gnarface | is there a way to disable pageflipping for the open-source amdgpu driver without restarting xorg? | 02:18 |
---|---|---|
gnarface | i think i've asked this before and the answer was no, but i'm wondering if it has changed | 02:33 |
adhoc | gnarface: why? (uneducated in such things) | 02:36 |
gnarface | attempting to debug a steam remote play bug that mirrors a nvidia bug but i'm too lazy to restart to test it | 02:38 |
gnarface | the nvidia drivers provide a nvidia-settings tool that allows you to toggle AllowFlipping on/off on the fly, amongst other things | 02:39 |
gnarface | i recall hearing there was an equivalent ATi tool back in the catalyst and pre-catalyst days | 02:39 |
gnarface | but only ever for their non-free drivers | 02:39 |
gnarface | i wonder if it could be disabled on a per-program basis with an environment variable like vblank_mode? | 02:41 |
gnarface | the amdgpu man page lists xorg configs but not such environment variables | 02:41 |
adhoc | ah, ok | 02:44 |
* adhoc has not had to deal with this stuff since setting mode lines manually | 02:45 | |
adhoc | erm ... last millenia | 02:46 |
malade_mental | [NoClan]GoAway: feeling old? isn't that normal on a channel aimed at veteran unix admins :P | 08:37 |
malade_mental | (on IRC...) | 08:37 |
Afdal | How do you use apt to check not what libraries a package depends on, but rather what packages a given library is a dependency of? | 08:53 |
jason1234 | hello | 09:09 |
jason1234 | how to install meet.jitsi server on ascii devuan? apt-get????? | 09:10 |
jason1234 | malade_mental: for people that cannot bare systemd killer | 09:11 |
jason1234 | salut | 09:11 |
malade_mental | <^_^> | 09:33 |
avih | hi, i noticed that the minimal-live iso (4.0) doesn't have the "devuan" user at suduers, and that network dhcp is not started automatically. while both are trivial to enable via root login, i think it would help if the iso boots with those too already in effect, similar to the desktop-live image | 09:46 |
avih | those two* | 09:46 |
avih | fwiw, other than the desktop-live having these two, the debian-minimal-live iso also have them enabled out of the box | 09:49 |
avih | also, it would be nice if README.txt at the download dir would list the devuan/root users and their password (because it doesn't login automatically, so it's hard to know which user to choose). e.g. this file https://ftp.nluug.nl/pub/os/Linux/distr/devuan/devuan_chimaera/README.txt | 09:50 |
luna-is-here | jason1234: Did you try to follow this guide <https://jitsi.github.io/handbook/docs/devops-guide/devops-guide-quickstart>? | 09:50 |
FatPhil | I know this is a bit of a bikeshedding question, but in my qemu of a live-USB stick, what's the "neatest" way to ensure my luser account has rw perms on the appropriate /dev/sd? Add me to a group? udev rule? simply sudo chown/chmod in a wrapper script that calls qemu? other better ideas? | 11:03 |
djph | FatPhil: in the VM, or ... what? | 11:28 |
adhoc | /etc/rc.local script ? | 11:33 |
gnarface | FatPhil: make a new group and a udev rule that assigns that group to a particular drive | 12:09 |
gnarface | well, depends on the disk i guess, since you could alternately just add the user's own permission to the mounted directory or some subdirectory thereof, but then you need something else to mount it | 12:11 |
FatPhil | gnarface: it's not for mounting, it's a raw boot image for qemu. Is it possible to teach udev to recognise individual usb keys, so only the ubuntu live-usb gets the special group? | 12:22 |
gnarface | yes | 12:25 |
FatPhil | harumph, qemu's still complaining, so I can't -host cpu, this time it's "could not access KVM module: Permission denied" | 12:28 |
FatPhil | I went into the bios and turned on intel's virtualisation. The kernel has kvm and kvm_intel modules loaded, and /dev/kvm exists, unlike yesterday's attempts. | 12:29 |
gnarface | there's a known issue with some motherboards where you have to also ivrs_ioapic[something... on the kernel command-line, but you do also have to have access to /dev/kvm | 12:30 |
gnarface | some bioses report them wrong so you have to specify the right ones | 12:31 |
FatPhil | annoyingly dmesg doesn't mention kvm at all by name | 12:31 |
FatPhil | grep -i virt yields: "Booting paravirtualized kernel on bare hardware" | 12:33 |
FatPhil | stracing qemu, I couldn't see the syscall that was bringing about the failure. | 12:37 |
FatPhil | yesterday, it was easy, there was an open("/dev/kvm") that failed, but I don't see that today. | 12:37 |
gnarface | you need to look up how to get it to spit out the values you need for ivrs_ioapic[]= | 12:39 |
gnarface | i just forget what the trick was | 12:39 |
gnarface | i think it might show up in POST | 12:40 |
FatPhil | POST's a nightmare on this machine, it seems to hop unpredictably between about 3 different BIOSes. (Lenovo x240) | 12:42 |
brocashelm | on ceres, i noticed apt version 2.3.12 won't show the "yes, do as i say" prompt for removing essential packages anymore (replaced by an error message, instead) | 13:15 |
brocashelm | you'd have to resolve the problems manually | 13:15 |
Kitty | Just installed a new machine, apt-get install nftables, | 18:29 |
Kitty | but there's no /etc/init.d/nftables | 18:30 |
Kitty | what am I missing ? | 18:30 |
Kitty | cp /usr/share/doc/nftables/examples/sysvinit/nftables.init /etc/init.d | 18:40 |
Kitty | update-rc.d nftables defaults | 18:40 |
Kitty | that's what I was missing | 18:40 |
used____ | Why switching to nftables? I am still on iptables | 20:53 |
[NoClan]GoAway | malade_mental: seems I didn't read the channel description. I merely felt nostalgic reading about ancient browsers and the years mentioned with them ;) | 21:09 |
phogg | used____: nftables is the new hotness. If you don't have any issues with the iptables command line interface there's not much reason to stop using it. | 22:13 |
used____ | I know it is the new hotness, I asked why ;) | 22:13 |
phogg | used____: the big headline items are mostly important for people with large rulesets | 22:14 |
phogg | an easier config language is pretty nice, though | 22:14 |
used____ | phogg: you can't scare me, I did cisco and junos configs. Shudder. Not anymore. | 22:16 |
phogg | iptables is pretty nice, IMO, but nft is nicer | 22:16 |
phogg | the down side is I have had iptables stuff memorized for decades | 22:16 |
used____ | Heh yes. Muscle memory. | 22:17 |
phogg | the benefits are slight unless you really need atomic updates and whatnot | 22:17 |
used____ | Does nftables support the no-brainer /etc/init.d/iptables save / flush / load ? | 22:17 |
phogg | it could be set up to work that way. To be fair even iptables doesn't do that, it's a distro choice | 22:18 |
phogg | Red Hat has some weird-ass thing for that | 22:18 |
onefang | Helps to learn the new hotness, coz when the next new hotness arrives, the old hotness might get dropped. Sound and firewall, make up your fucking minds Linux devs. | 22:21 |
phogg | In the case of sound I learned then discarded pulseaudio, because it was solving the problem the wrong way. I made a similar choice with systemd (except a bit less learning before discarding). | 22:21 |
phogg | Fortunately pipewire is 99% compatible with pulseaudio, so if you learned that then you're still mostly OK. | 22:22 |
phogg | the firewall honestly has not changed often. There have only ever been 3 and the new one is entirely compatible with the last one. Pretty good story. | 22:22 |
onefang | I think pipewire is the new audio hotness? I'm still on pulseaudio, it's always "just worked" for me. Is pipewire on Devuan Beowulf or later better? | 22:23 |
phogg | of course I am now suffering from having never really used pulse because I can't figure out how to troubleshoot pipewire-in-pulse-mode issues | 22:24 |
phogg | onefang: not sure, I am tracking pipewire master myself | 22:24 |
phogg | I say tracking, more like playing with it. I have real need for it even today, but it sure is cool. | 22:24 |
onefang | I much prefer sticking to what ever the distro supplies, plus a few extra apt repos for really important stuff. Makes updating everything trivial. | 22:25 |
phogg | disconnecting mechanism and policy? What a crazy concept! Not assuming there will be a GNOME systray applet available to configure it? Madness itself. | 22:25 |
phogg | this has kind of veered into more social chatter so I'll stop there | 22:25 |
used____ | I can't wait for the food fight between systemd controlling shutdown priority and gnome doing the same. | 22:26 |
onefang | That's why I added the "Is pipewire on Devuan Beowulf or later better?" question, to keep it on topic. B-) | 22:26 |
phogg | onefang: ah, well, to keep it topical as well: I would *love* to talk with anyone using pipewire on devuan (any version) | 22:26 |
phogg | most of the docs are sparse or heavily rely on "first start this systemd service" | 22:27 |
phogg | I can get as far as having pipewire running and pulse clients happily connecting and acting like they work, but not as far as getting audio out. | 22:27 |
phogg | My trouble is I don't just want it to work, I want to thoroughly understand how the plumbing works so I can hand-craft a setup tailored to my own tastes. | 22:29 |
phogg | I can walk backwards or forwards through most any OSS, alsa, or jack audio problems and I need to be able to do that in this brave new world. | 22:29 |
FatPhil | worrying thing since I upgraded from debian to devuan a few days ago - every time I leave the laptop idle for a long period (several hours, say, not really happened enough to narrow it down) it locks up and refuses to wake up again. I can't even SSH in to see what's happening. Hardware's still alive, the keyboard lights can be turned on/off, and Fn lock can be toggled. | 22:36 |
FatPhil | I can upgrade to a less ancient devuan, i guess, I'm oldstable now I think. | 22:37 |
phogg | FatPhil: sounds like a hibernation issue to me. I always disable all sleep/hibernate just so I don't have to deal with it. | 22:37 |
phogg | newer kernels tend to be have more fixes in them, so upgrading is always good | 22:37 |
FatPhil | I don't have anything automatic, PM-wise, more than screen blanking. | 22:38 |
FatPhil | The old debian never had any issues at all, alas, so the more modern kernel is being correlated with this brokenness. | 22:39 |
phogg | FatPhil: it's probably doing something automatically which was not happening before | 22:39 |
FatPhil | Hibernation itself mostly works fine. I've been rebooting into a usb-stick a lot recently, and just hibernate devuan when I want to that. (hence my attempts to get qemu running natively) | 22:40 |
rwp | FatPhil, https://en.wikipedia.org/wiki/Magic_SysRq_key might be useful to debug your system after it has locked up. | 22:47 |
rwp | It does seem likely that it is suspending and failing at that time. | 22:48 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!