14:56KitsuWhooa: karolherbst: since it doesn't look like I have too much time to look into how that code works (regarding the suspend issue), and since skeggsb_ doesn't seem to be around, should I try emailing them?
14:57KitsuWhooa: Unless I take this card with me, I'll only have access to it for about another week
14:57karolherbst: ehh skeggsb_ is already aware of the issue
14:57KitsuWhooa: and I don't want to leave this broken
14:57KitsuWhooa: I didn't know that
14:57karolherbst: I talked with skeggsb_ about it, but not sure if there will be a proper fix soonish
14:58karolherbst: but shouldn't be too difficult to figure this out now that we know what's broken
14:59KitsuWhooa: yeah, that's fair enough
19:02matteobin: Hi everyone. I've recently updated my 2009 laptop to Debian 11. I have a GeForce 300M GPU, which was supported by the deprecated nvidia-legacy-340xx-driver package, so now I'm using nouveau. However I've noticed that with nouveau driver my laptop freezes during shutdown or reboot. It stops at "Reached target Power-Off" or "Reached target Reboot". Is there something I can do to solve this problem, maybe a kernel parameter to sort things out?
19:02imirkin: matteobin: what makes you suspect nouveau is connected to this?
19:03imirkin: matteobin: what if you boot with "nouveau.modeset=0"? if the error persists, then nouveau has nothing to do with it
19:03matteobin: If I blacklist nouveau, shutdown and reboot work as usual.
19:03imirkin: ok cool
19:03imirkin: so then nouveau does have something to do with it :)
19:03imirkin: is this a macbook perchance?
19:03matteobin: It seems so... :)
19:04matteobin: No, it's an ASUS laptop.
19:04imirkin: i mean, clearly we're upsetting the EC somehow
19:04imirkin: calling too many ACPI methods, or not enough
19:04matteobin: What is the EC?
19:04imirkin: embedded controller
19:04imirkin: i.e. the thing that runs the system
19:05imirkin: handles things like power
19:05ccr: the one controller to rule them all
19:05imirkin: yes. definitely the seat of all evil.
19:05imirkin: maybe it was originally Evil Controller?
19:05imirkin: but then marketing came around
19:05ccr: "maybe .. something less .. evil?"
19:06karolherbst: could be that there is something crashing that late...
19:06imirkin: sorta like with "ax, bx, cx, dx" on intel
19:06ccr: "no, we need the evil. but something that SOUNDS less evil."
19:06karolherbst: wondering if netconsole _could_ catch it?
19:06imirkin: i'm a bit curious what produces such "Reached target xxxx" messages
19:07imirkin: so the kernel is in no way powered off at that point then
19:07karolherbst: yeah.. at this point userspace shouldn't exist anymore
19:07imirkin: if some rando userspace process is printing stuff
19:07karolherbst: imirkin: not after that message
19:07imirkin: karolherbst: ah ok
19:07karolherbst: that's like the third or fourth last message you see
19:07karolherbst: on a console
19:07imirkin:never used systemd
19:08imirkin:hopes to never use systemd
19:08karolherbst: it essentially means "I've done everything I was planning to do"
19:08karolherbst: imirkin: poor you
19:08matteobin: I can tell you that at that point the HD is not mounted anymore.
19:09matteobin: I don't get any orphaned node messages at the following startup.
19:09karolherbst: matteobin: I'd check if you see anything through netconsole, but that requires a second computer
19:09imirkin: yeah, but kernel has a lot of stuff to do
19:09imirkin: nouveau might be slow to unbind / deadlock or whatever
19:09karolherbst: imirkin: yeah.. probably
19:09karolherbst: or nouveau crashing ¯\_(ツ)_/¯
19:09karolherbst: but a crash would be shown...
19:10matteobin: How should I procede?
19:10karolherbst: matteobin: how long did you wait btw?
19:10imirkin: it definitely worked at one point or another
19:10matteobin: I've waited for a long time but it simply sits there.
19:10karolherbst: waiting like 2 minutes usually should hit messages
19:10matteobin: More than twenty minutes.
19:11karolherbst: matteobin: maybe boot with "debug" ?
19:11karolherbst: it could print additional things on shutdown/reboot
19:11matteobin: I'll try it right now!
19:15matteobin: Ok, now it stops at "Stopping disk", after "Synchronizing SCSI cache" (sda).
19:18matteobin: I have a second computer with Debian too. How does netconsole work?
19:18karolherbst: matteobin: very annoying to set up
19:18karolherbst: seems like debian scripted it
19:19karolherbst: and on the target you listen with netcat on the same UDP port
19:19karolherbst: but it's a mess to set it up correctly
19:20matteobin: So I shouldn't do it?
19:20karolherbst: well you should, but it's just annoying to get it right
19:21karolherbst: on the second machine: nc -klu -p 6666 --sh-exec "cat > /proc/$$/fd/1"
19:21karolherbst: and on your laptop: netconsole-setup +6666@ip_of_your_second_machine
19:21karolherbst: this might just work
19:21karolherbst: and then you should start seeing a log on your second machine
19:22matteobin: I'll do it. Be right back. Thank you very much for the help!
19:22karolherbst: that nc command is a bit hacked up so it won't break after one connection
19:25matteobin: Should I use the debug param for the kernel with netconsole?
19:30karolherbst: shouldn't hurt
19:30karolherbst: you might have to boot with loglevel=8 though
19:30karolherbst: otherwise netconsole won't be seending everything or antyhing over...
19:31karolherbst: (one of the reasons why netconsole is so annoying to set up)
19:31karolherbst: or do a "dmesg -n 8" might be good enough as well?
19:31karolherbst: anyway.. loglevel and debug should be good
19:33matteobin: Do I need to open some ports on the listening machine?
19:35karolherbst: uhm.. yeah you might have to
19:35karolherbst: if you have a firewall running
19:36matteobin: UDP on input, right?
19:40matteobin: Done, got the log. What should I do with it?
19:41karolherbst: is there any error at the end?
19:42karolherbst: or well.. anything which could indicate what is wrong?
19:42karolherbst: but you could pastebin it somewhere
19:46matteobin: I didn't notice anything significative...
19:46karolherbst: matteobin: ohh, I meant you have to shutdown/reboot after setting it up :)
19:47karolherbst: so you get all the logs from whatever your machine is doing there
19:47matteobin: I did it. Now the laptop is freezed at Stopping disk. This is all I got.
19:48karolherbst: matteobin: huh.. and that's all the log you got?
19:48matteobin: Yep, I've added debug and loglevel=8 in grub too.
19:49karolherbst: I bet netconsole-setup is doing something stupid den
19:49karolherbst: ehhh crap
19:49karolherbst: I am silly
19:49karolherbst: matteobin: that laptop is connected via wifi, isn't it?
19:49matteobin: Nope, cable.
19:50matteobin: The man page says it does not work via Wi-Fi.
19:50karolherbst: well.. it actually does
19:50karolherbst: but I suspecting the script to be overly stupid
19:50karolherbst: *I am
19:50karolherbst: sooo.. mhh
19:50karolherbst: let's try it outselfes
19:52karolherbst: boot with "netconsole=@source_ip/source_eth_device,6666@target_ip fill out the information and you should see the messages coming in after booting
20:03matteobin: It does not work. Should I put netconsole=@source_ip/source_eth_device,6666@target_ip in grub after loglevel=8?
20:09karolherbst: matteobin: anything in dmesg about it?
20:11matteobin: I've never used dmesg. What should I do "dmesg | less"?
20:12karolherbst: dmesg | grep netconsole
20:16matteobin: I got the following error: couldn't parse config at 'target_ip'. Maybe I need to use quotes after netconsole="config"?
20:16karolherbst: you need to fill in your IP addresses and evrything
20:17karolherbst: ohh wait
20:17karolherbst: maybe I forgot something..
20:17matteobin: Of course, I've put my real IP. target_ip is a placeholder.
20:17karolherbst: you have to put a / at the end..
20:23matteobin: Wait, I'm trying it...
20:26matteobin: It works!
20:29matteobin: This is after login, issuing shutdown from LXDE menu: https://bin.disroot.org/?3efcbc2bb5c48101#51hv1ZfmNH8Ueumu7NrjYaZ3x5q83uKmxRbizsE6FyVC
20:41karolherbst: mhh, strange
20:41karolherbst: does anything happen if you wait a while?
20:42matteobin: Nothing happens, I kept waiting until five minutes ago.
20:43karolherbst: I could imagine that userspace is blocking something
20:47matteobin: So what should I do now?
20:52matteobin: Be right back.
20:54karolherbst: matteobin: mhh yeah.... another thing to try would be to check if ssh is still running
20:55karolherbst: and figure out if there are silly processes hanging
20:55karolherbst: but usually sshd is already killed
20:55karolherbst: that's really a domain I don't really know how to debug if the shutdown process is stuck somewhere..
20:55karolherbst: normally systemd starts killing shit
20:55karolherbst: but it should also display it
20:57matteobin: Uhm... should I ask somewhere else, maybe another channel?
20:58karolherbst: mhh maybe systemd folks would know how to help
21:03matteobin: I'll try in the systemd channel. Thank you for everything, you've been very very kind.