[tj]: Thanks

Anyone know how to fix an invisible mouse cursor in xorg (KDE or XFCE both)?

f451: tsoome: remember my issue with zfs and mounting problem in jail? in fact you cannot mount a ZFS dataset as an unprivileged user, only root can do it

dh: there's no xfce4-desktop package in 14.2 p2

it's a bit strange to me that we have vfs.usermount, ZFS delegations, jailed dataset, etc and that only root can mount it, it takes away a bit of interest

also I don't see the point why you're allowed to create a dataset but not mount it, although I understand that it's different layers

mage would you want me to be able to mount a zfs data set of /bin

s/of/over/

crb: if I'm in a jail and that I explicitely allow it.. why not

crb: I'm talking about jailed dataset only

with delegations I guess you can do it on the HOST anyway

mmmh seems docproj package is missing too

the ability to mount a filesystem is the ability to put binaries in a place where they will be trusted to execute, that's a security hole. Even in a jail. At least to the things in the jail

Posterdati: You may want to try building it from ports then

mage this does sound like inconsistency. probably something for developers to think about.

dh: yes, but I cannot install ports, seems that there are other missing packeges too, like docproj

i don't mean to be vague (I do mean to bevague): DAE X11 memory leak on amdgpu?

wayland here

so is @world just everything in /var/lib/portage/world?

i tried sway, it was the only time the mouse has ever been jumpy and sticky when not under load

oh crap this isn't gentoo is it

my freebsd installs dont have x11 or wayland :)

be happy you are not in this hell

tsoome: do you think I should open a PR ?

sure.

either you should not be able to create dataset in jail or you should be able to mount created dataset in jail.

that sounds logical indeed

and the behavior should be clearly documented.

in restricted environment they usually disable mount to avoid attacks with corrupted file system access, but this is not the case with zfs mount.

looks like there is already one bugs.freebsd.org/bugzilla/show_bug.cgi?id=232021

huh, 11.2:D

ok, last comment refers to 12.2, I think it should be updated to current version;)

Hi there, I'm having issue with starting X at my machine. I posted in fBSD forum but haven't got any answer (forums.FreeBSD.org/threads/graphics-not-working-intel-gpu.97478)

I wonder if anyone can help me

tsoome: I've added some comments

hello, quick question, what happened to golang packages in 14-RELEASE? none of their versions show up, fallouts show some build errors, and now a go-compiled package I use is gone

well, if there are build errors, that's why they aren't available..

they are still available in 15 and 13 though

15 and 13 are built on different schedules though

I also went to bugs.freebsd.org but didn't find any ongoing things about it

also, I could build go123 on my freebsd box

so just wondering

which means I have the binaries if you need them :)

for x86_64, that is

lists.freebsd.org/archives/freebsd-ports/2025-April/007557.html tldr there is/was some bug in -CURRENT which made builds fail, builders were rolled back to an earlier version of -CURRENT, so now it means waiting for a new build of those repos and hoping that bug was the only reason go failed to build

bugs.freebsd.org/bugzilla/show_bug.cgi?id=285963 appears to be the PR that you couldn't find

thanks, subbing for that one

i found it on the freebsd-go list on lists.freebsd.org :)

I looked for the text www/go123, hence why I didn't find it

sorry, lang/go123

interesting that wildcards don't appear to work in bugzilla

btw are you guys doing your freebsd-related work all voluntarily, for free?

well, my "freebsd-related work" is complaining when stuff is broken and sometimes providing patches, I should really look into fixing that login.conf bug I opened, not being able to use " in strings is annoying

> The crash from Go seems to be in the garbage collector, I cannot reproduce the issue when I disable the GC with GOGC=off.

this sounds pretty bad

definitely subbing that bug

(and I definitly postponed sending that patch for the netsurf port too long)

dh: I'm compiling it from ports :)

Hi there, I need help with getting graphics to work for my system

I have a post in freeBSD forum ـدار اإلفــتــاء املصريّة. ودُور

اإلفـــتـــاء فـــي الــعــالــم الــعــربــي تخضع

لوصاية ماليّة وعقائديّة من النظام

الــســعــودي (بــاســتــثــنــاء مــفــتــي ليبيا

and will appreciate any help

sorry some rubbish were pasted by mistake, apologies

have you tried without any xorg config and loading i915kms before starting Xorg?

yes I tried without any xorg configs, and I start xorg manually after the i915kms been loaded

nimaje: I always get couldn't open module intel in the xorg logs

well, you don't want intel, you want modeset

Installed packages to be UPGRADED: firefox: 137.0_2,2 -> 137.0.1,2

hu? 2 to 1?

I removed the xorg configurations and that's what I get

intel would be the xf86-video-intel legacy driver

Xorg should try the modeset driver after it failed to use the intel driver, look if you get any warnings about modeset

if I put the xorg config files where I specify the driver to modesetting I get different error

what error do you get there?

(without a xorg config it should try all drivers until it finds a working one (probably in alphabetic order so it tries intel before modeset))

I paste above

are you in the video group?

yes

nimaje: in wheel and video

armin: 137.0.1 is newer than 137.0 regardless of the portrevision

ivy: I'm blind, thank you.

Rand0m: are you sure you don't have any xorg config left in /usr/local/etc/X11/xorg.conf.d ? there are some (**) Option lines in your log, which aren't in mine

nimaje: the logs in paste.debian.net/1368874 after I copied the configuration back

this [Rand0m> nimaje: I always get couldn't open module intel in the xorg logs ] was when I removed them

sorry for the confusion

so, try again to remove any xorg config you have and pastebin the logfile of trying to start X like that, the message that it failed to load intel driver is expected as it is the legacy driver you don't want to use and don't have installed, but it gets tried first

sure, will do now

nimaje: paste.debian.net/1368879

hm, why are you missing (==) Matched {intel,modesetting,scfb,vesa} as autoconfigured after glx got loaded? is that why it only tries intel? but why does it even try intel there?

I don't know

you installed the xorg meta-package and drm-kmod?

but nimaje when I added the configs as per the previous logs, it seems it was picking the modesetting. but why it wasn't starting?

what was the config you added?

paste.debian.net/1368874 that's the one with the config file

and in the config I had Driver "modesetting"

just Driver "modesetting" ? shouldn't it contain section or something?

yes it does let me pasteit for you

if you try without the BusID ?

okay, will do it now

without BusID paste.debian.net/1368886

hm, why does it fail to find devices? (and why does it try to use some fonts it didn't try without a config? and why does it try to load intel? and why does it set some Options it didn't before?)

the fonts are from a different config file, but that shouldn't be an issue though

please pastebin the complete config, not just some section you guess is relevant and please test with only that fragment you pastebinned before

okay I just kept the driver one without the bus and this is the log paste.debian.net/1368887

so this is the driver config without the busid

only

ok, thatlooks like what I would expect, except that it doesn't find any screens and still tries to load intel

so what to do?

if you try with BusID "PCI:0:0:2:0" ? in your config you only had one 0 before the 2, but there are two in the pciconf output in the forum post

okay let me try it, I will add BusID "PCI:0:0:2:0"

but in saying that, and as far as I know, this is only important if you have multiple cards installed. right? but I will try it now, one minute and I will paste the log file

xorg should autodetect it, but it doesn't for you for some reason, so lets try if it works if it is specified explicitly

nimaje: that with BusID with double 0s before the 2, paste.debian.net/1368891

Sorry got disconnected

my last message was: nimaje: that with BusID with double 0s before the 2, paste.debian.net/1368891

ok, progress, with the correct BusID it finds the screen, but for some reason the default screen section doesn't match it "Screen 0 deleted because of no matching config section." but before that "No screen section available. Using defaults." I would assume the default should match any screen that is found

When I discovered mdo exists i uninstalled doas

salute

then the next test would be giving it some explicite screen section, no idea how it should look like

I can boot nomadbsd and run xrandr

get the screen id from it

look if they have anything in /usr/local/etc/X11/xorg.conf.d or /etc/X11/xorg.conf.d

nope, I will boot nomadbsd from the usb (a bit slow though) and run xrandr. the monitor/screen configuration should look like paste.debian.net/1368894 with the correct identifier

will be back shortly

Hi nimaje, I pasted the monitor config and the logs paste.debian.net/1368897 Thanks

that is now on FreeBSD with the Monitor section added to that Device section we created before? the config complains about not finding a matching Screen section and according to the xorg.conf man page a Screen section is there to connect a Device to a Monitor

I just added screen (although I never had to create it before) but nothing changed

I will past the logs and configuration

yeah, no idea, autoconfiguration always worked fine for me

usually I add the monitor when I have multiple screens but anyway I will paste the logs now

nimaje: please find them in paste.debian.net/1368899

nimaje: should I be using drm-515-kmod or drm-61-kmod?

drm-kmod will depend on whichever you should use on your freebsd version

Rand0m: which GPU?

I'm using drm-616-kmod, since 515 will not work with my 6750XT

thedaemon: device = 'WhiskeyLake-U GT2 [UHD Graphics 620]'

either should work fine for that one I believe

8th gen?

yes

that's old enough it should be covered in 515

I tried both but none is working

oh?

so let me try again 515

have you loaded the kernel modules?

Do you know the steps involved?

yes, the kernel modules were loaded, I'm removing the drm-61 and installing 515 again for another try

personally I had to compile drm-616-kmod from ports

good luck

Do you reboot and it turns off our screen then you reboot into single user mode?

I did compile it from ports

:( I'm sorry

Well, lots of people use it

it works on nomadbsd but struggling to get it to work on freebsd

compiling the 515 shouldn't take too much time

Well, since NomadBSD is FreeBSD, should be a matter of figuring out what is wrong

I tried everything, here tonight we tried again

thedaemon: no luck :(

Thanks nimaje and thedaemon for all your help, It after midnight here and I will call a night

and probably a good time to tell everyone to upload their dmesg to dmesgd.nycbug.org/index.cgi and a probe to bsd-hardware.info

if you are running FreeBSD or interesting hardware, please feel free to upload you dmesg

cheers. for my understanding, should yesterday's FreeBSD-EN-25:07.openssl have been a security advisory instead of an EN? or is it not a vulnerability, depite the description? (i tend to take care of ENs much later, so i was surprised)

does anyone reccommend a data center? I'm looking for a bare metal server in Miami/Tampa or something near

Yaazkal: is that like renting a vps or ec2 instance? does the location of it matter that much?

not vps, looking for bare metal server. Looking for those locations because it's near to the users

oh yeah. it's a crazy world of people buying all kinds of cloud and server based stuff.

i didn't even know netflix had a bunch of CDNs around the country with freebsd just to serve content faster.

Yaazkal: i have friends that use freebsd on vultr, it appears they have a center in flordia

i don't recall if they have full servers or just vps, sorry

rtprio: they do have bare metal, but it's not clear if I can customize them (like adding more disks). Seems they did not understand my request because they just answered "If you need more space you can buy another server". I want to have a server with a mirror for OS and then mirror striped for jails (6 disks total)

not a recommendation but hostdime has orlando and miami locations.

I'm trying to get a quote with Hivelocity... but has not been easy too. Really their service is like... meh (bad experience recently)

isley: I'll take a look, thanks

they were okay

not great

rtprio they who? vultr?

hiveolicty was okay, i guess

Yaazkal: leaseweb.com/en

hernan604 thank, I'll check

dk: I would have expected that one to be a SA too, even if it says 'Systems not exposed to low-latency adversaries and systems not using "exotic" elliptic curve parameters are not affected.' same with the expat one

especially as low-latency adversaries just means that you need some second part that gives you RCE

Which is the Wireguard implementation I want on current system? wireguard-go from ports?

the one in the base system, maybe you want to install wireguard-tools to have wg-quick

^ wg-quick from wireguard-tools is what I use.