| al1r4d | ximon49, download from devuan repoistory | 00:28 |
|---|---|---|
| al1r4d | $ apt info telegram-desktop | 00:28 |
| al1r4d | Package: telegram-desktop | 00:28 |
| al1r4d | Version: 4.14.9+ds-1.1+b2 | 00:28 |
| al1r4d | ah, he's gone | 00:29 |
| amarsh04 | waiting for VAAPI hardware decode on radeonsi to work reliably without gpu lock-up | 05:28 |
| gnarface | amarsh04: which kernel and driver versions? which devuan release? which GPU? | 05:52 |
| gnarface | i'm not having problems with it here on daedalus, at least not over time periods of less than a few hours (haven't tested for longer than that recently) | 05:54 |
| gnarface | but my GPU is a post radeonsi model, i think... i'm wondering if you've looked into trying the special radeonsi specific kernel scheduler in an attempt to combat this | 05:55 |
| gnarface | i think there was also something about a range of cards that would auto-detect as using the radeon driver to X11 but work better with the amdgpu driver anyway | 05:55 |
| gnarface | i forget the specifics now | 05:56 |
| gnarface | make sure you're not just misdiagnosing a video player memory leak as a GPU lockup though, i was having some problems with mplayer leaking somewhat recently... | 05:56 |
| gnarface | this has all come up before but i can't remember if you were the one complaining about it or someone else | 05:57 |
| amarsh04 | gnarface, recent Firefox releases where VAAPI is enabled by default (137.X.X-X) with radeonsi (AMD Ryzen 5 5500GT with Radeon Graphics) and any version of mesa that supports VAAPI on radeonsi (mesa 24.2.X onwards including git head) and any kernel including Linus' git head. Confirmed by mesa developers: https://gitlab.freedesktop.org/mesa/mesa/-/issues/12528 | 06:03 |
| amarsh04 | confirmed with mpv as well as Firefox, was not aware of the radeonsi-specific scheduler. Do you have any more information on the radeonsi-specific scheduler? | 06:04 |
| gnarface | i just remember it's old and you might have to patch the kernel for it... if i understand right that you have a GPU built into your CPU, that sounds too new for it (though it sounds too new for radeonsi in general in that case too, so i'm a bit confused) | 06:05 |
| gnarface | do you also use amdgpu with this thing? | 06:06 |
| amarsh04 | no problem, mine uses amdgpu only | 06:06 |
| gnarface | it seems we have to disambiguate the 'radeonsi' component because it can refer to a kernel scheduler, a mesa library, a X11 driver, or a now outdated hardware series much older than that CPU | 06:06 |
| amarsh04 | mesa library in my case | 06:07 |
| gnarface | this is all stock daedalus versions of everything? | 06:07 |
| gnarface | all i can get in the search results when looking for the radeonsi scheduler thing i remember is just newer reports of high profile changes to some amdgpu scheduler instead that i wasn't aware of | 06:11 |
| amarsh04 | a lot of stuff is newer, I just keep re-running the test case when anything relavent is updated. | 06:12 |
| UsL | how do you "kickstart" an wifi adapter that has it's drivers, and is according to ifconfig and ifup alive and well. It won't show any wireless ap and when I manually add my router and save it it still does nothing after ifdown/ifup | 17:01 |
| UsL | this is on daedalus with xfce and networkmanager 1.30.0 | 17:02 |
| UsL | the chip is bcm43228 and according to https://wireless.docs.kernel.org/en/latest/en/users/drivers/b43.html it's supported | 17:04 |
| cousin_luigi | UsL: Is the firmware installed? | 17:05 |
| cousin_luigi | UsL: Also is there possibly a hardware or software switch involved? | 17:05 |
| UsL | the hardware switch does nothing (unk own keycode) but the applet allows me to enable and disable wifi and the led indicates as much | 17:06 |
| cousin_luigi | UsL: What does rfkill say? | 17:07 |
| plasma41 | UsL: The broadcom 43xx series is a real pain when it comes to firmware. See the firmware-b43-installer package in contrib to learn more. | 17:07 |
| cousin_luigi | I vaguely remember an open source firmware. | 17:07 |
| UsL | plasma41: that's the one I installed | 17:07 |
| UsL | and dmesg says it's okay | 17:08 |
| UsL | it just won't engage | 17:09 |
| plasma41 | "Have you tried turning it off and on again?" - The IT Crowd | 17:09 |
| UsL | well, yeah. With ifup/down as I mentioned. No cigar | 17:09 |
| UsL | but are there any other ways to make it go? | 17:10 |
| plasma41 | Are you sure the adapter isn't defective? Does it work with another system or OS? | 17:13 |
| UsL | it worked fine with an old mxlinux from 2018 | 17:14 |
| UsL | so I am a bit stumped to why actually | 17:14 |
| UsL | devuan doesn't have the broadcom-sta-dkms packages. Is it systemd dependent? | 17:14 |
| UsL | can't find it here: https://pkgmaster.devuan.org/bannedpackages.txt | 17:15 |
| plasma41 | UsL: The broadcom-sta-dkms package is in non-free. | 17:17 |
| UsL | np | 17:17 |
| UsL | I can take it | 17:18 |
| UsL | https://packages.debian.org/search?keywords=broadcom-sta-dkms | 17:18 |
| UsL | it's in bookwrom but not daedalus | 17:18 |
| UsL | but not in devuan non-free | 17:18 |
| plasma41 | https://pkginfo.devuan.org/cgi-bin/policy-query.html?c=package&q=broadcom-sta-dkms shows it in daedalus | 17:20 |
| UsL | ... | 17:20 |
| UsL | okay, so I didn't search for the exact name just "broadcom" and it didn't show | 17:21 |
| UsL | but now when I searched for exact package name it did show | 17:21 |
| UsL | weird. | 17:21 |
| plasma41 | https://tracker.debian.org/pkg/broadcom-sta | 17:21 |
| plasma41 | Weird indeed | 17:21 |
| UsL | I am following this: https://wiki.debian.org/wl | 17:33 |
| UsL | I know I said "np I can take it" but to be honest, I am not super comfortable with installing this. | 17:34 |
| UsL | okay, let's see if this was successful.. Bye, tand thanks cousin_luigi and plasma41 | 17:53 |
| UsL | well, finally. | 18:15 |
| UsL | I got a lot of [ 423.003710] WLC_SCAN error (-22) errors.. And I disabled the wifi and tried again. Four times. And voila. Here I am on glorious wifi. | 18:16 |
| UsL | kernel feels violated though: | 18:16 |
| UsL | [ 14.451914] wl: loading out-of-tree module taints kernel. | 18:16 |
| fsmithred | UsL, ifup/down should be used if you configured the interface in /etc/network/interfaces, but if you did that you should not have network-manager running. And vice versa. They conflict with each other. | 18:33 |
| UsL | I tried everything | 18:47 |
| UsL | but yeah, stay away from broadcom. Almost a dealbreaker. | 18:48 |
| UsL | also, I miss wicd. | 18:49 |
| greenjeans | we all miss wicd | 18:50 |
| djph | ^ | 18:52 |
| golinux | Several folks started to work on it but . . . | 19:14 |
| tempforever | i believe there's a working wicd ... not sure if i bookmarked it tho | 19:14 |
| golinux | There are somem beginning attempts in our git | 19:15 |
| golinux | IIRC the issue was python 2 to python 3 | 19:15 |
| tempforever | right, and some gtk issues i think too | 19:15 |
| golinux | tempforever: Hope you can find it! | 19:16 |
| tempforever | https://dev1galaxy.org/viewtopic.php?id=6842 | 19:17 |
| golinux | I avoided the problem by only having a static wired connection | 19:17 |
| UsL | golinux: haha, I was this close ' ' to do that as well. Before I tried this last modprobe lw | 19:20 |
| golinux | I think that plasma41 poked it for a while after to person working on it had to bow out. | 19:21 |
| golinux | to > the | 19:22 |
| golinux | Nice to see you UsL . . . it's been a while! Don't be a stranger!! | 19:32 |
| tsadok | For systems that don't need to be portable, a static wired connection is absolutely the way to go. But it's a bit problematic for systems that you carry from place to place. | 19:46 |
| golinux | tsadok: Thankfully, I don't own any! | 19:56 |
| tsadok | Heh. I have one laptop, and don't take it anywhere very often; but I think it's important to acknowledge that some people have a much more on-the-go lifestyle. | 19:58 |
| cousin_luigi | UsL: What is it, 802.11g? | 20:01 |
| UsL | golinux: oh, I never left :) . Only my physical body was busy for.. A couple of years. I'll be around : ) | 20:09 |
| UsL | cousin_luigi: Broadcom Inc. and subsidiaries BCM43228 802.11a/b/g/n | 20:10 |
| UsL | horrible. | 20:11 |
| cousin_luigi | UsL: I think I have one, somewhere. | 20:11 |
| cousin_luigi | Is it mini-pci? | 20:11 |
| UsL | burn it | 20:11 |
| UsL | no, this is a hp probook | 20:11 |
| cousin_luigi | so, what format? | 20:11 |
| cousin_luigi | I don't think I could even use it anymore, no slots. | 20:11 |
| UsL | ...? | 20:11 |
| UsL | laptop? | 20:12 |
| UsL | ah, yeah. | 20:12 |
| cousin_luigi | But it's got to be somewhere. | 20:14 |
| peterrooney | i was dealing with a b43-inflicted laptop last night, dell. "where the hell are the ethernet cables" escaped my lips | 20:21 |
| greenjeans | lol, i remember the old b43-wadcutter...yikes | 20:24 |
| Wonka | hehe | 20:26 |
| Wonka | I've been living with a BCM4333 for 13 years | 20:27 |
| Wonka | Android's USB Tethering was very nice when away from wired connections | 20:28 |
| cousin_luigi | Still, I'm fairly certain someone developed an open-source firmware for b43. | 20:34 |
| cousin_luigi | I packaged it for openSUSE, many years ago. | 20:34 |
| UsL | yeah, using the phone as a middleman. But I need to walk away with the phone, and I like being connected 24/7 with the laptop. So, if push had come to shove I'd drill down the floor into the living room with a cable : ) | 20:36 |
| fsmithred | usb wireless dongle $10-15 | 20:36 |
| UsL | but on the go when traveling usb tethering is gold | 20:36 |
| UsL | fsmithred: it's an alternative for sure, but the wifi in the laptop was paid for at one time and I guess I would think about that and get annoyed.. : ) | 20:40 |
| UsL | peterrooney: yeah, I was close. | 20:40 |
| cousin_luigi | UsL: That presupposes that your time is not worth much. | 20:42 |
| UsL | does it? | 20:54 |
| cousin_luigi | UsL: If you spend it to make an obsolete component work, a component that will also waste more of it by being slow, I suspect it does. | 20:56 |
| UsL | ah, but the point was for me that a device has a function that's been developed, manufactured, sold, and put into a laptop and sold again. It's purpose is to function. It would have bothered me If it didn't. I dislike being wasteful. | 21:01 |
| XFaCE | Hello all. I tried installing some ZFS backport packages today and have encountered service loop warnings. I was going to compare the old and new service scripts but I thought I should mention it here as well while I debug. This is default sysvinit. | 21:03 |
| XFaCE | (Daedalus) | 21:03 |
| XFaCE | From my reading, such loops are a sign of a package bug rather than user error. | 21:03 |
| XFaCE | This seems wrong... | 21:26 |
| XFaCE | # Provides: zfs-mount | 21:26 |
| XFaCE | # Required-Start: zfs-import | 21:26 |
| XFaCE | # Required-Stop: $local_fs zfs-import | 21:26 |
| XFaCE | # Default-Start: S | 21:26 |
| XFaCE | # Default-Stop: 0 1 6 | 21:26 |
| XFaCE | # X-Start-Before: mountall | 21:26 |
| XFaCE | Sorry, pastebin link: https://bpa.st/SUIQ | 21:28 |
| XFaCE | I'm thinking that definition is probably not right given that load-key is at higher runlevels than mount but it supposed to occur before mount | 21:29 |
| UsL | in my day and age I had flood prptection set to ten lines. But that was when all connected on dialup and connections were much more robust than today's fiber optical connections | 21:30 |
| XFaCE | Ha | 21:30 |
| XFaCE | UsL: Should I report this to the dev channel, assuming there is a devuan dev channel? | 21:32 |
| XFaCE | The ZFS backport issue that is. | 21:32 |
| UsL | the forum or mailing list is the best bet | 21:32 |
| UsL | since it's more static | 21:32 |
| XFaCE | Understood | 21:32 |
| UsL | you can also idle here as long as you wish | 21:33 |
| UsL | perhaps some zfs guru will return | 21:33 |
| XFaCE | Well I would prefer to get the word out so the issue is fixed. My first workaround instinct would be to put back the older service definitions | 21:34 |
| XFaCE | But I'm not as familiar with sysvinit or Devuan as a whole | 21:34 |
| XFaCE | UsL: I'm also not sure if it's really a ZFS issue as much as just a general service writing issue. I could share the difference between the stable repo ZFS package. | 21:35 |
| UsL | ^ | 21:36 |
| XFaCE | https://bpa.st/MV5Q | 21:40 |
| josh_ | Hi | 21:46 |
| XFaCE | UsL: Ah, my bad, these scripts are coming all the way from OpenZFS upstream... and it looks like they made a change to match systemd without proper checking, though it is such an obvious mistake that I can't help but wonder if I'm just reading incorrectly | 21:57 |
| UsL | XFaCE: that follows, systemd breaking or forcing other things to break | 23:40 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!