Anyone else use signal-desktop and have it hang from time to time?

when I tried signal-desktop it crashed my entire network and entire X GUI

i onl signal on my phone so my employer can't monitor it

glad to see I'm not the only one it is a hot mess for

SponiX: indeed, hangs from time to time and the emoji picker doesnt work

anyone else get this audio crackling problem after some days without powering down ?

after a reboot all is good

not sure whats the root cause

it could be that your sound hardware might actualy be haunted

could be

who knows

pretty sure my wifi is also haunted, occasionally it will not work unless my laptop is fully powered off

Martouf: your wifi works when the laptop is fully powered off?

that is strange for sure

yeah really strange

im telling you haunted!

Martouf: yeah

i need an exorcist

Hi, Im looking for the bhyve equivalent to qemu's (Linux) kvm=off option which basically hides the hypervisor signature to the guest os

megaTherion: i don't think such an option exists

rtprio: ok so one still needs these nvidia patches

what problem does hiding the hypervisor solve?

rtprio: that the driver loads, it wont if it detects an VM/hypervisor

oh- that's rather lame of them

stupid shenanigans but what can you do

save your nvidia cards for nonvirtualized machines, i guess

rtprio: nah, bhyve just needs some patches - it works on proxmox. I prefer VMs and room/electricity otherwise I'd need 10 machines here

🤷

rtprio: dflund.se/~getz/Notes/2024/freebsd-gpu

Haven't tested it for 14.2 but it applies cleanly so probably works

megaTherion: can you create a bug with a request for this and mention I asked you to do so in the description (thj⊙fo)

I've learned only committers can tag people so you can't assign it to me

Q. Under what circumstances would ZFS sort of .... revert automatically to a state a few hours ago, post-reboot?

I'm asking because I had the weirdest situation yesterday. My server was misbehaving weirdly (failed connections, weird SSL errors, etc), so I rebooted it as part of diagnosing the problem.

After rebooting, it turns out all the changes to a certain (*NOT* read-only) jail from that particular day were simply dropped/reverted, as it they had never happened.

Turns out the reboot wasn't going to help anyway, as the physical server the VPS was on was overloaded (could that somehow cause it?), but.

FTR, zpool scrub rapaired 0B with 0 errors.

DarkPlutonium: VM hosts having problems leads to all kinds of problems on VPSs

DarkPlutonium: my VPS reported defective harddrive sectors becuase of that

Amazing.

DarkPlutonium: and one time i've lost ALL my data because my hoster messed up that bad

I've had that problem too (with a dedicated, but)

Damn.

I had that too.

Both disks in a zfs mirror failed simultaneously. Fortunately I had backups.

They're migrating my server, at least.

i'm on contabo. i was very pleased with them for over a decade, until they started expanding like crazy

now i would not recommend them anymore

Incidentally, I'm on Contabo too. Ha.

Used to use Hetzner. Considering going back, despite having some issues historically (including those 2 HDDs)

if someone using steam-utils, I tried to compile from ports today and is not working pastebin.com/bi5WHqBN , I managed to make it compile passing -m64 but I wanted to confirm

where I could find the latest changes in Linux compatibility? I remember that work was being done to allow support for linux coredumps but I don't seem to find that

Anything interesting (for the better?) change with the pkg builders?

For the past three days I've been able to run `pkg upgrade` and there things to be upgraded.

Normally it's at least a couple days between having an updated repo.

Or are changes small right now such that rebuilds are able to happen within a day.

i have a poudriere question, i currently have it setup to get for jails but the documentation.. that i have found.. talks about disabling the freebsd.org and was curious is there a way to have the jail look at the poudriere packages, if it does not find it .. then default to freebsd.org pkg

never mind.. looks like priority is what i am suppose to use

is this feature or bug, nullfs mounting nosuid fs enables suid again?

megaTherion: so you can passthru gpu? man when will this get merged

if i mount_nullfs -o nosuid, now it's off but mount happily says (nullfs, local, noatime, nosuid, nfsv4acls) in both cases!

rtprio: I hope soon, it might work with AMD GPUs already (I've none) but nvidia requires special patches

also it works with Intel integrated but well who wants those for gaming

were those patches to allow gpu passthru, or obscuire the hypervisor, or both

afaik there are these change of the hypervisor ident and something to BAR 0 loading .. but Im not too deep into the technical details

the patch file isn't really that long

I can't believe I got xrdp configured on FreeBSD 14 and able to RDP from my Mac Mini to the FreeBSD VM, what a wild trip it has been.

that hard?

It was

I think I once had such configuration but Im not sure if it was on FreeBSD or Linux

for me it was, but for you.....donno

on one Linux box I simply use vnc, other than that my BSDs are headless anyways

It was hard in a sense that I kept mis-typing xrdp_sesman_enable="YES" as xrdp_sesman_enabled="YES" in /etc/rc.conf and Is kept on seeing SSL error messages in the console for hours late last night :)

the whole sesman was annoying with it

Late last night my eyes were very tired :)

22

unless I am going mad, I can not see a way to get the licence of a package using pkg-info

pkg-info(8) sorry

also I was looking for a quick neofetch alternative, and I came across github.com/Macchina-CLI/macchina and they list "FreeBSD" here, but yet its not within the port collection...

Anyone know why before I complain about misinformation, would hate to make a fool of myself :P

also for 14.1-RELEASE -p6 is called -p5 :/

someone forgot to bump the patch identifier?

I'm a bit surprised there is no Palemoon/Floorp/Arcticfox browsers in pkg&ports...

i've never heard of those browsers before

polarian: the userland needed the patch and the kernel didn't -- I think - that is why they are mismatched

SponiX: yes, that's right.

armin: ports/packages pretty much just cover the essentials...

SponiX: ah, that makes sense

armin: you are free to write a port yourself :P

if it means that much for you, but as far as I am aware palemoon is written for Linux/Windows and just dropping such a huge codebase onto BSD tends to not go too well

polarian: yes, that's how it works and we hope to get it fixed early in 2025. it's poor for vuln scanning.

dvl: "we" you part of the team then?

sorry I have no clue who was elected :P

apart from d ch

polarian: I'm the one complaining.

oh...

polarian: lists.freebsd.org/archives/freebsd-…questions/2024-December/006036.html

polarian: The issue you mention causes many false positives and alert fatigue. I'd like it fixed. :)

dvl: ah so it is considered a bug that freebsd-update doesn't reliable update all patch versions due to it only rebuilding the vuln section

ok cool

at least its been worked on :)

polarian: no, it is the design.

polarian: everything is fully patched

if you see the -6 in freebsd-version -ukr, you got it all.

I know it's confusing and it's awkward, but if we can fix it....

polarian: doesn't mean too much or something, I could definitely try how far I get :)

I see two -p5's and one -p6

:/

this could be due to me failing to rollback properly as 14.2-release broke my laptop

although as I rolled back to a snapshot it shouldn't be a factor

polarian: seems right, let me check.

[20:08 r730-01 dvl ~] % freebsd-version -urk

14.1-RELEASE-p5

14.1-RELEASE-p5

14.1-RELEASE-p6

I have noticed it over and over again since I updated, iirc the dmesg states booting FreeBSD-RELEASE-p5 instead of -p6 and I thought "well thats weird" and just assumed someone forgot to bump the release version

yeah same :)

ok cool

I am not vulnerable or anything, you scared me a second :P

the way you said it I thought you meant "all must be -p6"

I understand

polarian: yeah, updates only affect the version number for the components they touch

kevans_: yikes, that is confusing :(

it gets even worse in kernel-land

iirc a patch that only affects a module that isn't built into GENERIC won't bump the kernel version

Isn't that the case in this case :P