01:11:59 is RTL8811AU still not supported? I've got a Netgear A61000/AC600 01:19:20 well, there is a package in the opnsense community repo that I had to use just tonight for a new opnsense install. that was for realtek 8125b I think. the 2.5G ports weren't recognized until i installed that plugin 01:19:36 https://ibb.co/album/wyBtKN <-- my version of chicken parm. sorry for quality. 01:45:39 Matt|home: But, where's the parm!? 02:15:14 https://i.ibb.co/hxBHCxh8/IMG-20250919-210227908.jpg <-- right there ek 02:15:37 i tend to go light on the cheese since i can only afford that canned bullshit and it's .. quite a bit more salty than the blocks 02:48:27 Matt|home: That's fair. 02:48:43 The chicken should have melted parm (when possible). 03:19:26 Does the NVIDIA driver on FreeBSD support NVDEC & NVENC for hardware assisted transcoding of video? For use in Emby (which uses ffmpeg) 08:10:28 hello guys? 09:00:08 i guess NVENC requires CUDA, so quick answer is probably no 09:01:53 any freebsd people here? 10:55:13 scottpedia: Asking the real question usually helps. 11:43:44 Random hypothetical ... would something like L4Linux or MkLinux be a viable approach for a (better than current) Linux compat layer in BSD? 11:43:50 I just ran into L4Linux, and it made me think a bit, heh. 11:45:46 At that point, you'd pretty much be running Linux in a context similar to a xen guest, with FreeBSD as the domU. I think the main issue would be the requirement to run GPL'd code in kernel-space 11:46:56 But yeah, FreeBSD already has a hypervisor capable of running linux, bhyve 11:47:43 and it seems that as Xen too: https://docs.freebsd.org/en/books/handbook/virtualization/#virtualization-host-xen 11:48:59 And it's been so many years that I've last used Xen that I mixed it up, I meant to say dom0 instead of domU, domU are the unprivileged instances, so that's where you'd put Linux in this example 11:51:02 Xen used to be really useful in the days before CPU virtualization extensions were common, they allowed you to run multiple kernels regardless by replacing some of the CPU-specific functions with generic functions that could either access the real CPU or just a virtualized instance 12:03:14 hello, does FreeBSD has something like /usr/local/share/doc/pkg-readmes udner OpenBSD? 12:03:24 that's it, the hints after installing a package, but as a file 12:05:48 anthk_: pkg-info has the --pkg-message option 12:06:11 man pkg-info 12:12:15 DaliborFox: I was mostly thinking about devices without HW virtualization and such. 12:12:26 The nice thing about L4Linux is that it's paravirtualized. 12:12:47 Xen is paravirtualized too 12:12:51 (also *maybe* there'd be a way to get the two kernels to talk to eachother, to be able to use the exact same filesystem without relying on NAS, etc?) 12:14:34 I think Xen can do that too, but I'm not sure about the FreeBSD integration of that particular feature, thin provisioning of storage space 12:15:12 TL;DR, while most of my services run on FreeBSD, I have a handful that need a Linux kernel (and the compat layer isn't adequate). 12:15:23 There's also (potentially, mostly for others) Docker Linux images and whatnot. 12:16:03 It'd be nice to be able to actually mount a Docker Linux image's storage into a FreeBSD mountpoint, which you can't at the moment --- not without using networked storage of some sort inbetween such as p9fs or NFS. 12:16:41 DaliborFox: to be clear, I'm talking about actually doing (effectively) a null mount from one system into another, *not* about simply allocating a block of storage that's more or less opaque to the host. 12:19:21 Okay, looks like I was wrong: https://serverfault.com/questions/294070/xen-domu-shared-drive-alternative-to-nfs 12:20:04 divlamir: thanks 12:21:50 DaliborFox: ah well. 12:23:44 I wonder, is there a way to at least do a read-only mount of a FS that's actively used by another OS? (doesn't have to be any *particular* FS, as long as it supports all the usual Unix stuff like permissions, sockets/pipes/etc as files, and so on. (which rules out a few like FAT32) 12:23:50 In either case, paravirtualization doesn't actually bring much of a performance benefit with it anymore; in fact, according to the handbook, FreeBSD got rid of support for paravirtualization in FreeBSD 11, since hardware virtualization had better performance 12:24:13 DaliborFox: you missed the part where I mentioned HW that doesn't *have* hardware virtualization ^^ 12:24:30 Whoops, you're right 12:24:33 And anyway, the main point of me mentioning this was to have the two kernels talk directly to eachother (which is easier with paravirt) 12:24:45 But I recognize that it's a very niche use-case nowadays. 12:25:07 Could be an interesting research project, but I'd imagine it would take a fair bit of effort to maintain 12:25:58 Definitely. I feel like this is something Geode could make use of, for one. I mean, the idea of embracing paravirtualization "heavily". (maybe it does?) 12:26:59 To be honest, despite the scalability issues, I feel like something like XenFS would still have some very real use-cases. 12:27:14 (e.g. CI runners) 12:32:17 One thing that's quite interesting about virtualization is that it's actually been here since the 60s, and the time that virtualization extensions "disappeared" (how long it took for them to appear on consumer hardware) was actually pretty short 12:59:20 hello guys 12:59:49 how am I supposed to reach out to the team if I have an idea that might see potential implementation? 13:16:40 Not a FreeBSD developer, but I think the best way is probably to ask on the appropriate mailing list (see https://lists.freebsd.org/ for which MLs are available) for the area of FreeBSD this is about, or if none are applicable, maybe freebsd-hackers@. 13:19:23 ok thx will try V_PauAmma_V 14:07:55 i have a very weird network anomaly. i have bhyve-vm and wireguard on my desktop. currently no vm running, but two wireguard connections up. bhyve-vm has created that weird vm-public bridge-ish thing. 14:09:28 something, somewhere in this weird construct seems to redirect data from a public domain/ip to one of the wireguard hosts. anyone got an idea how this could happen because i sure don't… o_O 14:10:18 imagine my face when i did an nmap on this public customer domain, found port 9090 open and seemingly leaking the entire monitoring data of my *private* infrastructure… /o\ 14:12:51 i had a quick look at it with wireshark, and according to it, the data i'm seeing comes from the actual, public IP of the customers domain. but from any other host (i.e. without the weird vm-public thing), there's no open port 9090 on that host. 14:18:41 wireguard hosts are all in the 10.0.0.0/8 range, and it doesn't collide with the public domain in any way (that's in 91.0.0.0). 14:20:10 what's actually in the bridge here? 14:20:31 my local setup with vm-bhyve only has a few taps and uses natd -> wifi 14:20:50 (<-> bridge) 14:22:52 vm-public currently has no addresses, but re0 as a member. re0 itself has an address, i.e. the main one for the host itself. 14:24:38 how are the wireguard interfaces configured, exactly? manually, or with wg-quick / the wg rc script from wireguard-tools? 14:24:45 also, what allowed-ips? 14:25:21 rc script, i think that uses wg-quick in the background tho? 14:25:26 yeah, it does 14:26:02 and allowed ips are also in 10.0.0.0/8, but all in different subnets per wg interface. 14:32:18 wait, i think i got myself confused by too many different hosts here… it's not getting redirected to one of the desktops' wireguard hosts, but actually one of my homeservers/gateways jails. 14:35:17 ah yes, the joys of just winging changes to pf config when chaning network topology. apparently i'm just wholesale redirecting anything from the LAN trying port 9090 to the jail m) 14:35:28 whoops 14:36:17 i guess that's PEBCAK in a way 14:36:41 definitely, but at least the mystery has been solved.^^ 14:36:57 yep :-) 14:39:01 if i'm lucky, finding this might motivate me to redo my fw from scratch and add some egress filtering – have been meaning to do that for at least 2 years now… 17:35:17 phryk: who is doing NAT for you leakage ? 18:06:29 * mosaid pkg remove chromium 18:06:30 Finally 18:16:45 After 5 years of only using chrome.. I moved back to Seamonkey (a fork called Brassmonkey uses Goanna), surly it's slower than chrome, but works well 18:17:36 I use WaterFox, pretty happy with it. 18:17:50 I love full suite 18:17:57 So I am using it 18:18:04 Yeah, fair. 18:18:22 and brassmonkey supports NPAPI 18:18:34 with modern Html5 support 18:19:27 Imported all my bookmarks from 2011 til now, and passwords 18:20:00 * mosaid really want to get back his pre2011 lost ones.. 18:58:55 does anyone have access to real powerpc64le hardware and is comfortable building main? 18:59:09 or isn't comfortable but i can give you an ISO to test 19:00:05 mosaid: I got so sad about transferring and re-importing links that I just turned my bookmarks into a self-hosted webpage 19:00:37 sad :( 19:01:07 I nearly lost all my bookmarks before 2011 back in the day 19:01:52 Bookmarks are very useful for me, i have +5K 19:02:15 https://www.gluecode.net/blog/links.html 19:02:36 Get yourself a domain. Export what you have. Start typing when you get a new one 19:02:47 a new bookmark* 19:03:57 Nice, but my bookmarks contain some private stuff 19:04:15 also nice domain.. main is occlub.org 19:04:38 *mien 19:06:42 *my one 21:03:27 when can we expect a few packages for 16 to be available? 21:05:04 rtprio: 3,500 packages left, so i'd expect a few days + mirror sync time: https://pkg-status.freebsd.org/beefy18/build.html?mastername=main-amd64-default&build=p118fb2971704_s4ab64e34911 21:05:47 does the batch have to finish before any are available? i guess they would or a chunk of dependancies would be unavailable half the time 21:06:03 yes, the entire build has to finish before anything is copied to the repository 21:11:50 hi :D 21:18:06 how long does that copy take? 21:18:10 * rtprio bookmarks 21:32:20 ariadna, if you have a question about FreeBSD, channel custom is to ask directly without waiting to be recognized. 21:33:17 oh I see I accidentally sent the message here instead to a different channel 21:33:19 my apologies 21:44:30 No worry.