I have a dataset zroot/usr/home which is mounted at /usr/home. My homedir is under /usr/home, as I would expect it to be. When I create a new user with adduser I get a new dataset zroot/usr/home/NEW_USER and I don't want that behaviour. I'd like to have all new users be under the existing dataset. How do I achieve that?

Until now, I've been the only user on the system, so have never noticed this behaviour till today lol

mns: it's a relatively new addition; there' a -Z flag for it

see also adduser.conf(5) for a more persistent setting

ahh ok

kevans: thanks! I don't know how I missed this option when reading the man page :( I don't think it was mentioned in the Handbook

mns: that's not terribly surprising, the handbook doesn't catch up very quickly

I guess so. So this must be a really recent change. I'm on 14.3-RELEASE

yes, within the last year or so

cool, good to know

I was curious when this was added in: freebsd/freebsd-src 215c0a5 11 Apr 2024.

mns: I noticed that during a install recently. I attempted to try it without rtfm and failed. I just redid everything as normal. I was nervous about improting the pool.

well, having a dataset for each user allows more fine-grained control over them, you can give the user permissions to manage parts of their data set via zfs allow and you can have the individual homes encrypted and only decrypt on login via pam_zfs_key

nimaje: ty for info sounds exciting to mess with.

dch, ping

dch, absolutely lowest priority item but: did you get to check the poudriere-image cross-compile thingy yesterday?

yeah it worked for me

I didn't pick anything fancy, did you see the backlog around midnight?

so you got an image out of poudriere-image on an x86 host which works on an aarch64 target?

dch, aah no. I must have missed that - apologies.

I don't have cross-machine backlogs :<

no worries, its irc, this is expected :D

rtprio: slowly going up. SMART long test seems (?) stuck at 50%. Now 198 Offline_Uncorrectable ----C- 100 100 000 - 16 (was 8 yesterday). I'm certainly going to use a spare while I return the offending disk.

I have been strugeling with the Elantech touchpad on my thinkpad p53 since I got it, It has never worked on Freebsd but works just fine on Netbsd, Openbsd and linux. Never tried windowds. But on freebsd its just dead just as described in this bug report by someone else: bugs.freebsd.org/bugzilla/show_bug.cgi?id=266742

Today I accidentally held down the middle mouse button below the touchpad and while holding it, the touchpad works and so does the left and right buttons, letting go of middle button and they are all dead. Where do I go from here?

I am on 15.0-current, generic-nodebug kernel. I also have a custom kernel with nvidia passthrough patches that show the exact same behavor.

Should I make a new bug report of this or is it an "easy" fix?

Hello , I have a USB cd reader , but when I use it, I have a permission denied, my device is recognised like /dev/cd0, how to use it please?

dwho, you'll most likely want to setup a devd rule

err... devfs rule

jbo: ok , I have going to document it

I don't have a specific example for a USB optical drive for you but man.freebsd.org/cgi/man.cgi?devfs.rules

creek: why would you make a new problem report instead of adding your observations to the existing one?

jbo: thanks

nimaje thats what I was thinking but I cant be sure its the same problem, really, and the hardware is a bit different. But I was thinking of adding to it.

Freebsd 14.3, installed gnome desktop, all looks fine, but cannot log on back after wake from sleep

this frustrates me so much

every version I try to do desktop, it's gonna be something...

desktops are the last things I want to tinker with

Is it the "desktop" locked up? Is it still on the network?

imgur.com/a/Kxw0JJE this is the result basically

and I am pretty sure it's two small tweaks away but I just haven't seen it working out of the first try

ivy: bugs.freebsd.org/bugzilla/show_bug.cgi?id=288409

dacav: it already seems to be a setgid executable here as well

O_o wut

-r-xr-sr-x 1 root mail 68016 Jul 14 11:08 /usr/libexec/dma

Oh well, if that was already fixed, then the bug can be closed, I suppose

well, no

that's only part of the problem you're reporting

you're 100% correct that the permissions are too wide and should be fixed

Oh, I see. You're only looking at the binary permissions

yeah

the fact that the binary is already setgid just gives us a clear path forward

Then, according to the discussion we had the other day, probably the binaries should be changed to be owned by some dedicated user, say 'dma'

s/user/group

owned by root:dma

we have a `mailnull` user/group that dma would drop privileges to if run as root

I wonder if it'd be silly to just install the binary setuid/setgid mailnull to begin with, and fix the config permissions

How would a permission patch like this one affect the user's system during upgrade?

e.g. will the upgrade just fix the permission? Or will it query the user to notify them about the change?

s/user/root/

does anyone know if the the freebsd autofs is supposed to support nested mountpoints?

isn't there some UNIX/GNU/Linux X Window System (X) text editor that lets you edit by column, not just row? I know I used one years ago on Slackware and/or FreeBSD but can't find it...

edit by column?

like writting in Japanese ?

not necessarily

more like for writing/editing your own plaintext charts/tables or also ASCII/ANSI art

ahhh ok