dch: i can't get the nvme to boot, i can get into pxe if i use the iso

should that matter or will it confuse the loader

sigh.

ld-elf.so.1: Shared object "libpython3.9.so.1.0" not found, required by "python3" in a fresh jail

the file is definitely there, any idea what's missing? some sort of LDPATH thingy?

rtprio: I'm missing some context sorry,

this looks like iso forces you to BIOS/MBR boot, which works with pxe, but the nvme path doesn't and thus you end up with uefi ?

it seems /var/run/ld-elf.so.hints is missing stuff

what is supposed to create that file?

answer is `ldconfig /usr/local/lib` why that shouldnt have already been run in this jail is lost to me

that's wrong answer as it will replace the contents, losing any other non-standard directories; correct one should be `service ldconfig restart`

yuripv: ok, but if I already restarted the jail, shouldnt that have been done

yuripv: afaict it didn't overwrite it, according to `ldconfig -r`

do I need gpu drivers installed in order to control screen brightness?

I just found the answer to my question

Hello! (If this would have been more suitable for the freebsd-arm mailing list, I'm sorry)

I would like to dig into the world of FreeBSD, but unfortunately I'm bound to a Macbook M1 device.

So I will have to use virtual machines for using and working on FreeBSD.

First of all I tried if running the ARM64 port of FreeBSD inside a VM works on M1.

Using this gist.github.com/ctsrc/a1f57933a2cde9abc0f07be12889f97f it was only a matter of minute to get it running.

My next step was to build BSD from source, which turned out to be a little challenge but still solvable.

Title: Guide: Run FreeBSD 13.1-RELEASE for ARM64 in QEMU on Apple Silicon Mac (MacBook Pro M1, etc) with HVF acceleration (Hypervisor.framework) · GitHub

(wiki.freebsd.org/BuildingOnNonFreeBSD).

Now I have buildworld and buildkernel in a freebsd/obj but no clue how to actually create a VM image from that.

Is it even possible to create a image which I can just startup in qemu like the one I downloaded in the first gist?

I looked at multiple websites but the only promising ones was this adventurist.me/posts/0283

Title: Build a FreeBSD VM Image Release

Since I replaced make with ./tools/build/make.py in the build process I tried replacing it here to, where it just said, that bmake isn't capable of doing vm-release/install or whatever.

Afterwards I found this slide (papers.freebsd.org/2019/COSCUP/lwhs…%20with%20FreeBSD%20development.pdf)

and on side 17 it is stated to use the makefile again and use installworld, distribution installkernel now instead of build*.

I ran

env MAKEOBJDIRPREFIX=/Users/lockna/freebsd/obj time ../tools/build/make.py -j4 installworld TARGET=arm64 TARGET_ARCH=aarch64 DESTDIR=/Users/lockna/freebsd/destdir

from within the src/release directory.

But essentially I get this error

bmake[2]: "/Users/lockna/freebsd/src/Makefile.inc1" line 574: warning: "MAKEFLAGS= CPUTYPE=dummy /Users/lockna/freebsd/obj/bmake-install/bin/bmake -f /dev/null -m /Users/lockna/freebsd/src/share/mk MK_AUTO_OBJ=no -V CPUTYPE" exited on a signal

bmake[2]: "/Users/lockna/freebsd/src/Makefile.inc1" line 577: CPUTYPE global should be set with ?=.

I tried setting the CPUTYPE to native, aarch64, arm64 (as env variable) or setting it in the Makefile using ?=.

When I set it using env variable it changed accordingly in the command output but when setting it in the Makefile which was used it still was dummy.

So I decided to ask here how someone would create a vm image from self compiled sources on a non bsd host.

Thanks in advance :D

I'm not sure I noticed a question in there...

CPUTYPE comes from make.conf(5) with examples in /usr/share/examples/etc/make.conf on a FreeBSD system, is that what you were asking about?

lockna, have you looked at the release(7) manual page? I think it will add some of the pieces you're missing.

I was just writing that.. :D

development(7) is also a good resource, as is build(7).

Thanks, I tried using release.sh but I remember encounterting it complaining about syntax. Like unexpected token { in line ..

Well, nobody but re@ should need to use release.sh. ;)

yeah, when using the make.py it says "bmake isn't capable of doing release", well vm-release/install since that is what I would want

For reference, docs.freebsd.org/en/articles/freebsd-releng is the actual documentation that belongs with release(7) and release.sh

Title: FreeBSD Release Engineering | FreeBSD Documentation Portal

If someone wanted to do it properly.

hello

It's a little out of date in so far as it references svn, but translating commands isn't the hardest problem, and a patch is very welcome :3

It does contain an example of how to cross-build, though.

for incoming messages like user write or kernel error to my tty, can i redirect this in a way that i listen to a beep and it is saved somewhere else?

micttyl: I'm not sure what you mean by "user write", but it shouldn't be too difficult to monitor /var/log/messages (the system log where kernel errors are stored) via kqueue(2) and trigger beep(1).

any kind of message directly to my tty that ruins my vim display

any

Oh, hm. Maybe that's better handled via a vim plugin generating an audible bell event.

not only vim but also in generel

Have you considered not using ttyv0?

why no ;(

Well, ttyv1 through 8 don't have anything written on them.

hmm

screen or tmux wrap you mean?

No, I mean press alt-F2 to go to ttyv1.

vectors start from 0

The v in ttyvN isn't short for vector, it's short for virtual.

don't i get broad message even if i use other than 0?

system console ;-)

micttyl: broadcast messages are sent to all ttys, yes - but very few messages are broadcast.

Only one I can think of is for shutdown.

debdrup: thank you

No worries.

dch: following your example setting the bhyve vm to use nvme never boots into ipxe; but i can do it if i set it to boot from an iso of ipxe

aah ok its not doing a uefi boot then

but it does seem to load loader.efi okay

it was stuck because i forgot rpcbind on the nfs server

mmmmm

is the order of the pci devices in bhyve correct?

I'll take a look at it next week sometime, maybe it just needs a few updates

i'm using vm-bhyve which hides some of the guts

now it's trying to mount /198.51.100.1/ipxe.efi on the nfs server. (which happens to be the ip of the tftp/nfs server)

wondering if that's a thing baked into ipxe since i don't think i set that filename/root-path anywhere

ERROR: cannot open /boot/lua/loader.lua

ah, finally a usable error message

I'm not sure the best way to describe my issue, but I am running ISC DHCP and some devices on my network seem to 'refuse' the IP address being given out to them.

For example, I see this in my logs: "DHCPOFFER on 192.168.0.70 to d0:13:XX:XX:XX:XX (android-a......) via lagg0"

But certian client never respond back with a DHCPACK, and don't obtain an IP

Do you have a firewall between the DHCP server and the Network?

bad wifi?

you could try to tcpdump and see if your other device is responding to the DHCPOFFER at all...

No firewall, already tried rebooting the access point, but I don't think it's that since most other devices on the same AP work just fine. It's weird, there seems to be no pattern as to which devices are being "ignored"

Hello everyone, I have a program installed that comes with a service in /usr/local/etc/rc.d, I want to change the command the service runs from: /usr/local/bin/program to /usr/local/bin/other-program /usr/local/bin/progrram, but I get an error when trying to start the service, how would I do this?

oh mb, I didn't see that only some devices are affected

yaslam: how are you starting the 'other-program'

i suspect you'll have to do more than rename the script to get it to work this way; so i'd in general advise against it

rtprio: other-program should get started with program as an argument to it, that is all I want

what program/package? keeping these vague isn't helping

and what 'other program'

tuaris: I think you should use tcpdum or wireshark to look in to the traffic on your interface

rtprio: Sorry, I am trying to use proxychains with znc to route znc through tor, I want to launch znc as an argument to proxychains like: /usr/local/bin/proxychains /usr/local/bin/znc

Maybe you can identify the problem this way

Yep, trying to figure out what to filter for, it's noisy

if you dont have any fancy config, just for port 35 and upd, than grep for the MAC-address

you could try putting proxychains in the last line of /usr/local/etc/rc.d/znz like run_rc_command "proxychians $1"

tuaris: something like this 'tcpdump -i re0 -q udp port 53 | grep XX:XX:...'

ah, I was forgetting '-q'.. that's why I was getting nothing :)

rtprio: I tried what you mentioned: ix.io/4lRx and it fails with this: ix.io/4lRy

you might have to rewrite the whole rc script to do this

rtprio: I think I'll just leave it then because I don't know how to write rc scripts at all

rtprio: I'll just launch it manually

tuaris: Good luck with your problem, I am off for now :)

No harm in sharing this tcpdump filtered for just DHCP, right?

I think I narrowed it down. seems like something isn't working correctly with DHCP Snooping on my switch

Yeah, that'll do it.

Dang it :-( No manual page is available for lldb15 from llvm15 package :poop:

... similarly for lldb13

You could check what's different about the Makefiles and see if you can make corrections to the port, then send the port to the maintainer?

I could. I did want to complain about the lack of the manual page

Does Firefox 109.0_1,2 crash for anyone else when trying to use "Network" part of the "Web Developers tools"? (Oh wow, this time could not even open "Web ... Tools")

parv: with another profile (could use `firefox --ProfileManager` to create and start one) or user who has never started firefox too?

Kalten14, Will try.

(sometimes there is something configured wrong after a firefox update)

Kalten, Thanks for the nudge; Firefox has not crashed yet with a new profile.

Next step would be to try with your regular profile with all add-ons disabled

Then enable them one by one if it works without any add-ons.

debdrup, Ok, will do. Thanks

I had Firefox crash on anothe OS. At that time, I did create a new profile after seeing messages from it that it could not profile directory or some file in there. Don't know why did not occur to me to do same in this instance

Nonetheless isolating issue with extensions had not occured to me as debdrup had mentioned