| chomwitt | qqute | 00:34 |
|---|---|---|
| chomwitt | sorry | 00:34 |
| systemdlete | I'm having trouble getting a key (using apt-key or gpg) on refracta: It hangs. Whereas, on regular devuan, I have no hang; it comes right back and adds the key. | 01:44 |
| systemdlete | This is what I see when I run gpg: https://pastebin.com/kwcn8RzY | 01:46 |
| systemdlete | I did some web searching and got a hit on a "solution," but I could not figure out how to map that to my situation. | 01:46 |
| systemdlete | https://unix.stackexchange.com/questions/668976/gpg-agent-hanging-when-trying-to-access-private-keys, but it deals with systemd... | 01:47 |
| systemdlete | https://pastebin.com/8E44CQjD (when it has completed/failed) | 01:48 |
| systemdlete | (and the hang was much longer than 5s) | 01:48 |
| systemdlete | Some pages claim that merely re-installing gpg does the trick, but that sounds... a bit microsoftish to me. | 01:49 |
| systemdlete | Well, tried reinstalling gpg, but still no joy. | 01:54 |
| rrq | well in my mind "gpg" is not merely a program but a whole infrastructure of interdependent utilities that together brings serious convolution to encryption and keys and stuff, so a reinstall may well make a difference... | 01:56 |
| systemdlete | rrq: I just tried re-installing gpg. No help there, sorry... | 01:57 |
| systemdlete | ping fsmithred | 01:57 |
| fsmithred | what's up? | 01:58 |
| systemdlete | This only seems to happen on one system, which happens to be refracta. | 01:58 |
| systemdlete | thanks, fsmithred. Yeah, having a problem retrieving keys | 01:58 |
| systemdlete | Is there some difference in the way Refracta installs operate (vs regular devuan) | 01:58 |
| systemdlete | (I would not think so but Idk) | 01:58 |
| systemdlete | See https://pastebin.com/8E44CQjD | 01:59 |
| fsmithred | no difference that I know of | 02:00 |
| systemdlete | And, granted, maybe I did something wrong at some point. | 02:00 |
| systemdlete | ok, just double-checking. | 02:00 |
| fsmithred | I recall seeing that error about dirmgr or whatever | 02:00 |
| fsmithred | few years ago | 02:00 |
| fsmithred | I think it means something is missing, but I don't remember what | 02:00 |
| systemdlete | dirmngr is definitely installed, afaict | 02:01 |
| systemdlete | It seems to be installed with the rest of gpg | 02:01 |
| rrq | check ~/.gnupg/dirmngr.conf | 02:03 |
| systemdlete | there is no such file there | 02:04 |
| rrq | there's ~/.gnupg/ ? | 02:04 |
| systemdlete | yes, but no such file "dirmngr.conf" | 02:05 |
| fsmithred | I don't have that file, either | 02:06 |
| rrq | hmm is there ~/.gnupg/gpg.conf | 02:06 |
| systemdlete | I moved ~/.gnupg to a save directory and re-ran gpg | 02:06 |
| systemdlete | it is hanging again | 02:06 |
| rrq | the reason is probabky that it wants to locat a server that no longer exists | 02:07 |
| systemdlete | but rrq, if I copy-paste the command to another system, it works fine. | 02:07 |
| rrq | run gpgconf to see if it says | 02:07 |
| rrq | like: gpgconf --list-options dirmngr | 02:09 |
| systemdlete | says the same thing it says on another system where it does work fine | 02:09 |
| systemdlete | (I diffed the output) | 02:09 |
| systemdlete | hmmm | 02:09 |
| systemdlete | with those args it hangs | 02:09 |
| systemdlete | "gpgconf --list-options dirmngr" hangs, is what I'm saying | 02:10 |
| systemdlete | I wonder... maybe a missing library or some data file it needs? Maybe in a lost dependency package? | 02:10 |
| fsmithred | what about 'ps ax |grep gpg'? | 02:10 |
| systemdlete | (I'd hope not) | 02:10 |
| systemdlete | fsmithred: The gpg-agent is running | 02:11 |
| fsmithred | how many? | 02:11 |
| systemdlete | just one | 02:11 |
| rrq | ok; I'd start strace-ing now ... :) but I'm weird | 02:12 |
| systemdlete | wait, while I am running gpg, or when I am not running it? | 02:12 |
| systemdlete | good idea! | 02:12 |
| systemdlete | so it looks like it clones itself and then tries to read from an incore pipe | 02:13 |
| rrq | use -f to follow | 02:13 |
| rrq | follow forks | 02:13 |
| systemdlete | aha! | 02:13 |
| systemdlete | it opens /etc/gcrypt/hwf.den which does not exist! | 02:13 |
| systemdlete | sorry, got cut off. | 02:14 |
| systemdlete | /etc/gcrypt/hwf.deny | 02:14 |
| rrq | no /etc/gcrypt here | 02:14 |
| rrq | apparently gpg wants sockets in /run/user/1000/gnupg/ | 02:18 |
| systemdlete | https://pastebin.com/KLmcNLWN | 02:18 |
| systemdlete | (last two lines is me hitting control-c) | 02:18 |
| systemdlete | but this is for root, the key is for a repo | 02:19 |
| systemdlete | (sorry, should have mentioned this sooner) | 02:20 |
| fsmithred | shouldn't you be using apt-key add for that? | 02:22 |
| systemdlete | Well, I have tried it that way also, but it just turns around and runs gpg anyway... | 02:22 |
| systemdlete | same hang | 02:23 |
| systemdlete | and I'd aver probably same place it is hanging | 02:23 |
| systemdlete | also, using apt-key add warns that it is defecated and to use gpg instead | 02:23 |
| rrq | hmm my gpgconf starts /usr/bin/pinentry after having looked for hwf.deny | 02:23 |
| rrq | and I have pinentry-curses installed .. there are a couple of pinentry-* options | 02:24 |
| systemdlete | but does it hang? | 02:24 |
| rrq | nope | 02:25 |
| systemdlete | :) for you, :( for me | 02:25 |
| rrq | apparently the gpg programs forks in some way that prohibits stracing ... | 02:25 |
| systemdlete | oh... hold on | 02:26 |
| systemdlete | there is a switch for that... | 02:26 |
| rrq | I don't get that fork happening | 02:26 |
| systemdlete | it clones, not forks | 02:26 |
| rrq | that's -f | 02:27 |
| rrq | perhaps "pstree -ap" would tell about the child | 02:28 |
| * systemdlete opens a second root window... | 02:29 | |
| rrq | otherwise use lsof to follow up the socket (at /proc/$pid/fs/3) | 02:30 |
| systemdlete | fu for filtering just the process of interest please? | 02:30 |
| rrq | I tent to use "| less " and scroll | 02:31 |
| rrq | tend | 02:31 |
| systemdlete | oh, just add pid on cmd line | 02:31 |
| rrq | thanks. TIL :) | 02:31 |
| systemdlete | so, anyway, just 2 lines: Parent is gpgconf and child is dirmngr | 02:32 |
| systemdlete | no other news | 02:32 |
| rrq | hmm and nothing from lsof regarding the socket? | 02:33 |
| systemdlete | uhm... and now, it finally terminates and spits out a long list of info | 02:33 |
| systemdlete | ! | 02:33 |
| rrq | (seems weak programming to clone a child and just ignore it failing) | 02:33 |
| systemdlete | >.< | 02:33 |
| systemdlete | the program ended before I could trap that for you... hold on | 02:34 |
| systemdlete | no such file... ??? | 02:34 |
| systemdlete | /proc/pid/fs/3 does not exist | 02:35 |
| systemdlete | (where "pid" is the actual pid number, ok?) | 02:35 |
| systemdlete | and I double-checked the pid | 02:35 |
| rrq | ah $pid would be the pid of the parent (gpgconf) | 02:35 |
| rrq | it creates a pipe for fd [3,4] then clones, then reads 3 | 02:36 |
| systemdlete | so the child then? | 02:36 |
| rrq | yes the child if any would have that socket as fd 4 | 02:36 |
| rrq | the "write end" of the pipe | 02:37 |
| systemdlete | "fs" or "fd"... I think you mean fd | 02:37 |
| rrq | yes. (as per line 58 in your paste) | 02:37 |
| systemdlete | (you had put "fs" above) | 02:38 |
| systemdlete | (but that's ok, just threw me a bit) | 02:38 |
| rrq | right I tries to blame my keyboard for reordering key tops now and then :) | 02:38 |
| systemdlete | keyboards... they are always up to no good | 02:39 |
| systemdlete | But I do love my happy hacker! | 02:39 |
| systemdlete | there are probably much better options nowadays, but I've had mine for about 5 years, and before that, I had one for about 20 years. The keys were demolished! | 02:40 |
| systemdlete | now I am trying gpg again, to see what it does. | 02:41 |
| systemdlete | ok, it finally gets connection to dirmngr, but then the last line is: | 02:41 |
| systemdlete | gpg: keyserver receive failed: No name | 02:41 |
| rrq | my strace is same as yours up to line 58, where min starts pinentry while yours clone something | 02:42 |
| systemdlete | rrq: Are you on chimaera? | 02:42 |
| rrq | yes | 02:42 |
| systemdlete | and is it a devuan install or a refracta install (I mean fsmithred's thing) | 02:42 |
| systemdlete | distro, whatever | 02:42 |
| systemdlete | This one instance where it is hanging happens to be refracta, though I doubt that's the smoking gun | 02:43 |
| rrq | no I have an upgrade story | 02:43 |
| systemdlete | oh | 02:43 |
| rrq | plain devuan | 02:43 |
| systemdlete | beowulf->chimaera | 02:43 |
| rrq | yes.. and maybe not totally clean as such | 02:44 |
| systemdlete | mine should be so unclean... | 02:44 |
| systemdlete | at least it might be working now | 02:44 |
| rrq | hmm running s root I get cloning happening | 02:45 |
| rrq | it's usr/bin/dirmngr that gets started | 02:46 |
| rrq | with argument --gpgconf-list ... so you could check what "/usr/bin/dirmngr --gpgconf-list" does .. as root | 02:47 |
| systemdlete | crud. you won't believe this | 02:48 |
| systemdlete | I tried mit's keyserver, and it succeeds | 02:48 |
| systemdlete | but | 02:48 |
| systemdlete | but | 02:48 |
| systemdlete | it works fine on other systems here!!! | 02:48 |
| systemdlete | I wonder if the gpg config for refracta might be a bit diffrent than plain vanilla devuan | 02:48 |
| systemdlete | (which is why I asked fsmithred about this) | 02:49 |
| rrq | yes he'll have had to deal with it before... sometimes he doesn;t remember everything upfront though :) | 02:50 |
| systemdlete | funny, it still hangs using apt-key, but gpg came right back | 02:51 |
| rrq | any /root/gnupg/*.conf ? | 02:54 |
| * systemdlete thinks good grief... | 02:54 | |
| systemdlete | thank you for the assist, rrq | 02:54 |
| systemdlete | yeah, a whole slew of files | 02:55 |
| systemdlete | and subdirs | 02:55 |
| systemdlete | it looks sane enough to me | 02:55 |
| systemdlete | so Refracta's gpg does not like ubuntu's key server, but it is good with MIT's. | 02:56 |
| rrq | my /root/.gnupg is without any *.conf but has othere cruft | 02:56 |
| systemdlete | no *.conf in mine either | 02:56 |
| rrq | eeh just saw: isn't it keys.openpgp.org rathe tnan keys.openpgp.com | 02:59 |
| systemdlete | maybe | 03:00 |
| systemdlete | sometimes my fingers are too fast for my brain. | 03:00 |
| systemdlete | or get their wiring mixed up, sorry about that, rrq. | 03:00 |
| rrq | .. like my teenager's lips :) | 03:00 |
| systemdlete | I think the main point, though, is that Refracta apparently does not like ubuntu's keyserver. So I will use mit's from now on. Beside's it's a few less chars to type. | 03:01 |
| systemdlete | and it is so easy to mix up pgp with gpg | 03:01 |
| systemdlete | as far as your teenager, I cannot comment. I do not have kids. | 03:01 |
| systemdlete | But I was one once. | 03:02 |
| systemdlete | and sometimes, I still feel like I am. Even at 63. | 03:02 |
| systemdlete | :D | 03:02 |
| systemdlete | I can still "paint a room" if you get my meaning... | 03:02 |
| rrq | :) | 03:03 |
| systemdlete | :) | 03:03 |
| rrq | keys.openpgp.org does have the key but it's without UID so it doesn;t get donwloaded | 03:04 |
| rrq | keyserver.ubuntu.org rather responds but "indicates failure" instead of serving it | 03:05 |
| systemdlete | Maybe fsmithred decided to disable the ubuntu keyserver for symbolic reasons... | 03:05 |
| systemdlete | y'know. Team Devuan here. | 03:05 |
| systemdlete | For which I cannot say I'd have any disdain | 03:05 |
| systemdlete | Y'all probably get tired of hearing this, but I am so grateful for the friendly and constructive help I get here. | 03:07 |
| systemdlete | I totally mean that. | 03:08 |
| buZz | oxoxo | 03:08 |
| systemdlete | you too buZz | 03:08 |
| buZz | :P | 03:08 |
| buZz | we all make devuan awesome | 03:08 |
| systemdlete | lotsa love right back at you man! | 03:08 |
| systemdlete | (you are right!) | 03:08 |
| systemdlete | and now... I actually am trying to get something done here. Now that the software is upgraded, I can return to what I was working on before that whole yuck-up. | 03:09 |
| buZz | hf :) | 03:09 |
| rrq | cheers (rrq has to feed the cat) | 03:10 |
| fsmithred | systemdlete, I don't do anything to change gnupg in refracta | 03:19 |
| fsmithred | btw, I haven't been able to use MIT's gpg for a couple of years. | 03:27 |
| fsmithred | and I use ubuntu's keyserver | 03:27 |
| systemdlete | not sure why, but ubuntu's did not work for me | 03:28 |
| systemdlete | (on refracta, I mean. It works on all my vanilla devuan chimaeras) | 03:29 |
| fsmithred | weird | 03:29 |
| fsmithred | I just specify it on the command line | 03:30 |
| fsmithred | my own gpg.conf has a list of commented-out keyservers that failed me | 03:30 |
| systemdlete | where is this gpg.conf file? I don't see one under /etc | 04:08 |
| chomwitt | Goodmorning. Investigating an issue of boot messages thrown at my tty1 login terminal (no gui login manager installed) two messages were related to elogind and avahi both pulled in from my installation/tasksel choice of ConsoleProductivity | 08:15 |
| chomwitt | so i regret for being curious... | 08:15 |
| chomwitt | now i purged elogind and avahi but i still see one message in my tty1 | 08:16 |
| chomwitt | now i see ok dbus(piq:X) and X is process dbus-deamon --system --nofork --nopidfile | 08:16 |
| chomwitt | my first issue with ConsoleProductivity was not being able to install xorg! | 08:17 |
| chomwitt | also when using tasksel again to undo that selection it seemed that it remove a lot of packages but not elogind and avahi! | 08:18 |
| chomwitt | so i ponder the option to do a reinstallation being curious this time.. | 08:19 |
| chomwitt | so it seems to me i choosed by curiousity a less travel installation path (desktopless , consoleprocuctivity) and then installed xorg and awesomewm. That sums up that troubled Daedalus installation of mine. | 08:22 |
| chomwitt | speaking of tasksel shouldnt it help to have a help toggle with into tips ? | 08:26 |
| brocashelm | using exec startxfce4 --with-ck-launch, how to get policykit/pkexec (such as with synaptic as a regular user) to work without elogind (consolekit replacing it)? tried lxpolkit and mate-polkit, but neither did the trick and instead just froze my session | 10:20 |
| _Mystified | hi guys does testing come in a server edition. Would it be wiser to stick to safety of chimaera, for home media server | 11:58 |
| djph | _Mystified: I've got daedalus on a server or two (as "testing", but it's not misbehaved on me) | 12:18 |
| _Mystified | fantastic, which iso do I need, are you able to share the link please, I'm struggling to find it. | 12:20 |
| _Mystified | thanks | 12:20 |
| rrq | which mirror do you prefer? | 12:22 |
| rrq | https://www.devuan.org/get-devuan might be a good start | 12:24 |
| brocashelm | never mind what i asked earlier | 12:27 |
| brocashelm | forget about it | 12:27 |
| fsmithred | brocashelm, there was a discussion on d1g a few years ago about running xfce without xfce-session. That might help. I'll try to find it. I was on that thread. | 12:32 |
| fsmithred | oh, I was scrolled up and didn't see you were here now. | 12:33 |
| brocashelm | i tried that and still had problems with pkexec/xfce4-session-logout (only log out was available), even with --with-ck-launch appended to startxfce4 | 12:34 |
| brocashelm | removing elogind also kills kde packages | 12:35 |
| brocashelm | would using lxdm restore polkit functionality without elogind running? | 12:37 |
| fsmithred | I haven't tried consolekit in a long time, but last time I did, I couldn't get everything to work | 12:38 |
| fsmithred | not sure about lxdm. Maybe. | 12:38 |
| brocashelm | for synaptic, to get around the pkexec screen freeze, using lxqt-sudo in its place helped mitigate | 12:38 |
| fsmithred | that's one of the ones that work without dbus I think. | 12:38 |
| fsmithred | there's an lxqt-sudo command? | 12:39 |
| brocashelm | correct | 12:40 |
| brocashelm | and according to this: https://www.devuan.org/os/documentation/dev1fanboy/en/dbus-free-software.html | 12:40 |
| brocashelm | ...only nodm, ldm, xdm, and wdm are dbus-free | 12:41 |
| brocashelm | lxqt-sudo or lxsu work | 12:41 |
| fsmithred | I think I tested those recently and reported to you. | 12:41 |
| brocashelm | e.g. lxsu /usr/sbin/synaptic | 12:41 |
| brocashelm | that bypasses the pkexec bug that freezes your graphical session if elogind is not running | 12:42 |
| brocashelm | i was holding onto the deprecated gksu until i got pkexec to work, but now i could replace it with lxqt-sudo/lxsu for simplicity's sake | 12:42 |
| RhineDevil | I'm getting problems with drivers/video stack. mpv video acceleration works properly but if I run software using OpenGL I can't see videos at the same time (even if the game is paused), also if I see a video AFTER I used a software using OpenGL video acceleration crashes. What to do? | 15:08 |
| gnarface | RhineDevil: which video card? | 15:42 |
| gnarface | and which release of devuan? | 15:42 |
| RhineDevil | gnarface: testing, Radeon HD 8210 | 15:46 |
| gnarface | did you install firmware-amd-graphics? | 15:47 |
| RhineDevil | Besides it's not mpv in itself that crashes, the entire video stack crashes | 15:47 |
| RhineDevil | gnarface: yes | 15:47 |
| gnarface | run "dpkg -l |grep mesa" and compare the package list to this: https://paste.debian.net/1279292/ | 15:48 |
| RhineDevil | gnarface: libd3dadapter9-mesa isn't installed | 15:53 |
| gnarface | that's the only one? hmm, you may be hitting a actual driver glitch | 15:53 |
| gnarface | i don't think that one is important for this | 15:53 |
| gnarface | probably won't hurt to install just in case, but i think you've got a bug | 15:54 |
| gnarface | the mpv version you're using, it's not from another distro, right? it's from devuan? | 15:54 |
| gnarface | got any packages mixed in from 3rd party repos or other devuan versions? | 15:54 |
| gnarface | if you have time and some spare bootable media you should download a chimaera (stable) version of the live image and test that to see if it has the same bug or not | 15:55 |
| gnarface | RhineDevil: ^ | 15:56 |
| RhineDevil | gnarface: libopenvg1-mesa isn't installed too | 15:59 |
| RhineDevil | Yeah it's from devuan | 15:59 |
| gnarface | well try installing those two packages and give it a reboot and see if that magically fixes anything but don't get your hopes up. | 16:00 |
| RhineDevil | Oh also, I have mesa-opencl-icd installed but you haven't | 16:00 |
| RhineDevil | And you have libosmesa6-dev but I haven't | 16:00 |
| gnarface | mpv doesn't use opencl does it? | 16:01 |
| gnarface | libosmesa6-dev shouldn't matter unless you're compiling stuff | 16:01 |
| RhineDevil | Yeah | 16:01 |
| RhineDevil | mesa va vdpau and such stuff is identical in our setups | 16:02 |
| gnarface | i have more recent card though | 16:02 |
| gnarface | try the chimaera live cd. we need to figure out if this broke recently | 16:02 |
| RhineDevil | Yeah guessed | 16:02 |
| gnarface | which window manager, by the way? | 16:03 |
| gnarface | is it using xorg or wayland? | 16:03 |
| RhineDevil | gnarface: if I have to be honest, I think it's all about the reworking on AMD drivers for increasing OpenCL and Vulkan support | 16:03 |
| RhineDevil | xorg | 16:04 |
| gnarface | i wonder if your card can still use the pre-amdgpu drivers, and if they're still available in testing | 16:04 |
| RhineDevil | *increasing support on my card | 16:04 |
| gnarface | RhineDevil: actually, yea, i have an idea: check the xorg log and see which driver it's using. if it's actually using the amdgpu driver maybe you can try forcing radeon in the xorg config | 16:12 |
| gnarface | try the chimera live image too, but i figure that can give you something else to test while it's downloading | 16:13 |
| gnarface | i'm not sure exactly where your card lies in the timeline, but there were a couple generations of their hardware that could use either driver but ymmv on which one is actually better | 16:14 |
| RhineDevil | gnarface: which lines should I check for? | 16:15 |
| gnarface | the ones where it's picking a driver, but it'll try a few dumb ones usually before settling on the right one | 16:15 |
| RhineDevil | By the way I haven't enough free space for testing stuff I think | 16:15 |
| gnarface | it might be easier for me to look at it, just put it up at paste.debian.net | 16:15 |
| RhineDevil | gnarface: I'd prefer not, what does it say when it loads drivers? | 16:16 |
| gnarface | well, the output is kinda confusing these days, but what you can see it doing if you look carefully, is you see it loading every driver module at once, then failing on them one at a time until it finds one that works, then it unloads all the others and the one you're left with is the one it picked | 16:16 |
| gnarface | you'll see groups of lines starting with "LoadModule" then "Module" then "Loading" but like i said they're all loading at once so you'll see those groups of 3 split up by each other rather than all in a clean order | 16:18 |
| gnarface | unless the problem is much worse than we expected, it should be settling either on radeon or amdgpu | 16:19 |
| gnarface | if it ends up settling on fbdevhw or something like that then that's definitely off the railks | 16:20 |
| gnarface | *off the rails | 16:20 |
| gnarface | but whatever it settles on we can force it to use something else in the config | 16:21 |
| RhineDevil | gnarface: it loaded radeon driver already, ati driver was available too | 16:23 |
| gnarface | hmm | 16:24 |
| gnarface | so incidentally that means you're not even using amdgpu, and i think that means you don't even need firmware-amd-graphics | 16:24 |
| gnarface | i don't know for sure whether that's still the case, but you might want to try pitching it overboard to see whether there's any change | 16:25 |
| RhineDevil | gnarface: firmware-amd-graphics is installed already | 16:25 |
| gnarface | yea but you shouldn't need it | 16:25 |
| gnarface | and if something is using parts of it maybe that's causing this bug | 16:25 |
| RhineDevil | That's possible | 16:26 |
| gnarface | i could be wrong here, like i said, but it was my understanding that the radeon and ati drivers were fully free, and only the amdgpu driver needed the firmware | 16:27 |
| RhineDevil | No idea here | 16:28 |
| gnarface | but in a situation where they are using this type of "throw it all at the wall to see which parts stick" method of brute-force auto-detection across a dozen or so generations of cards, some of which are supported by multiple drivers... eh, i think things could go wrong, so it's another thing worth testing | 16:28 |
| RhineDevil | So should I remove firmware-amd-graphics? | 16:28 |
| gnarface | yea just purge it and reboot then check the xorg log again, but be ready to install it again if that causes some unexpected massive regression | 16:29 |
| gnarface | if we're really lucky though maybe that'll just actually fix the problem | 16:29 |
| gnarface | if not, i think the only recourse is to report the bug upstream | 16:29 |
| RhineDevil | Do you think mesa-opencl-icd needs firmware-amd-graphics in order to work? | 16:30 |
| gnarface | i don't know, but do you think your card even supports opencl? | 16:30 |
| gnarface | is that something you use? | 16:30 |
| gnarface | if it requires it, it should complain when you try to uninstall the firmware | 16:31 |
| gnarface | (complain or also try to uninstall the opencl libraries) | 16:31 |
| gnarface | sigh, i possibly should have made sure RhineDevil had a way to get back here if it all went sideways | 16:51 |
| gnarface | does this problem look familiar to anyone? | 16:52 |
| RhineDevil | gnarface: tried doing as you said | 17:28 |
| RhineDevil | It introduced a major regression, X didn't even boot up | 17:28 |
| gnarface | oh, bummer | 17:29 |
| RhineDevil | Anyway now I'm really sure the problem is launching stuff using OpenGL and then opening a video | 17:29 |
| gnarface | sorry about that | 17:29 |
| RhineDevil | In particular I used -v on mpv | 17:29 |
| RhineDevil | Dw | 17:29 |
| gnarface | did you test the same sequence of programs on the chimaera live iso yet? | 17:29 |
| RhineDevil | When it loaded vaapi the screen started flickering | 17:29 |
| RhineDevil | The program is a nonfree game | 17:30 |
| RhineDevil | Well, all nonfree games using OpenCL I think | 17:30 |
| RhineDevil | To be fair what these games have in common is using unity3d | 17:30 |
| gnarface | test the chimaera live iso, we need to know | 17:31 |
| RhineDevil | I don't know how to replicate the issue with free programs | 17:31 |
| gnarface | did you try ioquake3? | 17:31 |
| RhineDevil | No what's that | 17:32 |
| gnarface | eh, you can still run your non-free programs from the chimaera live iso | 17:32 |
| gnarface | ioquake3 is a open source port of quake3, just add maps | 17:32 |
| RhineDevil | Yeah I know but they are on an encrypted partition | 17:32 |
| gnarface | i don't think it'll use opencl but it'll at least use opengl | 17:32 |
| RhineDevil | gnarface: do I have to download non-free stuff? | 17:32 |
| buZz | gnarface: quake3 itself is already opensource ;) | 17:32 |
| RhineDevil | Pardon | 17:32 |
| RhineDevil | Using OpenGL | 17:32 |
| RhineDevil | I said OpenCL by mistake | 17:33 |
| gnarface | buZz: well, yes, but ioquake3 is the debian-packaged one that's most faithful to the functionality of the commercial build | 17:33 |
| buZz | right right | 17:33 |
| buZz | there's also a game using that engine with only free data | 17:33 |
| buZz | cant remember the name though | 17:33 |
| gnarface | xonotic? | 17:33 |
| buZz | no | 17:33 |
| buZz | its in debian/devuan | 17:34 |
| RhineDevil | Whatever game if it uses OpenGL is fine | 17:34 |
| RhineDevil | I THINK | 17:34 |
| buZz | 'dreamchess' is in debian and uses opengl | 17:34 |
| buZz | gnarface: i think it had 'arena' in the name | 17:34 |
| buZz | not sure though | 17:34 |
| buZz | oh yes! openarena | 17:35 |
| buZz | thats the one | 17:35 |
| RhineDevil | Let's try if it crashes again | 17:35 |
| buZz | openarena/stable 0.8.8+dfsg-5 amd64 | 17:35 |
| buZz | fast-paced 3D first-person shooter | 17:35 |
| gnarface | RhineDevil: well, you can test basic opengl with glxgears from the mesa-utils package, but it doesn't even have textures. i mentioned ioquake3 just because it's the open source thing i could think of with the most intensive opengl reuiqrements | 17:35 |
| gnarface | buZz: openarena is just the other open source quake3 codebase, but they've not made a specific commitment to keeping it like the original | 17:35 |
| buZz | i think debian's ioquake3 packages include a tool that can download the demo/shareware files | 17:36 |
| gnarface | RhineDevil: (openarena should be just fine too) | 17:36 |
| buZz | openarena includes all new levels, which are foss too | 17:36 |
| gnarface | i can't think of much that would require more opengl features, pickings are slim with the open source stuff | 17:36 |
| buZz | mpv -vo gpu | 17:36 |
| buZz | :D | 17:36 |
| gnarface | heh, yea but mpv is our problem child here, we're trying to isolate a triggering cause | 17:37 |
| buZz | hehe ok | 17:37 |
| RhineDevil | Ah anyway | 17:37 |
| gnarface | need a way to test on chimaera to see how far back this bug goes | 17:37 |
| RhineDevil | As I said | 17:37 |
| RhineDevil | When mpv loaded vaapi the screen started flickering | 17:37 |
| gnarface | RhineDevil: actually, that would be a good test. see if just using "glxgears" is enough opengl to trigger the problem in mpv | 17:38 |
| gnarface | i don't think it will be but if it were that might shed some light | 17:38 |
| RhineDevil | When mpv said it decided the decoder (h264_vaapi) the video stack totally crashed | 17:38 |
| gnarface | hmmm, you know, my card can do vdpau... i wonder if yours can't | 17:39 |
| RhineDevil | No it isn't glxgears | 17:39 |
| RhineDevil | I mean | 17:39 |
| RhineDevil | glxgears doesn't trigger anything | 17:39 |
| RhineDevil | mpv still works fine | 17:39 |
| gnarface | interesting. so opengl alone isn't enough to trigger it.... i only know of vaapi libraries for intel hardware, i wonder if they'd work for your card though... | 17:40 |
| gnarface | RhineDevil: do you have any of these packages installed? libvdpau-va-gl1 intel-media-va-driver i965-va-driver | 17:41 |
| RhineDevil | Stuff still works | 17:41 |
| gnarface | libvdpau-va-gl1 - VDPAU driver with OpenGL/VAAPI backend | 17:41 |
| gnarface | specifically i'm wondering if this might be useful | 17:41 |
| gnarface | (since it's the one of the 3 not specifically labeled as for intel hardware) | 17:41 |
| gnarface | i don't have any of those installed here, but like i said that's because this generation of cards can use vdpau directly, but i don't think that was always the case with earlier generations | 17:42 |
| RhineDevil | I have it installed | 17:43 |
| gnarface | (my thought was just, grasping at straws here mind you, but if the issue is actually some bad interaction with vaapi maybe you can simply avoid it in favor of vdpau) | 17:43 |
| gnarface | you'd still want to report the bug of course in that case, but at least then you'd have a workaround | 17:43 |
| RhineDevil | Yeah | 17:44 |
| RhineDevil | Should I try if it happens with openarena? | 17:44 |
| gnarface | yea, go for it | 17:44 |
| gnarface | isolating the exact driver features that are causing this is important for diagnosis | 17:44 |
| _ds_ | Hmm. AFAICT, HD 8210 is GCN, so should be possible to use amdgpu with it (it'll be GCN1). | 19:34 |
| _ds_ | Much of §2.2 at https://wiki.archlinux.org/title/AMDGPU#Enable_Southern_Islands_(SI)_and_Sea_Islands_(CIK)_support is applicable. | 19:35 |
| _ds_ | (RhineDevil ↑) | 19:37 |
| RhineDevil | _ds_: which is it southern islands or sea islands? | 23:04 |
| _ds_ | I initially thought SI, but I wouldn't rule out CIK… | 23:05 |
| gnarface | RhineDevil: what was the result testing openarena? | 23:12 |
| RhineDevil | gnarface: still haven't tested, will do soon | 23:45 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!