IRC channel logs

<star_sapphire>thats awesome you got the x220 to boot hurd. one of these days soon im going to try on mine as well

<damo22>yeah libacpica is something i rolled

<damo22>it contains the acpica code from upstream + hurd port

<damo22>its difficult to upstream because all the other ports bake it into monolithic kernel

<damo22>whereas we use it as a userspace lib

<janneke>damo22: nice; yes i figured/saw something like that

<janneke>it was a bit of a puzzle with all these *acpi* packages and initially i was a bit confused seing an almost empty debian/rules, but no makefile :)

<janneke>i wrote a meme for guix to apply debian's patches which works nicely here too

<janneke>star_sapphire: thanks, yeah pretty happy with it!

<janneke>the recipe for rumpnet -- https://www.gnu.org/software/hurd/hurd/rump/rumpnet.html

<janneke>gives

<janneke>root@debian:~# settrans -fgap /dev/rumpnet /hurd/rumpnet

<janneke>settrans: /dev/rumpnet: No such file or directory

<janneke>ACTION tries: mknod /dev/rumpnet --mode 660 c 0 0

<janneke>oh, and now:

<janneke>settrans: /dev/wm0: No such file or directory

<damo22>touch /dev/wm0

<janneke>damo22: OK, and what about /dev/rumpnet?

<damo22>the node has to exist

<janneke>yeab but is: "mknod /dev/rumpnet --mode 660 c 0 0" OK?

<damo22>im not sure, i only ever used touch to create an empty node

<janneke>OK

<damo22>the error you saw was because /dev/wm0 did not exist on the filesystem and you tried to attach a translator to it

<janneke>ok

<damo22>are you trying to run rumpnet on x220?

<janneke>yes; ifconfig gives: no devices found (or something)

<janneke>on the x60, i think i got at least an eth0 there that could be setup

<damo22>inetutils-ifconfig?

<janneke>yes

<damo22>it could work with rumpnet

<damo22>lspci -nn | grep -i net

<damo22>probably as root

<janneke>crap, got no lspci; bbl :)

<damo22>libpci?

<janneke>i just didn't install it

<bjc>is there a list of how many drivers rumpnet contains?

<damo22>1

<bjc>oh

<damo22>well one family

<bjc>that explains the lack of virtio =)

<damo22>check wm(4) on netbsd man

<bjc>neat. thanks

<damo22> https://man.netbsd.org/wm.4

<bjc>i'll have to see if qemu's e10k works. seems like it should

<janneke>re lspci (run from linux, but yeah)

<janneke>00:19.0 Ethernet controller [0200]: Intel Corporation 82579LM Gigabit Network Connection (Lewisville) [8086:1502] (rev 04)

<janneke>03:00.0 Network controller [0280]: Intel Corporation Centrino Advanced-N 6205 [Taylor Peak] [8086:0085] (rev 34)

<bjc>how can i capture console output from settrans?

<bjc>i don't see anything in /var/log post-boot

<bjc>in /hurd/console --help it says i can set the number of scrollback lines, so presumably there's a key combo that does it

<bjc>the answer appears to be lshift+arrows or rshift+pgup/dn, except the latter doesn't work for me

<bjc>sorry, left-alt+arrows

<bjc>and it doesn't seem to work anyway

<bjc>screen scrolls, but the scrollback buffer is wrong

<bjc>the advice in https://www.gnu.org/software/hurd/hurd/debugging/translator/capturing_stdout_and_stderr.html isn't working for me, either

<bjc>though maybe i'm capturing the wrong translator. is the pci probe output when starting rumpnet from rumpnet itself, or is it coming from pci-arbiter (or something else)?

<bjc>and it won't work for pci-arbiter anyway, since there's no /bin/sh at that point in the process

<bjc>ugh

<bjc>well, i got output on the serial console, so that'll work

<bjc>interesting. rumpnet isn't seeing the device at all

<bjc>lspci has: 01:00.0 Ethernet controller: Intel Corporation 82545EM Gigabit…

<bjc>but nothing at all from rumpnet

<bjc>i wonder if it has to due with its memory address being 32-bit rather than 64

<janneke>damo22: as the x220's nic is intel (see above), i guess that's increasing my chances of getting rumpnet to work eventually?

<damo22>it should work with intel yes

<janneke>grand! then i/we just have to find the rigth memes for the x200

<bjc>fwiw, i have yet to be able to start rumpnet with any intel driver that qemu supports, even though wm(4) claims support

<bjc>not 82540em, not 82545em, and not 82576

<bjc>they're not even showing up in the rumpnet logs, though they do show in lspci

<janneke>*x220

<bjc>yeah, none of the nics in the intersection of wm support and qemu support work

IRC channel logs

2026-05-10.log