IRC channel logs

<rsmarples>hello!

<rsmarples>I am trying to install hurd on a vm and having a hard time. I have two hosts, NetBSD x86_64 and a macbook neo arm64. Both hosts when trying to boot a hurd netinst cd get to a grub prompt and no matter which installer type i pick i either get black screen or no more input

<rsmarples>this is over vnc. is there any way to debug this easily?

<sam_>rsmarples: what options did you launch qemu with? (or are you using another hypervisor?)

<rsmarples>sam_: qemu-system-x86_64 -rtc base=utc,clock=host -drive file=hurd.qcow2,if=none,id=disk13 -device virtio-blk-pci,drive=disk13 -object rng-random,id=rng13,filename=/dev/urandom -device virtio-rng-pci,rng=rng13 -netdev tap,ifname=tap13,id=net13,script=no,downscript=/home/roy/qemu/qctrl-tapdown -device 'virtio-net-pci,romfile=,netdev=net13,mac=52:54:00:73:00:0d' -vnc ':13,password' -monitor

<rsmarples>'telnet:127.0.0.1:4413,server,nowait' -pidfile /home/roy/qemu/run/qemu-hurd.pid -daemonize -m 4G -smp cpus=2 -cdrom debian-hurd-2025-amd64-NETINST-1.iso -boot d

<rsmarples>sam_: so it seems I need to close my VNC session and reconnect to detect a screen change after booting

<rsmarples>ah, does hurd have support for qemu virt-io devices?

<etno>There are no drivers to support virtio devices in Hurd, AFAIK. It was recently mentioned that rumpkernel could provide those at some point, though.

<azert>indeed it is one of those projects where it should require just to patch a couple of makefiles and it will probably work out of the box. But it is not something very interesting since other driver work on qemu, so nobody will be doing that

<azert>just check how other people use qemu please

<rrq>rsmarples: you might consider using crosshurd-image (in debian/sid) instead of netinstall

<damo22>i think his host is netbsd, its not easy to cross compile everything

<rrq>yes, I missed that one.

<azert>netinst always worked nicely for me

<janneke>i'm trying to debug console-run hanging upon second boot (in guix)

<janneke>i'm suspecting either get_console or mach_open_devstream to hang/be the problem

<janneke>quite naively, i added

<janneke> fprintf (stderr, "20\n");

<janneke> stderr = mach_open_devstream (consdev, "w");

<janneke> fprintf (stderr, "30\n");

<janneke>obviously, upon first boot, which works, i only see the "30" and not the "20"

<janneke>is there any way to add some printout before we have stderr?

<janneke>hmm, it seems that when either the /hurd/ symlink exists, or /dev/console is not a "touched" file but a translator, console-run hangs

<bjc>i did the same thing, and i haven't been able to get output

<bjc>(from fprintf in startup.c before stderr is re-created)

<bjc>i tried changing the conditions in those branches to spin loop, rather than _exit, so i could at least see if the cpu usage pegged

<bjc>think i screwed that up, though, because it only went up to like 10 or 20%, rather than 100%

<bjc>now i'm trying to debug the messages that get sent to startup and i've hit a pretty hard wall

<bjc>in ‘demuxer’, there are a bunch of calls like ‘routine = *_server_routine(inp)’, and i cannot find where those routines are defined

<bjc>grep returns nothing, there's nothing in the headers. i even did a grep through all of /gnu/store looking for ‘server_routine’ and the only matches were library files

<janneke>would it be possible to do something like:

<janneke>fsysopts / --writable

<janneke>rm -f /hurd # a symlink

<janneke>fsysopts / --readonly

<janneke>in console-run.c (ie, from C), even if the file-system is unclean/has errors?

<janneke>(this is all to avoid/work around console-run hanging in guix upon second boot when using -)

<janneke>*when using --readonly with ext2fs.static

<janneke>IWBN to understand the core problem, ie, why console-run hangs when the /hurd symlink exists

<janneke>but yeah

<janneke>okay, so having /dev/urandom, pointing to a read-only /var/lib/random-seed makes console-run hang for me

<janneke>how does debian get around that?

<janneke>okay, so bjc found the solution: using active translators for /dev/*random, it works!

<janneke>thanks all for listening!

<bjc>is it possible to spawn any translator from grub multiboot as long as it's static?

<bjc>i want to try to mount a tmpfs at boot time

<bjc>also: if someone can recommend docs on how grub's multiboot works, i'd appreciate it

<etno>bjc: are you referring to multiboot itself or the way the Hurd uses it ?

<etno> https://www.gnu.org/software/grub/manual/multiboot/multiboot.html

<bjc>both, i guess

<bjc>oh god this is so much more than i need

<gnucode>hello friends!

<gnucode>well I'm hopefully will be running the ext2 journal in an hour or so. *fingers crossed*

<gnucode>well I believe that I am currently running the ext2 journal...or "ext3/4fs"

<gnucode>but I am getting this error on real hardware.

<gnucode>ext2fs: part:5:device:wd0: warning: [JOURNAL] Device flush failed: (os/device) invalid operation

<gnucode>not sure what that means, but I guess I'll find out eventually.

IRC channel logs

2026-05-12.log