dautor: probably not without two or three handstands. the easiest way to associate an interface with a jail is to put it into the description

Xlib sure does come with a lot of default compose sequences‽

Does anyone know if the keycodes given by xev, and those used in the vt keymap files (/usr/share/vt/keymaps/) would be the same, or different ?

Compare the codes in /usr/share/vt/keymaps/us.unix.kbd (or your favorite keymap) to what xev produces. I think they are different.

i think the keycodes should be the same, aren't those the raw numbers send by the keyboard?

Thanks. I'm trying to map the keyboard of a brazilian thinkpad. Not sure how I can obtain the correct 'scan code' for key that's not working.

is there a decent Mp3tag-like app in ports? for mass tag-editing of music library

Uh, is it just me or is reddit being weird for everyone with locked posts?

"id3mtag-0.81 Mass tagging utility for audio files" courtesy of "pkg search id3".

(no idea about "decent")

nevermind, after a refresh

(That was freaking weird)

V_PauAmma_V: no id3v2.4 support :-/

but i was also looking for something like a spreadsheet-style UI... apparently musicbrainz picard is what most people use

i wonder if the windows version of mp3tag would run under wine, it's good enough that might actually be worth the hassle

I've been using easytag for that kind of thing. doesn't just do mp3s, but mp4 and what not. And it is in in ports, audio/easytag

also found "puddletag" which looks interesting docs.puddletag.net/screenshots.html

doesn't *just* do, I meant to say :/

Title: Screenshots — puddletag

BinGOs: this looks good too, thanks

(my library is mostly FLAC, not MP3s... "Mp3tag" is kind of a legacy name, it can tag almost anything)

picard or beats

i tried picard before mp3tag and really didn't get along with it, but maybe i'll give that another try too

For editing id3 tags. There are a few other CLI tools that I've used before, and one of them worked well, but can't remember the name of it now.

Gah, I really need to setup my knowledge base + search engine. Too many notes/snippets/references/etc that are scattered all over the place.

i really wish we had proper manpages for shell builtins like trap

i just want to know how the builtin works in every shell people actually use, i do not care about a comparison with csh

so i went with shotman for Wayland screenhots but it's actually not very good. is there a screenshot tool that can either save the screenshot, or pipe it to an external program?

I've been using scrot happily for years, but I've got no clue if it supports Wayland.

mason: this seems like a dead project so i guess not...

heh

Wildly popular dead project, but I guess no Wayland then.

it's apparently poplar enough that it's been forked by a project called "resurrecting dead open source projects" or something

s/poplar/popular

what the fuck. errors: No known data errors

this pool has had a phantom "error" for the last 2 weeks and now when i did a zfs send | zfs recv it's just vanished

there is something seriously broken with zfs in 14.0, i am sure of it. maybe only affecting encrypted pools

might update my fileserver to main for shits + giggles. can't really be any more broken than releng/14.0

Well... You did upgrade to a .0 release so you knew what you were getting into.

rwp: feh nonsense

rwp: if a .0 release is not supported to be suitable for production systems then it should say so on freebsd.org. which it doesn't.

s/supported/supposed/

otherwise we'd all turn into those people who are still on 11.x because "it's old so it's stable"

11.x… don't want 7.x/8.x/9.x? :-D

i'm on the verge of reinstalling with NetBSD and using RAIDframe because i really don't trust zfs anymore

that ossl(4) bug really turned me off it

like, i love the idea of zfs, but in practice it's so practice there's so many things that can go wrong

A lot of people play it safe and wait for the .1 release and that's when they upgrade.

used the wrong kernel encryption implementation? whoops, your entire pool is corrupted, sorry, better restore from backup

i will probably not do that though because i'd like to actually find the bug here...

I have 12 barematal servers with ~40 pools. Some of them created on 8.x. I had issue on one old test server only with broken RAM module.

I don't use the encription.

i do but i'm kind of wishing i didn't, i suspect my current issue might be related to that

i have obtained a much more serious nickname, now you must take me seriously when i complain about zfs

lol

meena: that's a good idea, thank you

i've ran zfs since it was introduced and haven't had any problems

i lost a pool when th chassis lost power but there's no avoiding that

rtprio: that shouldn't happen... with any filesystems, really, not just zfs

losing data on power failure is basically unheard of nowadays

what cool things can I get FreeBSD to do that I cannot on Linux as a desktop?

lw: shouldn't. but it was an esata external chassis

AumShivaya: /usr/sbin/spkrtest

Nixkernal: is right paste.scsys.co.uk/19448 to update bootloader?

Title: magnet_web paste from "ferz" at 217.168.150.38...

Outside of du/df can you find the memory usage of tmpfs?

after bumping ports this morning, I get this weird dbus error

ld-elf.so.1: /usr/local/lib/libdbus-1.so.3: version LIBDBUS_1_3 required by /usr/local/lib/libck-connector.so.0 not defined

skered: wrong with df?

and no desktop. zfs rollback "fixed" the problem

s/wrong/whats wrong/

elirco: I assumed there was a better way

Or some way to cram it in to top or vmstat

df -h /var/run | tail -1 | awk '{print $2}'

prints the size human readable (-h option can be left out if this is not required)

maybe that helps

df -t tmpfs would be better

true, my fault

But I don't need to know how to find it via df... I just thought there was a better way

Less roll your own

true... I darkly recall some memtop utility, but cannot find it atm

arg, so my poweredge with 2x broadcom BCM5720 keeps dying, major packet loss, i've bounced the network switch 4 times this weekend. i've disabled lagg0 failover, i've removed every extended feature on ifconfig (no hw accel, etc), and it's still happening

basically early on, it'd run for a week or two then start losing 50% packets

now it's every few days

bouncing the switch or ifconfig down then up seem to temporarily fix it

cgit.freebsd.org seems to be down for port 443

f451: indeed it is. h/w problem, it will reappear in due course

f451: you can use the github.com/freebsd mirror, or codeberg.org/freebsd in eu

Title: The FreeBSD Project - Codeberg.org

both src and ports as required

or wait for h/w stuff to happen

:D thanks

Hello! I have an old host with FreeBSD 14.0-RELEASE amd64, Pentium(R) Dual-Core E5300 @ 2.60GHz and nVIDIA `C73 [GeForce 7100 / nForce 630i]' as graphic card. I followed the instructions here docs.freebsd.org/en/books/handbook/x11/#x-graphic-card-drivers . However, when booting, the display keeps having a very low resolution (probably 640x480) till the user login prompt. kldstat(8) shows

Title: Chapter 5. The X Window System | FreeBSD Documentation Portal

that the `nvidia' module is correctly loaded. What else may be wrong?

(the installed driver is x11/nvidia-driver-304)

kevans: don't you think it's just a use-after-free or so?

kevans: but tty_rel_free also removes it from the tty_list, where it's still in

rockyh: do not know about that specific Nvidia driver version, but latest driver requires also the kernel module nvidia-modeset. maybe worth a try.

elirco: reading the Handbook, it seems that x11/nvidia-driver-304 requires the kernel module nvidia instead of nvidia-modeset.

(docs.freebsd.org/en/books/handbook/x11/#x-graphic-card-drivers)

Title: Chapter 5. The X Window System | FreeBSD Documentation Portal

I guess (at least) it's intended this way

rockyh: ic... and later driver does not support your graphics card I assume...

elirco: uhm no, my card model is not in the list of the supported devices

do we have an `eject` equivalent? in base? canonical/cloud-init #4732

Title: [enhancement]: List `eject` as a suggested/recommended dependency for the deb · Issue #4732 · canonical/cloud-init · GitHub

4732 – ac command works incorrect with old wtmp files bugs.freebsd.org/bugzilla/show_bug.cgi?id=4732

what's the patch, Steve??

meena: camcontrol eject may be?

yuripv: will that work on CD drives?

that's exactly for cd drives

rockyh: how do you start xorg? startx or a display manager?

man-dev.freebsd.org/camcontrol.8#EXAMPLES it's literally the first example

Title: camcontrol(8) - FreeBSD Manual Pages (Dev)

yuripv: i assumed it would eject a drive from a SCSI chassis. because that makes sense…

:D

as you can clearly see, i don't often touch hardware, but when i do, it explodes

elirco: I still did non install a wm. I would like to start it with startx, though. But I would like to have the driver active and a higher resolution in the text console before starting Xorg

ascreen: it's hard to say without malloc debugging options enabled

ascreen: could be use-after-free too, but without UMA trashing the distinction isn't necessarily easy to make

(thus we'd also be more likely to catch UAF back on development branches)

cgit.freebsd.org down?

yep

Title: cgit.freebsd.org not reachable? | The FreeBSD Forums

Title: The FreeBSD Project - Codeberg.org

Mirror

can you statically link libc?

statically link libc or link to libc?

Hi. I'm trying to get VLC to work in GNOME with my USB headphones. It seems that VLC doesn't respect the audio output preferences from GNOME. Is this a known thing? Is there an easy fix I just haven't found yet?

gh00p: what says 'cat /dev/sndstat'? Maybe you need to switch your default audio, since vlc directly uses OSS afaik.

can you compile libc statically into a freebsd program?

elirco, this: bsd.to/ybDE  VLC plays on the Realtek device, firefox plays on USB Audio, which I want. But FF is not a dlna client. :)

Title: dpaste/ybDE (Plain Text)

VLC allows me to select "pcm4:play" and "pcm4:virtual" as output devices, but the headphones are silent.

elirco, this: bsd.to/ybDE  VLC plays on the Realtek device, firefox plays on USB Audio, which I want. But FF is not a dlna client. :)

Title: dpaste/ybDE (Plain Text)

and wb. :)

gh00p: sry my irc connection seems unstable atm

gh00p: run 'sysctl hw.snd.default_unit=4' as root and restart VLC. That should do the job.

if I want to install freebsd with zfs on root on a ssd, and I want the swap partition to be outside of zfs, where should I create that swap partition? Is it ok to create it at the end of ssd drive ?

so I would have: 1. boot partition, 2. zfs on root partition. 3. swap partition

meandrain the default installer does that!

swap on zfs used to be discouraged

(I will not use the default installer)

not sure if it is still so

meandrain what will you be using?

Soni, I haven't done this myself but look at how the program /rescue/sh or any of the 148 hard-linked programs there are built. Those are statically linked.

desktop work with firefox and lots of opened tabs

meandrain: I describe what you want here: wiki.freebsd.org/MasonLoringBliss/ZFSandGELIbyHAND

Title: MasonLoringBliss/ZFSandGELIbyHAND - FreeBSD Wiki

mason: & guys, thank you!

oh

oh btw, uh, does freebsd have anything like LVM?

Soni: ZFS encompasses what LVM does. Outside of ZFS, I'm not aware of equivalent tooling.

mason: Is it important to make swap second partition instead of last ?

meandrain: Not at all, no.

Soni, ZFS is like LVM+mdadm+dmintegrity+more all together. ZFS FTW!

meandrain, On an SSD there should be no actual first versus last since all data blocks are mapped and in motion internally. Not like a spinning drive where one end is actually faster than the other end.

rwp: noted, thank you!

mason, (meandrain) I might suggest adding swap encryption to those instructions. (nudge, nudge) Since that is easy and only very slightly different. :-)

if swap will be used alot will that eat a specific portion of the ssd ?

rwp: Ah, good call. Thank you. I will. I should rework a few of my wiki things.

futune: yes, but afaik that applies to all file systems, why would you put the swap inside a filesystem if you could put it in an extra partition besides the filesystem?

mason, It's cool how loading a driver and adding .eli to the swap name and everything Just Works. That's so nice.

rwp: Wait, Isn't this happening already?

meandrain, "eat"? Whatever you allocate for swap is allocated and not usable by other partitions. Or do you mean damage? I have really beat hard on SSDs and have never had a Samsung or an Intel SSD die yet. Have a big stack of lower tier SSDs like OCZ that are just not good and failed almost immediately.

I thought referriing to swap.eli was already doing throwaway encryption.

consume/wear

yes, damage

mason, For encrypted swap don't you need the swap name to have a .eli to trigger things? I did not see .eli in your instructions. My /etc/fstab includes this line "/dev/mirror/swap.eli none swap sw 0 0" for example.

meandrain: Remember also that physical sectors are kind of abstract on an SSD.

mason: I see

rwp: check out the post-install instructions where we set up fstab

If I'm missing something I'd love to know.

nimaje, I would rather just give all my drives to zpools and use zfs to divide it up (if this setup was safe the swap zvol could share free space with the datasets)

meandrain, If you are actively thrashing then you will find the system very slowed down and will either need more RAM or smaller processes or fewer processes running. But having some swap that the kernel can offload unused blocks can free up that RAM to be used by useful other things on the system.

on a laptop with 16G of ram is it ok to use a swap in size of 4G ?

meandrain, Honestly I don't know how hibernate works on FreeBSD and for that purpose I cannot answer. Perhaps someone else can comment about use on a laptop. But for other purposes yes that should be okay.

well thank you guys, I'll start installing freebsd ;)

meandrain, On your laptop you will subsequently want to "pkg install powerdxx" and enable it "sysrc powerdxx_enable=YES" and start it "service powerdxx start". That will frequency scale your cpu and save a lot of power and battery.

meandrain we don't have hibernate, so don't worry about the size of swap.

After you install.

pretty sure you would need to give the zvol for swap a fixed size and couldn't have a sparse zvol, so no sharing of free space

nimaje, Strongly suggest not putting swap on zfs as it may deadlock under memory stress. schmidp.com/2014/01/05/freebsd-10-does-swap-work-on-a-zvol openzfs/zfs #7734 bugs.launchpad.net/ubuntu/+source/zfs-linux/+bug/1847628

Title: philipp schmid - FreeBSD 10: does SWAP work on a ZVOL?

nimaje: I won't use swap on zfs

I used on a hdd drive and when RAM was full swap was going crazy, If I did not free memory it would slow down the system so hard I could not use it

meandrain: that was meant regarding futunes comment

my problem with swap on root was with freebsd 12.x

meandrain really? I never had issues, even when using bsdinstall's non-interactive methods

mason, Apologies! I did a Control-F looking for .eli in that page and did not see it. But I now see it in there ("/dev/mirror/swap.eli none swap sw 0 0") and don't know why I did not before. My bad.

mason, I read through and I see "load aesni if we're on FreeBSD < 13" and I had a question that the other day in another channel. The "man aesni" still says to do that. I don't see a module loaded. Is that built into the kernel now, is that why it is no longer needed?

rwp: maybe that "ZFS Volumes as Swap" paragraph in zfs-create(8) should be expanded a bit

rwp: I'm going to have to struggle with bad memory there. Hopefully I left a comment when I made that change.

rwp: So, that came from here I believe. Let me dig through logs.

I don't know why I didn't find it with control-f ".eli" when I looked either. I am failing all of the time at simple things it seems.

Yeah: 03:55 < angry_vincent> aesni is now compiled in into kernel, not module ( since

13.0 )

in 2021

Credited in the commit comment. (Yay, commit comment!)

A user in #freebsd-irc who can't enter this channel was having a question about the aesni module the other day. And other than the man page information I didn't know anything. And on 13 after install I couldn't deduce any performance reason to enable it.

Ah! That explains things. Compiled in now. So no longer required to load as a module. Here is a newbie question. Is there a list of modules compiled into the kernel available in /boot/?

There isn't so far as I know any equivalent of a Linux /boot/config-foo

That said, if you look at the kernel naming, that should show you which kernel to config to look at in the source. GENERIC by default.

rwp: So, github.com/freebsd/freebsd-src/blob/main/sys/amd64/conf/GENERIC and we see an aesni line in there

I need to start compiling a kernel locally just so I can understand the process. It's a big hole in my knowledge base.

Title: freebsd-src/sys/amd64/conf/GENERIC at main · freebsd/freebsd-src · GitHub

oh good our freebsd HDD boots in a VM, last touched in october 25, 2019

how do we uh,

That appears to be just the device node though. Does that automatically compile in the module?

recover this

Soni, Recover? It boots? If it boots does it need recovery?

rwp: I don't think it's a device node. Pretty sure it's a device driver.

rwp: it boots in qemu

it does not boot on real hardware

mason, The comment above says "Pseudo devices." Makes me go, hmm... Ambiguous information as I read it. But perhaps the author thought it was clear by their thinking.

rwp: I had precisely the same doubt and confusion as you, but looking at the file overall, everything's either an option or a device.

rwp: github.com/freebsd/freebsd-src/blob/main/sys/conf/NOTES has an orientation of sorts

Title: freebsd-src/sys/conf/NOTES at main · freebsd/freebsd-src · GitHub

Thanks for that NOTES reference. I am scrolling through it now.

I used to build my NetBSD kernels by hand as a rule but honestly since VNET's been included by default in FreeBSD, I've had no need to build those.

Soni, Hmm... Boots in a VM. Does not boot on actual hardware. Could that be a mismatch of Legacy BIOS versus UEFI?

rwp: nope, had to set up UEFI in the VM so it's definitely UEFI

mason, I have been blissfully using the precompiled binary everything and it's been simply wonderful. Makes me lazy.

I'm right there with you.

Though it means I am getting the same experience day by day and not learning new stuff. I am starting to push myself out of the comfort zone so that I can expand and learn new things.

Soni, Hmm... I don't know. I would boot the hardware using rescue media like the bsd installer iso image and then, if I knew how, I would refresh the boot loader, "gpart bootcode -b /boot/pmbr -p /boot/gptboot -I 1 ada0" type of thing. Is that for UEFI or for Legacy BIOS? I forget! Refer to the handbook for sure. And then see if that allows it to boot.

Wups! That -i to -I is my automacro expansion because I am lazy and never type shift-I anymore. Should be lower case option.

rwp: You copy in UEFI bootloaders by hand. gpart bootcode is for legacy

Looking at the Fine Mason docs I see it is cp /boot/loader.efi /boot/efi0/efi/boot/bootx64.efi

A new RTFM for me is Read The Fine Mason docs. :-)

hah

They need to be updated.

For instance, there's old cruft that's no longer needed in my jail docs.

efi/freebsd/loader.efi is the one we'll primarily use these days, but if we're the only OS then you should update that one as well

The steady state of disks is full and documentation in slight need of update.

rwp: the bootloader is fine

rwp: hah, yes

it gets stuck when loading the graphics drivers or something

also this disk is from 2019 and we got a new GPU since, so that's something to consider too

Soni: Again, my memory is shoddy here, but some graphics drivers can't load during kld_list and need to be in /boot/loader.conf directly, and can do what you're describing.

kld_list

kevans: With the one caveat that you need a boot variable for the vendor directory version.

we create that by default

but not for the fallback

Well. Yes. Still. Just saying, it'd be an issue if you, for instance, move your disk to a new system.

yes

this is why i said they should both be updated

Soni, Maybe reduce /boot/loader.conf and /etc/rc.conf to minimum things needed?

you can also loose boot vars on, e.g., firmware updates

Yeah.

lose, even

Cry havoc, and let loose the boot variables.

so it's good to keep the default one around, but the vendor one is what we're using on a day-to-day basis so it's better to not be lured into a false sense of security with just updating one

around + updated

Only then I remember it's "let slip" and the joke falls down.

rwp: so basically take out kld_list and jackd?

Maybe. For me I am loading kld_list="radeonkms" for example. If I were to swap out cards then maybe remove that and let it boot and reconfigure later with the appropriate thing then.

As you can tell I don't know the answer and am simply trying to work through the problem with you in order to help. You said it starts to boot and then appears to get stuck loading the graphics driver. So that seems like a good place to start.

(this is so old, too... is there a way to upgrade freebsd?)

Soni: What version have you got?

mason: whatever was -RELEASE by september 11 of 2019

allegedly 12.0-RELEASE?

oh we downloaded that on july 29, 2019 apparently

the good ol days

Soni, "freebsd-version -kru" is three useful tidbits. Expecting -k kernel installed and -r kernel running to be the same. The -u userland base might have a different patch level from kernel.

After an upgrade and before rebooting the -k kernel installed and -r kernel running might be different due to the recent install.

rwp: that would require booting into it and now we're doing something else

It's just information. Which can be useful at the time it is useful.

If you do boot it and decide to upgrade I would upgrade it to 13 and follow the release notes for it before upgrading further if you decide to do that. Binary upgrade: "freebsd-update -r 13.2-RELEASE upgrade"

Title: FreeBSD 13.2-RELEASE Release Notes | The FreeBSD Project