine reference copy such as one found at /usr/freebsd-dist/base.txz assuming that exists for you as it does for me when using freebsd-update.

Or you can always do a bsdinstall to a directory and create an unpacked directory tree of pristine files and use it as a reference.

I've never changed anything in /etc/default/

Right. You have not changed anything there. But changes have been made in the released system bits. And so now for an upgrade you will need to upgrade your /etc bits to match.

ferz: uh, that's weird. i would keep whatever freebsd-update says is the new content, then grab the actual 13.2 /etc/default/rc.conf from git and copy it over (or at least diff it)

I'll dream a little and say that since we don't make changes to these files that I think the freebsd-update utility should be better about just performing the update without our help for those things. And I remember that in 12 it was better but for example now it is always prompting me that sshd_config is changed which is just noise to me.

ferz: freebsd-update is not supposed to work like this, it should just replaced unmodified files, so either it's a bug or you edited it by accident at some point

NixKernal No I´am not naked :)) If I open the windows I will only have more cold air inside :((

kevin-oculus: Just leave the windows open to let the cold air out, it's the best and most economic method

kevin-oculus, Similar temperatures here. I am warm next to my running equipment but the fans are a little noisy here. :-)

unixwitch: if I answer 'no' when freebsd-update ask if it's resonable, how can I continue without to make it from the beginning?

most of the times it's not visible the filename using it from a console and ssh connection.

NixKernal Thats for summertime not wintertime with cold air outside

freezing fog

is FreeBSD cgit web frontend down for everybody, or just me?

zear: not loading here

thx

zear: if you need src, there's a copy on github github.com/freebsd/freebsd-src

Title: GitHub - freebsd/freebsd-src: The FreeBSD src tree publish-only repository. Experimenting with 'simple' pull requests....

according to the forums, the relevant people are aware of the issue.

kevin-oculus: Sorry i got that the other way around ;)

ascreen: t_session->s_count is plausible if something underflowed, but the rest is garbage so it seems more likely that something else overflowed their own allocation and took out this particular session

hard to say based on what's been written what that something else might be

I always use cgit.freebsd.org when I am discussing code and pointing other people to a specific line but for myself I always use my local checked out sandbox copy.

Do we have a way to "start service X after service Y" ? I know I can add the service next to the REQUIRED keyword in the rc.d script, but I wanted a way without modifications

i think i'm getting bored of audio/sublime-music before the port has even been committed... it has weird issues/bugs and development seems a bit dead

I don't think I understand it enough to say it but this is documented in rc(8) and rcorder(8)

And I know that for example this will print the order: rcorder /etc/rc.d/* /usr/local/etc/rc.d/*

antranigv: if you do a `man 8 rc` and do a forward-slash search for "REQUIRE", that might be what you're looking for.

rwp true

kenrap yes, but I want to do that WITHOUT modifying the rc.d script

say I want to run nginx after named

so in nginx's REQUIRE I add named

all good, except when nginx's rc script changes

depending on how things are, either my changes will be gone, or I need to look at diff, or the file might not me touched at all

I think I need to add <name>_<flag>_<action>

for rc.conf that is

I am reading the docs and I don't see a way to add a local override off to the side of the file and modifying the file seems to be the expected process. (mergemaster would then merge it on the next upgrade.)

so someone could do `nginx_require_add="named"`. well that's bad design but you get my point :D

I'll dig into the code and see how to create a dependency tree for rc

anyone runs freebsd on a thinkpad x250|? i cannot wake it up after sleep

Seems like for example nginx the entire script from /etc/rc.d/nginx could be copied to /usr/local/etc/rc.d/nginx and then that copy modified as needed.

mane, I have a similar problem and have not resolved it for myself yet but the laptop wakes up but the display driver is dark. I can ssh into the system but I haven't figured out the driver issue.

you need to reset vidseo

I think it likely that your x250 is awake but just without a working graphics display.

hw.acpi.reset_video=1 is what you need

rwp it has nothing to do with a graphics driver, the system does not wake up at all, it just remains asleep

i cannot resume it entirely

...I was thinking that, e.g. copying the "nginx" rc file as "nginx-custom" and *THEN* modify the latter's contents without interferring with the original pkg provided file could be a way to go about, but that wouldn't solve needing "diff" every now and then...

whatever

this was fixed in a recent commit on main right? /usr/include/netinet/tcp.h:82:8: error: unknown type name 'inline'

maybe i didn't installworld recently

mane, I am pretty sure I tried that hw.acpi.reset_video=1 and it did not work on my Thinkpad x201 that I was using for testing. But I don't know. I'll try it again.

hmm ok apparentlty it was not fixed

wait no. it was. i just forgot to update the jail

This is kinda a Linux-question, but not fully. Anyone know how to install libfetch on Linux?

I can't seem to figure that out...

unixwitch: i finally starting looking at sanoid, holy moly is it easier than zfs-periodic or other zfs send | zfs recv things..

fikran: that depends on your linux distribution, i suggest asking in #debian or #fedora or whatever whether it's been packaged

^

voy4g3r2: right? like i said when jbo was talking about this the other day, life's too short to write this stuff by yourself

exactly!

you just made me shit my pants

once i can figure out this flock /var locking, i think i am good

just pushed a commit and got an IRC highlight a few moments later... thought I fucked up.

apparently audio/cmus is also broken on freebsd, the file browser doesn't work properly... what do people use to play music (mp3/flac)?

unixwitch: do you use that command line utility to "push" the snapshots to a remote server.. or you do it in the cron job?

voy4g3r2: i have syncoid in cron

unixwitch: pine player on my mac, typically flac.. i loothe mp3

unixwitch: yeah, me too.. i am trying to figure out the remote part but i could just be reading things wrong.

jbo: nah, unixwitch was just complementing your skills and sanoid

jbo: I hope that was a figurative :)

kenrap, this time it was :D

phew

my blood pressure is certainly higher since I became a ports committer.

unlike src you push a commit and stuff breaks immediately :p

voy4g3r2: le-fay.org/tmp/30d/syncoid-backup.sh

jbo: I'm already getting high amounts of anxiety just from hearing about your anxiety :/

ohh

unixwitch, I need to talk to you about sanoid at some point btw. I was playing around with it at home before putting it to use on servers and I had some weird behavior where syncoid wasn't able to successfully push "updates" after the first initial sync :/

jbo: i am temporarily not using it because of a different issue where my zpool is reporting random corruption, triggered by snapshots...

well that is helpful, thank you unixwitch .. i am trying it under a backups user

those two things, you both bring up, does not give me much faith

but whatever.. can't fix something if you do not break it first

my issue at least is not caused by sanoid/syndoid itself, and i haven't seen anyone else report it

i might de-encrypt the pool and see if it's caused by that, or maybe try with GENERIC kernel

well my issue right now is.. trying to save a /var/run/sanoid/snapshot.lock file when backups can not write to that directory

Welp, it would be time for me to experiment with the new drm-61-kmod package and see if and how I can get my Arc A770 GPU to work

kenrap, Thank you for your work re GPU (I do not have it; for future)

parv: No problem! I'm thinking of making a reddit post guide if I'm successful

(assuming it's a bit of a process)

Going to jump back into 15-CURRENT first and learn to take advantage of BEs before getting started

huh, bectl isn't all that complicated...

Probably because of "newcomer" anxiety from half a year ago.

i do have to say sanoid code is quite nice to read

This may be heresy, but I find the per-application volume/mic controls in pulseaudio really helpful. I am curious whether that is actually what folks use these days for more advanced sound configuration, or is JACK still a thing?

rtyler: i used to do this but i found trying to mix every application in a way that works consistently was just too difficult. it's easier to just adjust my hardware volume control as needed

i do use pulseaudio though, connected to jack -> oss in bitperfect mode, that's mostly so i can use jack with ardour (which i still haven't gotten around to trying out)

success, we got sanoid backing up stuff and wow that was fast

I think sndio does something similar to pulseaudio but it's a simpler design. Although I suppose it really comes down to app support, so stick with pulseaudio

that's pretty much why i went with pulseaudio. i don't have any particular preference, if everything supported sndio i'd probably use that

fair enough, Iwasn't clear based on the handbook how/if pulseaudio was in good shape atopi FreeBSD

i haven't had any issues but i'm not doing anything complicated with it

rtyler: it's pretty good, pactl(1) is your friend also :)

and pacmd(1). This one is more useful because of it being a repl shell for pulseaudio

I forgot to uncomment the `KERNCONF=GENERIC-NODEBUG` and `WITH_MALLOC_PRODUCTION=yes` lines in my /etc/src.conf...

Is there a way to get JID from a ng_ether node?

(I am using VNET jails)