| - mtm (QUIT: Ping timeout: 248 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 02:04 | |
| - chomwitt (QUIT: Ping timeout: 244 seconds) (~chomwitt@2a02:587:dc09:3900:f527:3ca7:b122:fd0c) | 02:42 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 03:53 | |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 04:10 | |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:17 | |
| + Ar|stote|is (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:18 | |
| - Ar|stote|is (QUIT: Client Quit) (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:19 | |
| + Ar|stote|is (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:19 | |
| - Ar|stote|is (QUIT: Client Quit) (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:22 | |
| + Ar|stote|is (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:22 | |
| - Ar|stote|is (QUIT: Client Quit) (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:25 | |
| + Ar|stote|is (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:25 | |
| - Ar|stote|is (QUIT: Client Quit) (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:25 | |
| + Ar|stote|is (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:26 | |
| - Ar|stote|is (QUIT: Client Quit) (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:26 | |
| + Ar|stote|is (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:26 | |
| - Ar|stote|is (QUIT: Client Quit) (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:26 | |
| + Ar|stote|is (~linx@149-210-4-123.mobile.nym.cosmote.net) | 05:27 | |
| Asmadeus | did anyone try https://github.com/bootlin/libva-v4l2-request instead of the unmaintained v4l2-request ffmpeg patches I backported last year? ffmpeg can use libva which can use v4l2-request so it should somehow work out and might be better than keeping a patched ffmpeg around | 09:44 |
|---|---|---|
| - robin (QUIT: Ping timeout: 244 seconds) (~robin@user/terpri) | 12:34 | |
| sigrid | i tried and failed with void linux | 12:35 |
| sigrid | used clapper instead | 12:36 |
| eery | I think I got it to build once but never got anything to work with it | 12:36 |
| + robin (~robin@user/terpri) | 12:47 | |
| Boostisbetter | eery: you are running VMs on your Reform? | 12:57 |
| eery | well, not all the time with 4GB of memory, but I've tested Windows + NetBSD + Ubuntu so far and all boot/are functional | 13:01 |
| eery | NetBSD in particular only uses up 20MB of memory from the base image, so it's more practical than the others | 13:01 |
| Boostisbetter | Wait you got Windows to boot in a VM on the Reform? | 13:05 |
| eery | Yup, minute did too a while back, posted about it on social media I think | 13:08 |
| eery | it peforms fairly well, but you need to give it ... basically all the RAM, so it's of limited use | 13:08 |
| Boostisbetter | This is WIndows 10? Running x86 or ARM? | 13:09 |
| Boostisbetter | either way that is ok course impressive. | 13:09 |
| eery | Win11 for ARM, though win10 should also work (the images are just harder to get now) | 13:10 |
| eery | it does have a built-in x86 translation layer, which works reasonably well | 13:10 |
| Boostisbetter | I suppose in theory that means you could boot Win11 cold | 13:10 |
| eery | well, if someone implemented UEFI for Reform | 13:11 |
| eery | UEFI and ACPI I guess | 13:11 |
| eery | I have no idea if that's even remotely possible, I assume it's technically possible but an enormous time sink with little benefit (besides booting Windows) | 13:12 |
| eery | and it'd make booting random OSes without custom u-boot scripts easier, but still, enormous time sink to implement | 13:12 |
| Boostisbetter | yeah, I mean if you can run it via hypervisor who really cares. | 13:12 |
| Boostisbetter | I just think it is cool that the SoC can do it. | 13:12 |
| Boostisbetter | Very impressive. | 13:13 |
| eery | Yeah, as crappy as the A53 cores are they do at least support e.g. virtualization extensions | 13:14 |
| eery | obviously it'd be more useful on one of the alternate SoMs with 8GB+ of memory :) | 13:15 |
| + ben_ (~ben@71.168.125.51) | 13:46 | |
| - ben_ (QUIT: Client Quit) (~ben@71.168.125.51) | 13:50 | |
| - buckket (QUIT: Remote host closed the connection) (~buckket@pdp8.buckket.org) | 13:57 | |
| + buckket (~buckket@pdp8.buckket.org) | 14:02 | |
| - mtm (QUIT: Ping timeout: 256 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 14:04 | |
| - qbit (QUIT: Quit: WeeChat 3.5) (~qbit@h.suah.dev) | 14:06 | |
| - ex-parrot (QUIT: Ping timeout: 268 seconds) (~fincham@user/ex-parrot) | 15:00 | |
| + ex-parrot (~fincham@user/ex-parrot) | 15:02 | |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 15:09 | |
| lastebil | hmm did we get netbsd running on the reform? | 15:28 |
| sknebel | afaik eery was trying it, but none of the BSDs have code to run the GPU or sth like that? | 15:32 |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 15:37 | |
| Boostisbetter | yeah minute has a couple of boards on deck to help increase the amount of RAM the Reform is able to use. I think 4gb RAM with an NVME drive is actually pretty good. Your NVME will act as a swap and you will very rarely have any slow downs due to paging. | 15:38 |
| lastebil | yeah, the gpu is likely the issue with the bsd's, could likely run serial, but ... I have sbc's for that (: | 15:44 |
| lastebil | I'll come back to this in a few hours and finally get mine set up properly (finally have time, it's what, 1.5 years now?) | 15:45 |
| sigrid | gpu is irrelevant to getting a framebuffer | 15:48 |
| sknebel | I guess yes, on the imx the pixel pushing is technically not part of the GPU | 15:52 |
| eery | there's no support at all for anything graphics-related on any BSDs for i.mx8, sadly | 16:04 |
| eery | FreeBSD looks like it supports the pixel-pusher bits for the i.mx6, but not 8 | 16:05 |
| eery | they work well on qemu/kvm/uefi though | 16:09 |
| + qbit (~qbit@2602:ff16:3:0:1:3a0:0:1) | 16:12 | |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 16:20 | |
| Boostisbetter | eery: Do you think you have time to kind of walk me through how you got Win11 running in a vm? I don't think I need it anytime soon, but as a Windows developer, it certainly wouldn't be a bad thing to know. | 16:55 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 17:17 | |
| eery | Sure - if you're okay using libvirt/virt-manager I think it should be as easy as downloading the Win11 arm64 installer CD and installing a VM with it, I'll try it and see if it works | 17:22 |
| eery | when I tried it before I spent a lot of time figuring out why tianocore/edk wasn't booting, and that bug seems to have vanished | 17:22 |
| - Boostisbetter (QUIT: Ping timeout: 268 seconds) (4a410829d7@irc.cheogram.com) | 17:43 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 17:53 | |
| Boostisbetter | so libvirt/virt-manager is the hypervisor you'd use? | 18:03 |
| Boostisbetter | I'm thinking I'll give gnome-boxes a shot first, as it feels bit more GUI based and friendly. | 18:07 |
| jfred | gnome-boxes is libvirt under the hood | 18:16 |
| Boostisbetter | well that is perfect then! | 18:16 |
| Boostisbetter | good to know | 18:16 |
| lastebil | ACTION would absolutely NOT use libvirt, but, he has been doing virtualization since Xen 1.7 or so, and finds practically everything with libvirt overly complex and convoluted, along with unneccesary | 18:22 |
| lastebil | that said, most folks don't want to actually know what their machines are doing, so... this sort of thing works great for them. | 18:22 |
| lastebil | - I would just use qemu, straight, and add the things needed in a file... | 18:23 |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 18:25 | |
| kfx | I'm also a professional virtualizer and I think it's a complete waste of time to learn qemu's 54,305,429 command line flags, much less all the ridiculous linux networking mess concomitantly required | 18:26 |
| eery | I don't mean to start an argument, but I use qemu "raw" on occasion and have large "shell scripts" for doing arcane things like PCIe passthrough of an output-less laptop GPU to Windows and all manners of snapshot weirdness, and I still generally prefer libvirt for day-to-day things | 18:44 |
| Boostisbetter | I'm a SE and just don't have the time to learn too many things. I like when things are simplified some. I bet Qemu is great as well, just don't have to the time to mess around with it. | 18:45 |
| eery | IMO virt-manager+libvirt is great, for the "easy path" of Linux distros or Windows the setup process is pretty seamless, and you have a fairly high level of granular control if you need it | 18:46 |
| kfx | eery: don't get me wrong, I've learned qemu's massive ridiculous interface, and I do plenty of dumb networking tricks as well... I just think it would have been a huge waste of time if I weren't a systems administrator for a living | 18:46 |
| eery | oh, sure | 18:47 |
| eery | or screw around with "exotic" architectures ;) on both aarch64 and ppc64le I've run into problems with libvirt where you need to start setting lots of QEMU flags manually, and it's just "easier" to shell script it | 18:47 |
| eery | but that's an extremely niche reason in the grand scheme of things lol | 18:48 |
| eery | also, using qemu directly opens up the possibility of using it all the OSes with supported VM monitors/hypervisor layers, so Linux/Windows/MacOS/NetBSD | 18:49 |
| eery | with a ... somewhat consistent interface | 18:49 |
| kfx | on the flipside learning libvirt is useful if you stumble into a job running ovirt infra | 18:49 |
| kfx | and there's a lot more on-prem cloud out there than people think | 18:50 |
| eery | I introduced libvirt at work and my team's been pretty happy overall | 18:50 |
| Boostisbetter | I have been using VMware Workstation for over 10 years. It just sucks that they are basically discontinuing the product. | 18:50 |
| eery | beats running vmware workstation on windows laptops and being stuck with all crappy options for remote management, vs SSH over wireguard on linux | 18:51 |
| Boostisbetter | hmm I've been using ssh and vnc server myself. | 18:54 |
| Boostisbetter | it has been fine. | 18:54 |
| eery | some of the BSD hypervisors like vmm/vmd and bhyve are really nice on their "happy path", but break down fast if you need something weird | 18:54 |
| eery | Yeah, but you still need a GUI session, compared to being able to pop into a virsh shell and type a few commands | 18:54 |
| Boostisbetter | BTW is anyone else here using Chromium on their Reform? Are you finding that it is hard crashing when doing file operations with downloaded files? | 18:54 |
| eery | uhhh, I'm not sure, I don't know if I've actually tried opening any files from Chromium itself | 18:55 |
| Boostisbetter | eery: can't you just run the ssh server on the guest VM? | 18:55 |
| Boostisbetter | I'm finding Vivaldi way more stable on the Reform, but because it is so top heavy it runs really slow. | 18:55 |
| eery | I mean for managing the VMs themselves, e.g. starting/stopping, changing config, opening a serial console | 18:58 |
| eery | I know back in the old days you could run vmware GSX and get that kind of remote management, no idea if its in any version of Workstation | 18:59 |
| eery | now that's a truly terrifying piece of software -- vCenter | 19:01 |
| Boostisbetter | hahaha, nope always just used Workstation and honestly that client is very good. It is why I have been paying for it. | 19:04 |
| Boostisbetter | I've messed with Virtual Box as well | 19:04 |
| Boostisbetter | Gnome Boxes seems like it will work just fine. | 19:04 |
| Boostisbetter | not worried. | 19:04 |
| eery | Well, if it doesn't, you can always use either virt-manager or the CLI shell virsh to manage the same VMs | 19:05 |
| eery | They all work on the same XML config files | 19:06 |
| kfx | virtualbox is by far the worst. every single release manages to break something fundamental, unless you're running windows or linux I wouldn't touch vbox with a ten foot pole | 19:10 |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 19:24 | |
| eery | ...why windows or linux lol | 19:52 |
| eery | the last time I tried to run through setting up vbox to help a coworker who insisted on using Win11 home (so no hyper-v), I ran into a bug that breaks any VM with more than one logical core | 19:53 |
| eery | the only mitigation was just limiting all your VMs to 1 logical core or downgrading several versions | 19:53 |
| eery | absolutely ridiculous | 19:53 |
| eery | Like, a regular release (not some development build) of virtualbox was incapable of multiprocessing on any VM on windows | 19:55 |
| eery | I figure it works "best" on linux but I would definitely not trust it one second | 19:56 |
| kfx | mostly people come to our project having had vbox break their install via an upgrade, then insist that their ubuntu and windows VMs still worked | 19:57 |
| + chomwitt (~chomwitt@2a02:587:dc09:3900:9994:71e5:52c9:ca3d) | 20:52 | |
| + ajr (~ajr@user/ajr) | 21:02 | |
| - chomwitt (QUIT: Remote host closed the connection) (~chomwitt@2a02:587:dc09:3900:9994:71e5:52c9:ca3d) | 21:26 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 23:29 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!